Enabling the drying process to save energy and water, realising process efficiency in the dairy chain

Final Report Summary - ENTHALPY (Enabling the drying process to save energy and water, realising process efficiency in the dairy chain)

Executive Summary:
Within the European community there is a growing concern regarding sustainability, which is set forth in the objectives of the Europe 2020 strategy. Therefore the saving of water and energy within European industry is an important target. The objective is to significantly reduce the amount of energy and water use, therewith increasing the competitiveness of the European industry
The Enthalpy project has targeted the dairy industry for saving water and energy by introducing innovations and introduce the further development of existing technologies in the dairy sector. Up till now this was not feasible due to technical limitations. The dairy sector is chosen since it is one of the main sectors regarding energy and water usage within the European food processing industry. The dairy industry consists of a wide variety of companies, varying from small SMEs to large industrial companies and all sizes in between. The companies are highly intertwined and all equipment is directly related to each other for input and output. Therefore an overlapping approach is needed, since compatibility to the whole chain needs to be accounted for when changing equipment. The current processes and techniques in the sector have been matured in the past decades. In order to obtain significant improvements in this sector new breakthrough innovations are needed, which are proposed in this project.

Within the Enthalpy project successful implementation of the following technologies was performed in a pilot facility
- Mono disperse atomisation
- Enzymatic cleaning
- Radio Frequency heating
- Membrane distillation and membrane contactor
- Solar thermal energy usage

In order to implement these technologies and make them successful the following competences were used to support the development
- Process systems engineering
- Modelling of drying behaviour
- Food and powder quality analysis
- Life cycle analysis

With these combination of technologies an 45% energy reduction could be obtained compared to conventional process in the conducted benchmark study.
The technologies were demonstrated at two pilot facilities to show their potential and an industrially relevant scale. Beyond the project the steps will be taken to have these technologies adopted in industry and further developed to potentially other sectors, like pharma.
Project Context and Objectives:
The project “ENTHALPY” aims at a significant and simultaneous saving of water and energy in one of the most energy and water consuming sectors of the food industry, the dairy sector. Using a systematic approach the “ENTHALPY” project focusses at innovations at those parts of the post-harvest chain with the highest energy and water consumption. This will ultimately transform the conventional drying system to a fully reorganised and optimised closed loop drying plant.

The following innovations are proposed for obtaining this optimised closed loop dryer plant.

• Atomization based on ink-jet technology to create mono disperse droplets, eliminating fines from the system and enabling step-change improvements in process control and product quality
• Combining both existing and innovative membrane technology for pre-concentration (Reverse Osmosis and Membrane Distillation) and recovery of heat and water from the exhaust air of the spray dryer (Membrane contactor)
• Pre-treatment of dairy feed by application of renewable energy and radio frequency heating
• New cleaning protocols based new process characteristics and breakthrough innovations are introduced
• Integration of the innovations with the help of systematic process engineering (PSE)

Concept and objectives

Within the European community there is a growing concern regarding sustainability, which is set forth in the objectives of the Europe 2020 strategy. Therefore the saving of water and energy within European industry is an important target. The objective is to significantly reduce the amount of energy and water use, therewith increasing the competitiveness of the European industry

This proposal targets the dairy industry for saving water and energy by introducing innovations and introduce the further development of existing technologies in the dairy sector. Up till now this was not feasible due to technical limitations. A three year project “Enabling the drying process to save energy and water, realising process efficiency in the dairy chain” “ENTHALPY” is proposed.

The dairy sector is chosen since it is one of the main sectors regarding energy and water usage within the European food processing industry. The dairy industry consists of a wide variety of companies, varying from small SMEs to large industrial companies and all sizes in between. The companies are highly intertwined and all equipment is directly related to each other for input and output. Therefore an overlapping approach is needed, since compatibility to the whole chain needs to be accounted for when changing equipment. The current processes and techniques in the sector have been matured in the past decades. In order to obtain significant improvements in this sector new breakthrough innovations are needed, which are proposed in this project.

The dairy chain can be subdivided into different segments in the following manner:

One approach of saving energy and water in this chain could be to investigate all steps of the chain individually and try to make improvements. Within this proposal it is believed that the most impact is obtained by introducing innovations in that part which contributes the most to the energy and water usage and focus with promising innovations where the biggest strive forward is made, instead of making smaller increments along the complete chain.

Within “ENTHALPY” energy savings of 63% and water savings of 18% are expected

The energy consumption throughout the dairy chain is not distributed evenly. There are two distinct big contributors, namely the feed and the processing. The energy in the feed is required for the growth and living of the livestock and therefore cannot be significantly reduced and the call aimed at post-harvesting chain. Therefore the Enthalpy project aims at the processing plant.

The processing plant consists of different operating steps (unit operations). All different unit operations have separate water and energy chains, from pasteurization, evaporation, homogenizer, pre-treatment, spraydryer, fluidized bed, cyclone with bag filters to final product and exhaust air. In this conventional processing plant most energy is lost in the form of heat in the exhaust air, the dried product and in the fines. Water and resources/product are lost in the humid exhaust air. Furthermore resources are lost due to fouling throughout the system. Currently the exhaust air cannot be processed, since it contains fines which prevent heat recovery due to fouling of heat exchangers and condensation on the fines. Furthermore fines introduce the risk of dust explosions.

The “ENTHALPY” project addresses the different unit operations of the processing plant which ultimately will lead to closing the loops of energy and water. This will be achieved by elimination of fines in the spray dryer, recovery of the energy and water present in the used drying air and a more efficient pre-treatment process.

Furthermore cleaning of the system is an important part of both the energy and water consumption. The “ENTHALPY” project envisions both direct and indirect improvements for cleaning. By using newly developed enzymes the use of water and energy for cleaning all different equipment in the processing plant is tackled, while also the cleaning intervals will be addressed. By using mono disperse printing better control of the droplets is obtained in the dryer, which will result in less fouling and therefore decreasing the cleaning frequency and thereby indirectly reducing the amount of water and energy required. Furthermore membrane technology is used to win clean water out of the processing plant, reducing the need for fresh water intake.

In order to improve the competitiveness of the European dairy industry also the aspect of food quality will be taken into account. The mono disperse atomising technology provides better process control in the spray dryer and therefore opens the opportunity for accurate modelling of the behaviour of the droplets in the spray dryer and as a consequence control of the desired powder quality. End users within the project will define the desired properties related to the food product and how this correlates to the powder properties. By this coupling it will become possible to predict and tune the powder properties in order to obtain a desired product.

Within the scope of the “ENTHALPY” project both novel as well as existing technologies (which can be implemented since barriers, e.g. presence of fines, have been removed) will be further developed, with the aim of being applied in both existing as well as new dairy/food processing facilities.

The technologies will be demonstrated at a pilot scale facility to show industrial relevance and prove the technologies work in such an environment compared to lab scale proof of concept validation.

The corresponding objectives on the different involved topics are
“ENTHALPY” innovation regarding cleaning
• The development of enzymatic detergents specifically directed to remove fouling formed in dairy processing equipment

“ENTHALPY” innovation regarding pre-treatment
• Low fouling during heating
• Renewable energy sources

“ENTHALPY” innovation regarding atomization
• High viscous jetting to create a mono disperse distribution
“ENTHALPY” innovation regarding drying tower
• Coupled droplet drying to CFD modeling
• Inline spray monitoring

“ENTHALPY” innovation regarding membrane technology
• Closed loop dryer with full heat and gas recycle, and water recovery
• Recovering fresh water from waste streams
• World’s first membrane contactor for gas/liquid operating at high temperature
• Full replacement of evaporator by RO/MD

“ENTHALPY” innovation regarding process system engineering
• ENTHALPY delivers a framework that integrates process systems engineering and LCA
• A demonstration on how the LCA impact of a total production system is reduced by process optimization

Project Results:
WP1: Process systems engineering

This task concerned the search for most efficient use of the innovative processing methods considered in the project (reverse osmosis, membrane distillation, and closed-loop monodisperse droplet drying with air dehumidification). From a superstructure all possible processing routes from milk feed to skimmed milk powder were defined.

In order to define the superstructure a logical overview of all potential combinations and routes has to be made. The choices for energy efficient unit operations are not only driven by the functionality of the unit operations, but also by the operational conditions. Superstructure optimization is therefore a mixed-integer nonlinear problem (MINLP). Integer variables (0 or 1) are used for the selection of the unit operations while non-integer variables are used for the estimation of the operational conditions. Compared to non-linear problems (NLP), which have only non-integer decision variables, MINLP problems are more complex to solve. Several techniques are developed to solve these problems.

For solving MINLP and NLP problems a range of numerical solvers are available. For unconstrained problems gradient methods are successful in finding the optimum. Problems which are constrained by decision and system variables need advanced algorithms. Interior point methods, based on an iterative search within the feasible solution area, are most recommended. Other algorithms are based on sequential quadratic programming, branch and bound, trust region reflective, etc. Modern optimization software combine algorithms in one software function and can apply an automatic selection of the algorithms. Genetic algorithms, derived from the principles of natural selection, are another popular class of optimization algorithms which prove to be effective in computational time. All algorithms have a risk for ending in a local optimum and therefore a check by changing the starting points is always required. By defining the operational constraints and variables for all unit operations and the product related constraints the problem is defined and the optimization process can be performed.

For the Enthalpy project a set of 16 different processing routes is defined and analysed for both energy and water consumption. Where in case of energy the results are specified in primary and direct energy use and for water the obtained water that is extracted and can be reused is expressed. With these simulation model tool a clear analysis can be made been different concepts and the desired processing route can be determined.

The main aim of the ENTHALPY project is the reduction of energy and water consumption in milk powder production. Reduction of energy and water consumption, however, do not necessarily imply a reduction of the environmental impact. A lower energy consumption may require the use of larger equipment and higher amounts of consumables with a possible negative effect on the environmental impact. Therefore the evaluation of all proposed production scenarios is extended with a life cycle assessment (LCA) (deliverable D9.3).

In task 1.2 (Analysis of process configurations with innovative technologies derived from superstructure optimisation) milk powder production scenarios, including conventional and innovative technologies, were optimized for the lowest energy consumption. Moreover, the scenarios were evaluated on water consumption/production and their investment costs. In this task the production scenarios are evaluated and optimized with respect to a combination of environmental impact indicators, resulting in a minimal environmental impact.

The objective functions that are minimized were a combined score of 13 environmental impact categories (single score analysis) and the total annual production costs. The results are compared to the results of task 1.2 where the primary energy consumption was minimized. For each unit operation all the environmental impact categories were determined for the applied operational conditions.

The magnitude of the different impact categories is very different and also the importance of the impact categories varies. Hence a normalisation step is applied in the impact categories. By normalising the data the impact of each impact category is translated into a relative impact on national, regional or even global level. From literature derived normalization factors for Europe and the global system. The normalization factors refer to the reference situation of the extractions and emissions in the year 2000. Not all factors were covered by the literature reference; the missing factors were taken from another. As the scope is a European project, the European normalisation factors were used, completed with global normalization factors for the missing values. As a last step a weighting factor is added to determine the relative importance of each category

Examination of the combined score for environmental impact showed that process scenarios which include membrane distillation need significant cooling, resulting in a major impact by water depletion. However, the production of water in these scenarios compensate water depletion for a major part, and therefore water depletion is excluded from the single score.

In the last decades life cycle assessment (LCA) has become an increasingly important aspect or the design of products and processes. LCA has become also an important issue in policy making. Despite the ISO description of requirements for LCA, there is no common agreement on how to interpret some of those ISO requirements. As a result different approaches have been developed over the years. Nevertheless LCA is the most used method to assess a products or process’ environmental impact.
In current practice of process development, process design is based on product requirements in combination with cost, energy and resource efficiency. This results in process systems with a combination of unit operations and specifications for the operational conditions or replacing process equipment. LCA is performed after the design phase, and identified hot spots in the production chain might be improved by adjustments of the operational conditions or replacing process unit operations. This iterative work process is repeated until the decision maker is satisfied with the outputs. This iterative two-step method may lead to sub optimal design in terms of LCA if marginal improvements are applied instead of considering alternative processing options.
The four steps of LCA can be integrated in the design process in the following manner:
1. Combine the formulation of the production objectives with the objective and system definition, such that the system boundaries and objectives with respect to LCA and process design are formulated at the same time.
2. Define process functions, material, and energy flows and selection of unit operations with the inventory analysis.
3. Apply process simulation and impact assessment simultaneous and apply subsequently optimization for selection of unit operations and optimization of operational conditions.
4. Joint interpretation of process performance and LCA.

During the ENTHALPY project, pilot installations have been developed and these installations are experimentally evaluated. Task 1.5 concerns a projection of the obtained experimental results to production scale and a comparison of the new technologies with conventional processing.

The new technologies in the ENTHALPY project were enzymatic cleaning, radio frequency heating for pasteurization/sterilization, alternative energy supply by solar heating, membrane distillation for product concentration, dryer air dehumidification by membrane contactor, monodisperse droplet drying and inline monitoring of powder qualities. The aimed projection of the technologies concerned the benchmark production facility processing 10000 kg raw milk per hour into 900 kg skimmed milk powder per hour.

At the end of the ENTHALPY project experimental results were available for enzymatic cleaning, the solar heater, radio frequency heating, monodisperse droplet drying, membrane distillation and inline monitoring of powder qualities. The projection of the results from the experimental work to the benchmark production facility is reported in this deliverable.

The pilot experiments for the membrane contactor were hindered due to membrane leakage. The available data does therefore not cover the full range of operation that was aimed for the ENTHALPY benchmark production facility. The application of this technology in the evaluated range of operational conditions for the benchmark and an extrapolation to the aimed operational conditions was evaluated.

For all the technologies a comparison was made based on energy and water usage. And the resulting conclusions are:
- Enzymatic cleaning reduces energy consumption for cleaning with 60-70% and water consumption with 30-50%.
- For a medium scale factory with conventional equipment located in Spain, 1 ha of solar collectors is required, while for a factory equipped with the new technologies 0.38 ha satisfies. The solar heater must always be combined with an additional boiler for night operation, and for days and seasons with low radiation.
- Radio frequency heating proved to be a fast heating system, and reduces the residence time in heating equipment. As a result there is potential for lower degradation of product and a lower degree of fouling. The primary energy consumption is about 2.5 times higher compared to fuel heated boiler systems, due to electricity usage.
- The energy requirements derived from the pilot plant experiments indicate that membrane distillation does not yet compete with evaporation. The main limitations are the high amount of cooling energy required, and the low flux obtained in the pilot scale experiments. Certainly, there is potential for improving the technology to compete with the energy level of evaporators. Strong advantage of membrane distillation is that the systems can use ‘waste’ heat. For example from the air dehumidification unit.
- The tests on the use of monodisperse droplet drying were too limited for conclusions on the energy reduction. Mass and energy balance data of the Bodec drying towers obtained for standard products have been used to calibrate the dryer in the simulation tool. With this result the tool can be used to predict the performance for industrial open- and closed-loop dryers
- The experiments with the membrane contactor were not yet successful to draw conclusion about the energy recovery of this technology in closed-loop drying.
- The direct contribution of the inline monitoring systems (droplet size and moisture content) is small, but these systems are essential for a successful application of monodisperse droplet drying.

The most successful chain for milk powder production consists of the following unit operations: standardization, conventional pasteurization, reverse osmosis, multi stage evaporation, and monodisperse droplet drying in closed-loop with zeolite dehumidification. This system results in a 45% reduction of energy consumption for the considered factory compared to the benchmark.
WP2: Enzymatic cleaning technology
Fouling may be defined as the unwanted formation of thermally insulating materials or deposits from processing fluids on heat transfer surfaces. Selecting the correct cleaning strategy requires an understanding of fouling and the interrelationships between the surface, the food product and fouling. Some studies related to the fouling process during the transformation of milk have been published, but in general the way a milk facility is cleaned is based more on empirical practice than on scientific studies. The entire food processing plant must be cleaned in order to ensure its hygienic operation, but cleaning is not well-understood. Milk is a complex biological fluid whose several components include whey proteins, calcium and lipids. Under 100ºC thermal treatments in heat exchangers induce fouling on stainless steel surfaces that proceeds mainly from the denaturation of whey proteins.
Another important aspect is microbiological contamination, the formation of biofilms and the contamination of the end product by microorganisms that grow on the fouling when the milk is treated at relatively low temperatures. If the number of bacteria is high, the initial count in 3 logs units can be reduced by a pasteurisation process. A biofilm may include billions of microorganisms and this elevated number of bacteria may affect the microbiological contamination of the processed milk, reducing its expected shelf life or increasing the risk to consumer health. A correct cleaning process will result in minimal processing costs for food products compared with the potential economic losses and the risk to health associated with incorrect cleaning processes. There seems to be general agreement that thermal denaturation of the whey protein β-lactoglobulin plays a major role in the fouling process, certainly when the temperature is below 90ºC.
In order to improve the cleaning procedure the composition of the fouling and the effectiveness of enzymes on individual components of the fouling has been investigated in the previous tasks. In this task a cleaning formulation was developed and tested. A solution of alkaline water (W) (pH=8.5) alkaline surfactant and a mix of enzymes were used as an enzymatic formulation. To work with the solutions of enzymes, a stock solution concentrated 10-fold was prepared.
This enzymatic formulation was compared against two chemical treatments, namely an alkaline cleaning in a one-way treatment and an acidic cleaning treatment.
The enzyme activity is maintained for 12 months in the final product.
A new model for fouling attached to surfaces has been developed. We could then perform a direct analysis of a material, staining the organic components of the fouling and viewing their physical localisation. According to this new model and the staining of fouling residues, a specific enzymatic product has been obtained. This product has been evaluated at a laboratory scale and its effectiveness and durability demonstrated. Moreover, this new enzymatic cleaning product ensures better effectiveness and an acceptable durability, applied as a one-way treatment. It appears, therefore, that enzymatic treatment will mean a reduction in the time and energy consumed in the cleaning process.
The developed enzymatic cleaning formulation from the previous tasks was analysed in a pilot scale facility taking into account the energy and water consumption. The formulation of detergents with enzymes is very common for domestic purposes. However, in the food industry it is relatively new and specific analyses are needed to ensure the best effectiveness at a reasonable price.
Cleaning efficacy has not been very well-evaluated previously, as disinfection has been considered more important. Each country has specific regulations related to the use of disinfectants, but current cleaners have not aroused such interest. The EU started with the regulation of chemicals in 2007 and the publication of the REACH register (Registration, Evaluation, Authorisation and Restriction of Chemicals), the aim of which is to “ensure a high level of protection for human health and the environment, including the promotion of alternative test methods, as well as the free circulation of substances on the internal market and the enhancement of competitiveness and innovation”.
In general, the best methods to detect enzyme activity are those where a colour change is produced. Therefore, a standard reagent must be used for an easy procedure and the temperature, pH and time must be monitored. Colour intensity is directly proportional to the enzyme activity and the specific activity is quantified by means of a spectrophotometer and a specific wavelength.
Plate thermal interchanger
A Plate Thermal interchanger has been used for cleaning. The machinery was installed in the pilot plant of IRTA in Monells (Girona – Spain). The PTH was adjusted for pasteurization of the milk at 110ºC. The process volume was of 600 L / hour. Each assay included a pasteurization process during 2 hours. Then, an assay let a total consumption of 1.200 L. All assays were performed by duplicate. An alkaline, an acidic and the enzymatic cleaning process was performed.
Spray dryer cleaning process
Alter the drying process, the spray dried internal surfaces were swept away, to eliminate the dry residues unattached to the stainless-steel. Then, foaming products were used to remove organic residues. An alkaline, an acidic and the enzymatic cleaning process was performed.
For each of the cleaning processes and residue analysis was done.
WP3: Pre-treatment
Simulation of processes allows studying and improving processes faster than using experimental approaches and contributes to save money, because tests can be carried out on a computer and no operation of equipment is required. Simulation can be also useful for determining equivalent thermal treatment conditions (time, temperature) for different processing technologies.
In this task the mathematical models developed for RF simulation and for lethality efficacy simulation, and its implementation and validation in the experimental equipment are developed. In addition, a specific tool, created to calculate the solar thermal contribute to the savings of water and energy consumption is developed. The tool calculates energy balances of each process step using the registered data from the monitoring databases of the pilot plant. The results show the electrical and thermal energy consumption as well as the water consumption for a complete process
In RF heating the heat is generated as a result of the interaction between the milk and the electric field generated by setting different electrodes at two different electric potential. This process is modelled.
The model was validated by using data from several experiments. Three different heat treatments were considered for the validation: Low temperature pasteurization, high temperature pasteurization and sterilization. The following data was recorded for each experiment: inlet temperature, outlet temperature, flow rate and the applied voltage on the electrodes. This data was used for defining the boundary conditions of the model. For each treatment, the voltage in simulation (simulated voltage), which allowed obtaining the same temperature of the milk at the outlet, was determined. This was obtained by trial and error using the applied voltage in the experiment as the starting value. The voltage was increased/decreased depending on whether the resulting outlet temperature was higher or lower than the experimental value.
The main objective is to evaluate the energy and water consumption of the different heat treatments and to calculate the solar thermal contribution. For this purpose, a specific tool has been created. The tool calculates energy balances of each process step using the registered data from the monitoring databases of the pilot plant. The results show the electrical and thermal energy consumption as well as the water consumption for a complete process. Furthermore, the share of thermal and electrical energy consumption related to each process step is represented graphically. A crucial indicator of the tool is the solar fraction. The principal parameters of the buffer tank are represented in dynamic tables at one-minute time step.
The tool can analyse either an individual process or a continuous period. The last feature allows the study of the behaviour of the solar system during a period of time of several days in order to observe the solar contribution and the thermal inertia of the buffer tank.
Radio frequency heating is a dielectric heating process in which electric fields are applied to the product. Energy is transferred to the product in the form of electromagnetic energy which increases the product temperature. This process is governed by the dielectric constants of the product, which depend on the ability of the material to store electrical energy and dissipate it into heat. Once heat is generated inside the product, the speed of increase of the product temperature will depend on the heat capacity of the product. This parameter measures the amount of energy required to increase the product temperature by 1 ºC. These two properties must be determined for modelling and simulating the process.
For this reason, the dielectric constants and heat capacities of the different types of milk were determined in this task for RF heating. Several methods exist for measuring dielectric properties of materials. The open-ended coaxial probe method has been used extensively for measuring dielectric constants of foodstuffs. This technique allows to measure a broadband of frequencies and has minimum noise. For this reason, this technique has been selected for this work.
Dielectric measurements were performed with an open-ended coaxial-line probe (Model 85070E, High temperature probe, Agilent Technologies) connected to a network analyser (Agilent E5071C network analyser 300KHz-20GHz) with a coaxial cable (UT-250C-TPLL,Micro Coax) special for high temperature applications. The electronic calibration kit Agilent N4691-60006 300KHz-26,5GHz module was used to increase the accuracy of the measurements and to reduce signal noise.
A small autoclave has been built for measuring dielectric constants above 100 ºC (max. 140 ºC). In previous works, a pressure-proof dielectric test cell built by Wang et al. (2003) has been extensively used for analysing dielectric properties above 100 ºC for several foodstuffs. This autoclave has a capacity of 750 ml and has been built in stainless steel (316 L-S). It has three inlets. One for dielectric probe, one for a temperature probe for measuring liquid temperature and one for connecting a compressed air hose. The vessel could be pressurised at a constant pressure during the experiment. An oil bath heated by an electrical resistance was responsible for heating the liquid inside the vessel. The milk and the oil were separated by a wall through which heat was transferred. The electrical current applied to the resistance was controlled by a potentiometer which could be adjusted manually
Enzyme and acid coagulation are milk properties that are strongly affected by thermal treatments. By applying temperatures over 70 ºC, whey proteins denature, and aggregate on the micelle caseins surface, which modify the milk proteins for enzyme and acid coagulation. Enzyme coagulation properties are decreased if whey protein are attached to the micelles, but on the other hand, acid coagulation texture is increased in this conditions. Enzyme coagulation is a key step on cheese making, and acid coagulation is related to yogurt elaboration.
Coagulation curves were obtained by Optigraph (AMS Alliance), which detects changes in milk sample during clotting, due to changes in the caseins structure. Full fat homogenised milk samples were studied. Raw milk and milk heated by RF or by indirect interchange heating system, with treating temperatures from 70 ºC up to 152 ºC, were analysed.
There were no statistical differences for enzymatic coagulation between milk samples treated at the same temperature for Indirect heating or RF. Milk samples treated above 115 ºC could not be analysed due to lack of changes on the coagulation curve. These results suggest that RF and Indirect interchange heating have a similar effect on milk protein structure. Further studies may be done with non-homogenised milk and pilot plant cheese making proves to validate this conclusions.
Also textural analysis were performed. The results show a similar pattern to enzymatic coagulation properties of RF milk. Few differences were observed for the same conditions between RF and Indirect Heating. In any case, RF seemed to cause acid gels with slightly stronger gels as compared with the Indirect heating system. The enzymatic and acid milk coagulation results suggest that RF has a similar effect on milk protein structure than Indirect heating process.
In work package 7 the demonstration activities are described and in this task the results of the demonstration activities are correlated to the WP 3 activities to evaluate the pre-treatment technologies. The initial hypothesis is that RF heating causes less undesirables changes to milk than traditional UHT processing. This is the initial thesis of the RF machinery producer Cartilgliano, according to the data available in an internal study. It’s known that RF heating may result in a high uniform increase of temperature through the pipe flow, due to its wave characteristics. The uniform heating may result in an absence of cold spots, resulting in combination of high microbial inactivation and improved maintenance of nutrients. In order to validate this hypothesis, an experimental plan was designed, including both pasteurisation and sterilization conditions.
Raw cow’s milk from a nearby farm was collected and filtered. The content of fat and protein was analysed, and, if required, fat content was standardized to 3,5%. Samples for microbial and physical-chemical analysis were taken. Milk was always heated up to 60 ºC, before the homogenization process was applied. Then, using the Inoxpa prototype, the milk was heat treated at the selected temperature (by indirect heat exchange or by RF). After the heat treatment, milk was collected in aseptically conditions, in an aseptic and labelled bottles, using an aseptic sampling cabin. Milk samples were then kept in refrigeration before the different analysis were performed.
The samples were analysed on Microbrial analysis, Micronutrients, Physical characteristics, Thermal indicators, Effect of protein solubility.
All results of the comparative experiments between RF heating and Indirect Interchange Heating lead to the conclusion that both heating systems have similar effects on milk processing. Microbial reduction, either at low or high pasteurization, or at sterilization temperatures, is similar for both technologies. Moreover, effects on micronutrients, thermal indicators or protein solubility are similar, for a similar thermal profile, for RF or for Ind. Heating.
RF heating is a suitable heating technology for milk processing, although it is foreseen that its application may be more beneficial for high solid content liquid foods. Heat treatments for this kind of food is very complicated with Indirect Heating equipment, due to poor thermal food conductivity, and high fouling rates.
WP4: Atomiser
A set of models is used to describe the whole printing process to be able to predict droplets formation as a resultant of input parameters. A robust system design is necessary to make sure process stability is guaranteed during the process. Small variations in input variations should not lead to a large difference in droplet formation. The internal liquid flow and droplet break-up are modelled in code that was developed in-house. This has been coupled to an electromechanical model of the piezo system interaction with the liquid, so that the whole system is described from input to droplet formation.
The internal flow is modelled including the effect of the piezo movement.
The electromechanical behaviour of the piezo is modelled including the mechanical damping effects of the print head housing. This model is fitted on experimental data measurements performed with the actual print head and piezo element.
These two models are combined and used to simulated the break-up behaviour of the flow out of the nozzle. In order to validate this numerical model it is compared to two cases in literature.
This resulted in a complete break-up model that can be used to predict the behaviour of the Enthalpy print head.
In the previous tasks the stability and behaviour of the single nozzle inkjet print head was investigated. This behaviour and the flow through the print head was modelled in order to gain insight in the print head operation. Based on this experimental and modelling capabilities an improved design was made and this was up scaled to a multi-nozzle print head.
In the end, the new active multi nozzle print head should be as much a drop-in replacement as possible for the classic passive high pressure spray drying nozzles that are currently the default in dairy spray drying towers. The operational principle of the new print head however is quite different from the classic spray nozzle. In a classic spray nozzle, feed is pumped though a small nozzle. This results in a spray of droplets with a large size distribution. This spray nozzle is fully pressure driven. The new print head on the other hand is a flow driven system that will operate at a much lower pressure. The input flow is evenly distributed over an array of nozzles. At every nozzle this flow produces a fluid jet. The piezo actuation forces the break-up of the jet at a desired frequency leading to a desired droplet size.
This different way of providing the feed, means that in-stead of a pressure pump, feed needs to be provided by means of a flow pump. In the case of multiple print heads, care must be taken that the feed lines to all print heads are symmetrical, so that the flow is provided evenly to all print heads. Alternatively, each print head should be provided with its own flow controller, but that would rather increase the system cost.
After the homogenizer, a flow pump will be placed. This pump will be able to provide enough flow for the capacity of the drying tower. It will also be possible to run this pump at the relatively low flow of 20 L/h that is required to test the first, multi nozzle print head module. It can provide a pressure of up to 25 bar though normal operation is expected to operate at < 5 bar. The pump can be controlled by an analog voltage.
The operation of the new print head is in principle more sensitive to clogging. If one nozzle would get clogged, the other nozzles get a larger percentage of the flow. This seems redundant, but this also shifts the operational settings of the head. A higher flow per nozzle should be accompanied by a change in drive frequency to keep the same droplet diameter and make sure it stays in a stable droplet formation range. So to automatically correct the frequency, we would need to measure the droplet size from multiple nozzles as part of a control loop. Although we do plan to monitor droplet size in the development phase, in a real world implementation of the multi nozzle print head that would be too complicated and costly at the moment.
So it is very important to prevent clogging and make sure that all print heads remain in the same operational regime. During the single nozzle lab experiments, it was also seen, that the formation of coagulants is much more likely to happen with increasing solid content of the milk. It is difficult to find a good reference, but Figure 2 taken from [1] shows the state diagram of whole milk. Although not based on skimmed milk as will be used in Enthalpy, this research does give an indication of what may happen.
It is difficult to find a good reference for a state diagram of skimmed milk and therefore a state diagram of whole milk is used as reference. From this state diagram it can be learned that at 50% solid content and a temperature of 60 ºC, the formation of lactose crystals already starts. They can in turn act as nuclei for the formation of caseine coagulants. Therefore it was decided to run the plant at 45% solid content. At this level, single nozzle experiments in the lab with reconstituted skimmed milk have shown to run smoothly for over one hour without clogging.
In order to prevent clogging an extra filter is placed right before the entrance of the print head(s). This filter has a pore size of 5 µm which is a factor of 10 lower than the diameter of the nozzles.
A second issue observed from the state diagram is that the temperature of the feed is also important. In the current plant, the feed in the product tank is kept at 60 ºC. No measures however are implemented to ensure that the temperature does not drop below this value before it reaches the spray nozzle. For the classic nozzle there is no need for this because any possible coagulant is violently destroyed by the high pressure jet through the nozzle.
With the new multi nozzle head, measures have to be taken to make sure the temperature does not drop below 60 ºC along the feed pipes. Re-heating at the end, just before the print head is not an option, as it takes too much time to dissolve any lactose crystals that may have formed. In addition, the formation of caseine coagulants is quite an irreversible process so that has to be prevented.
For the design of the multi-nozzle print head the dead volumes inside the print head were minimised. The dead volumes are zones in the print head were feed material is present, but this feed is not flowing. This causes risk of microbial growth and have to be eliminated. Further a homogeneous pressure distribution to all nozzles has to be obtained to ensure that all nozzles produce the same droplet.
For the printhead to be able to exert enough energy on feed to create a disturbance in the jet coming out of the nozzle a set of piezo excitators is need that operate synchronise and exert the same force on all the nozzles. Therefore a specific excitatory stack is designed.
The multi-nozzle can be equipped with different nozzle plates varying in the amount of nozzle, rows of nozzles and the nozzle size. This will result in different droplet sizes and throughput. By stacking multiple print head together the throughput can be further increased.
In experiments with the multi-nozzle print head droplets were print into the spraydryer and dried to powders. These powders were analysed and result in a smaller powder size distribution and no fines are created.
In the first trials at the pilot facility one print head will be used for operation and if this is successful, three more print heads will be produced. This is the maximum number of heads that will fit in the inlet of the dryer in the pilot facility. That brings the maximum capacity of the tower with the new heads to about 80 liter wet feed in per hour.
All print heads have identical nozzle plates, so the same drive frequency can be used for all heads. This implies that there is no need for a calibration factor per print head. Every head will however get its own piezo amplifier since these do not come with multiple outputs and the output of a single amplifier is probably not powerful enough to drive all piezo’s of all heads. As piezo amplifier, a E-617.00F high-power piezo amplifier OEM module from PI will be used.
At first the flow signal will come from a manual input. Maybe later on, a specific output of the plant’s system controller could be used. A DC value between 0 and 5V represents the full flow range of the plant. It will probably need some scaling to match the input of the flow pump (signal adapter). A dedicated signal generator provides the required waveform, amplitude and frequency to drive the piezo amplifier. Waveform and amplitude will be tuned and fixed after the trials. So basically the signal generator will be a voltage controlled oscillator. For the trials, lab equipment will be used. After the trails simple but dedicated electronics will be designed and made.
WP5: Dryer modelling and inline monitoring
The main objectives of Task 5.3 defined in the Description of Work (DoW) for the present deliverable are the following:
Reduction of OVGU SDD model
The advanced SDD model developed by OVGU and described with details in D5.1 and D5.2 needs to be transformed into a simpler form which can be implemented into the CFD solver.
Development of CFD model of spray drying tower (air flow + particles)
Determination of boundary and initial conditions for CFD simulations. Development of CFD skimmed milk spray drying model which takes into account: behaviour of the continuous phase, particle drying, changes in particle diameters, and heat losses from the tower to the environment.
Test of the CFD model
The developed CFD model need to be used for simulations of skimmed milk spray drying under different operational conditions (air temperature and spray mass flow rate).
Implementation of kinetic models for quality into the CFD solver
Implementation of kinetic models developed by LBORO for protein degradation into the CFD solver and simulations in parallel with the drying model.
In order to build up these spray drying models five models were developed, starting from the simplest (water evaporation, mono disperse spray) to the most complex with full account of two stage particle drying.
Developed models can be divided into two groups: Standard evaporation models built into the Fluent solver, which allows only to simulate evaporation from one component liquid droplets, and models implemented by the user as an additional evaporation law for multicomponent droplets.
One of the most important parameters in product optimization is the PSD of received powder. Changes in droplet/particle diameters have significant influence not only on the drying process but also on particle trajectories. Knowledge about the changes in size would allow to better predict the behaviour of the discrete phase and to more reliably assess the influence of changes in dryer geometry.
The spatially distributed SDD model produces automatically the information concerning droplet/particle size (on physical grounds regarding droplet shrinkage, by empirical extension in case of particle inflation/deflation). The same holds for particle density or porosity. It is, therefore, important that such information does not get completely lost in the course of reduction to the CDC model, but can, at least partially, be retained or restored in the CFD environment.
To confirm its operability, the developed CFD spray drying model was tested for different process parameters. Each result was analysed to check if the model responds to changing conditions properly. Parameters used for the tests were chosen in correspondence to future experimental validation of the CFD spray drying model.
All tests prove that the CFD model for the spray drying process works properly. Couplings between the phases (mass, heat and momentum transfer) appear to be set correctly. However, to finally show model accuracy, results of CFD spray drying simulations need to be compared to experimental data, which has been done in the next tasks.
The main objectives of Task 5.4 defined in the Description of Work (DoW) for the present deliverable are the following:
Spray drying experiments
Experiments of skimmed milk spray drying were carried out in the co-current spray tower constructed in OVGU laboratory and described with details in D5.2. Product samples were analysed upon moisture content, particle morphology (measurements performed by OVGU) and quality loss (measurements performed by LBORO).
Validation of the CFD drying model for skimmed milk
The CFD spray drying model developed by OVGU, described with details in D5.3 needs to be validated on the base of data obtained from the spray drying experiments performed under different operational conditions (air temperature and spray mass flow rate).
Validation of the CFD quality loss model
The kinetic model of protein degradation during the spray drying of skimmed milk developed by LBORO and described with details in D5.2 needs to be validated on the base of data obtained from the spray drying experiments performed under different operational conditions.
Updates and tunings of the final CFD model
The complex CFD model needs to be updated according to the new sets of data obtained during spray drying experiments.
Measurements of drying kinetics during the operation of the spray dryer are difficult and complicated to perform. One of the issues for the measurements of particle moisture content and size distribution is the separation of falling particles from the drying air. This operation needs to be performed fast and carefully in order to not overheat the samples or mechanically change the particle size (agglomeration or breakage). Our first idea in regard of the powder sampling method was to suck the particles with air from the drying chamber and separate them by the outside cyclone. After a first set of experiments performed by the cyclone powder separation it has been observed that the moisture content of the collected powder was much lower than it was expected. In fact, only the smallest and dry particles were collected by the cyclone. Wet and heavy particles stuck inside the sampling pipe. This was confirmed by analysis of the powder PSD, which verified that only the smallest fractions of particles could be caught and transported to the cyclone container. Therefore it was decided to use a much simpler way of powder collection. Powder was caught by a metal container (350 x 70 mm) which was inserted into the drying chamber at different heights of the tower. The container has a metal cover which can be moved during the measurement. To decrease the possibility of sample overheating and measurement mistake powder was gathered for only 30 s. After this time the container cover was closed to protect the sample from being blown away from the container. Next the sample was collected and the container was cleaned and cooled down. To avoid any mechanical interaction with the powder, particles were removed from the container by pressurized air. This procedure was repeated until two samples with mass 3–5 g each were gathered. One sample was used for moisture content measurement performed by MA100C analyser (Sartorius, Germany). The second sample was analysed by the CamSizer XT (Retsch, Germany) to validate the particle size distribution model.
The changes in air temperature inside the drying tower for different atomisation rates and initial temperatures of drying air were modelled. The temperature pattern in the drying chamber shows a symmetrical temperature distribution. Two characteristic zones of changes in air temperature profiles inside the tower can be distinguished.
The first zone of air temperature changes can be observed in the atomisation area at the top of the column, where intensive moisture evaporation takes place. In the drying tower axis, where evaporation is most intense, we can observe a rapid air temperature drop. The air moving close to the chamber wall beside the spray envelope has higher temperature. The profile of air temperature changes inside the atomisation zone overlaps with the evaporation rate changes.
The second zone begins around 2 m below the nozzle; as a result of heat losses to the environment the temperature in the dryer axis is higher than the temperature near the dryer wall.
To verify the correctness of heat transfer calculations, air temperature at the level 0.85 m from the bottom of the tower was measured by a thermocouple. Because of small fluctuations of the air temperature for the comparison time averaged values were used.
Besides the temperature comparison also the powder properties and shape are analysed as an example the protein activity level was compared for both the CFD model as well as the experiments.
Optimization of existing spray drying installations is a difficult and expensive process. When designing new processes not only reduction of energy consumption is the objective. Also production efficiency, spray drying chamber size, stability of the drying process and quality of the produced powder must be taken into account in the optimization process. To avoid wrong design solutions that would lead to financial losses, all designer decisions should be supported by calculations and simulations to minimize the cost of proposed new technical solutions. CFD modelling seems to be a perfect tool to support the development of new spray drying chambers in this sense.
In this task the operation of spray drying chambers fitted by the multi-stream monodisperse atomiser developed by TNO was simulated by the previously developed CFD model described in D5.3 and D5.4.
In the first part a series of CFD simulations were performed for geometry of the spray dryer constructed in OVGU laboratory described in D5.2. In those CFD simulations the influence of different nozzle positions, two initial droplet diameters (180 μm and 167 μm) and the way of air introduction (vertical or swirling with 30° or 60° angle from the dryer axis) on the drying process were checked. Parameters like drying and particle residence time, wall deposition, inter-particle collisions, protein deactivation, air velocity and temperature profiles were compared for each case. Additionally, reference CFD simulations of the same dryer with a two-fluid nozzle were performed to show differences in spray drying process conducted by the new monodisperse and commonly used atomisers.
In the second part CFD simulations of a spray dryer installed at BODEC and fitted by the monodisperse atomiser were conducted to analyse the drying performance of the dryer with this innovative nozzle.
Series of CFD skimmed milk drying simulations of the OVGU spray dryer fitted by multi-stream monodisperse atomiser were performed.
Moreover simulations of spray drying in the BODRC dryer have been performed. Instead of a standard two-fluid nozzle, the monodisperse multi-head atomiser developed by TNO was used to produce skimmed milk droplets. In this simulation drying conditions were not scaled for the new atomiser capacity, which shows how sensitive the quality of powder is to high temperatures.
Simulation of BODEC and OVGU dryers shows that for the new monodisperse atomiser the best construction is a high and narrow co-current spray dryer. Axial air flow without recirculation prevents agglomeration, wall deposition and overheating of particles. Co-current spray drying allows for better control by changing flow the rate and temperature of the drying agent. The optimal diameter of the tower depends on the number of installed printing heads. However it is recommended that each printing head should be separated by an air inlet to increase drying efficiency and prevent collisions between the particle streams.
CFD simulations of skimmed milk spray drying conducted by monodisperse atomiser show big advantages over other drying processes performed by standard nozzles. Produced powder is expected to have high quality and uniform particle size which can be of crucial importance in pharmaceutical, cosmetic and food powders production.
Prototypes have been developed, constructed and made operational for the inline monitoring of particle size distribution and particle moisture content in spray dryers by OVGU and AVA, respectively.
The prototype for inline monitoring of particle size (and in the perspective also of particle shape) detects optically the projection areas of particles moving between a light source and a camera. The light source shall be placed at the wall of the spray dryer, with the camera inserted to an adjustable extent from the opposing wall into the dryer and the, also adjustable, position of the focal plane (observation plane) in between. This corresponds to a transmitted light method in cross-section setting, anticipating that optimal spray dryers with printer atomisers will be of different geometry than conventional spray dryers (relatively narrow and long tubes with concurrent flow of air and droplets/particles, instead of the wide and short conventional vessels with large recirculation zones).
Development of the particle size monitoring prototype included the following steps:
- Rejection of the option to adapt commercially available sensors with one-dimensional arrangement of light beams;
- Definition of challenges by spray dryer conditions (low concentration, relatively small size and relatively high velocity of droplets/particles in the dryer, high operating temperature) and their translation into measurement requirements;
- Component selection, hardware testing, selection/development of control software and tools for image analysis and statistics;
- Successful offline testing by means of a falling curtain of particles, using as reference standard equipment for offline particle size distribution measurement (trade name: Camsizer).
The prototype for inline monitoring of particle moisture content has been developed on the basis of a sensor offered commercially by AVA under the trade name SAMPIN for inline application in fluidized bed dryers and granulators. The successful operating principle of SAMPIN (repeated cycles of filling a sample holder with solids, capacitive moisture content determination of the sample, and emptying of the sample collector) has been pertained, but the device has been essentially modified to account for spray dryer conditions in respect to the following:
- Cooling of the sensor adapter by refrigerant agent to allow operation in high-temperature environment;
- Decrease of the measurement volume (sample volume) by reconstruction of the sensor head to cope with low particle concentrations;
- Implementation of a new method for discharging the sample after measurement from the measurement volume (by pivot drive movement, i.e. rotation, of the sensor head, instead of the use of pressurized air for blowing particles out of a rigidly mounted sensor head).
All new components and operating features have been tested offline, separately as well as in their combination in the fully assembled new sensor.
Inline testing of both monitoring devices in the new spray dryer at OVGU has been conducted. Experiments with the DIA confirmed the required thermal durability under high gas temperature conditions (200 °C). It was shown that adverse effects by fouling of the probe window can be avoided either by hardware (installation of a metal shield) or by software (development of an adaptive binarisation algorithm). The adaptive binarisation algorithm is applied locally and can also accommodate for the random movement of particles in the gas flow and differing distances from the optical focal plane. In total, DIA inline particle size results were found to agree well with offline reference measurements, provided an appropriate and unbiased sample collection for the latter. Particle shape can also be evaluated by the DIA, so that the appearance of undesired agglomeration can be detected; (large agglomerate particles are less spherical that small primary particles resulting from single droplet drying). As a conclusion, the provided results fulfil the deliverable objectives of particle size and shape inline measurement.
Concerning particle moisture, an existing capacitive instrument has been redesigned and made operational for spray drying conditions. For this purpose a cooling jacket has been designed, which enables the sensor to operate at high gas temperatures (200 °C). The emptying mechanism of the sample holder been changed into a turning mode to prevent compressed air (which is loaded with oil) from reaching the process and polluting the milk powder. The measurement volume of the sample holder has been scaled down to achieve a faster filling for the small particle concentrations which appear in spray dryers. Trials which compared data from the inline moisture measurement to the reference method (offline drying oven) have shown inconsistent behaviour for the raw material, but a good agreement if the milk powder is compacted in the sample holder (attributed to the avoidance of variations in bulk density).
WP6: Membrane technology
In WP6 innovative routes are developed for the concentration and drying process. These aim at recuperation of at least 45-85% of the total heat which is consumed by the concentration/drying process. The steam-consuming multi-effect evaporators will be fully replaced by a combined Reverse Osmosis (RO) and Membrane Distillation (MD) process. In this case the heat for driving the MD process is taken from a membrane contactor (MC) process. It should be noted that the high energy efficiency is obtained by enabling closed-loop drying.
The membrane contactor (MC) to be developed will combine several functions in one device:
- Removal of water vapor from the gas by absorption in a hygroscopic salt solution (brine) at high temperature;
- Heat exchange between the brine and the gas stream;
- Complete physical separation between brine and gas, ensuring zero contamination risk of produced powders; and
- Ideal counter-current flow conditions, enabling a good performance at relatively low salt concentrations and therewith mild regeneration conditions.
The first series of experiments were carried out on lab scale with an MC, which had two channels, one for air and one for the brine. At the same time calculations on the heat integration in the total process (described in D6.1) showed the benefit of a third water channel. Four bench scale MC modules were built with three channels.
First small scale lab experiments with a Membrane Contactor test unit are executed at TNO. This unit was designed according to the original plan, in which the MC has two compartments (air and brine).
The FP1 membrane was used with an area of 0.075 m2. The module consists of one air and one brine channel with a flow path length of 500 mm. The brine was recirculated over the module. The weight increase of the brine was measured using a balance and the temperature at the inlet and outlet was monitored. The air passed through the module once and at the inlet steam was added to increase the moisture content. Online measured were the temperature and RH of the inlet and outlet air.
Four tests were done in which the experimental conditions were adjusted to increase the water flux. In the last test the air flow rate was 7.8 m3/h and the brine flow rate 17.4 kg/h. During the experiment the brine concentration was increased till 57% DW at the end of the experiment.
The experiments on lab scale showed that the principle worked; the air was dried and the LiBr adsorbed water. A high %DW of the brine was required to obtain enough driving force. A flux of 3 kg/m2.h was reached. The heat loss in the lab scale setup was high. It was decided that scale up and improvement of the experimental setup were required
MC with three channel
Calculations on the heat integration in the total process (described in D6.1) showed the benefit of a third water channel. The air is dried, the moisture is adsorbed by the brine and all latent heat is transferred to the water. Ideally the air and the brine stay constant in temperature. The air and brine are operated counter current, the brine and the water are operated counter current.
Four different modules were made and tested. In the MC modules preferred flow occurred. By placing the modules horizontally and introducing a meandering flow pattern in the brine channel, the mass transfer coefficient could be improved by a factor 9 (from 0.002 to 0.018 m/s). This is still a factor 2 lower than the theoretical value (0.035 m/s). The highest average flux reached was 1.5 kg/m2.h.
Also on bench scale the heat losses to the environment were high. In practice these losses will be much lower since the modules will be stacked. The internal losses were lower and in practice will be even lower since the MC will be operated with a brine and air flow of the same temperature (80ºC).
The second part of the closed loop for membrane technology is the Membrane distillation technology. For this technology the experimental work was split on the following topics:
- In the first series three MD membranes were tested to select the most suitable MD membrane for the concentration of skim milk;
- In the second series the selected Fluorinated polymer (FP1) membrane was tested on lab scale for the concentration of skimmed milk;
- In the third series the FP1 membrane was tested on lab scale to monitor how fouling affects the performance;
- In the fourth series of experiments the setup was scaled up to bench scale testing modules. Pasteurization tests and skimmed milk concentration tests were carried out.
Three flat-sheet membranes were tested on lab scale. The membranes were tested in a MD system set up. Membrane area was 0.05 m2. The module was operated co-current. Temperature and conductivity were measured at the in- and outlet flows of the module. Transmembrane flux was continuously measured by recording the weight gain of the distillate, which was re-circulated in a closed system. Duration of the experiments was between 30 and 50 h.
The distillate conductivity of the PVDF and FP1 membrane were comparable with an average of 25 µS/cm. The conductivity of the PTFE membrane distillate showed the same behaviour for the first 5 hours, but thereafter increased linear with time, indicating micro leaking of the membrane. Proteins have surface active properties and could be the case of the wetting of the membrane. The highest distillate flux (14 L/m2.h) was found for the FP1 membrane. The fluxes for the PVDF and PTFE membranes were lower and decreased more over time, indicating possible membrane fouling.
The same MD setup was used for the lab scale concentrate of skimmed milk as for the membrane selection tests. The experiment started with tap water for 2 h, during which the clean water flux was measured. The next day skim milk was concentrated for approximately 8 h and cleaned with NaOH (1.5 h). The next morning the clean water flux was measured (1.5 h) before the skim milk was further concentrated for 7 h. At the end of the day the membrane was cleaned with NaOH (1 h). On the fourth and last day the skimmed milk was further concentrated and at the end the membrane was cleaned with NaOH.
Throughout the experiment the distillate flux, conductivity in the feed and distillate, temperatures at the in- and outlets were measured. The temperature of the skimmed milk (input in the membrane channel) was maintained at 60ºC and the temperature of the distillate (condenser channel) at 12ºC throughout the experiment. The distillate flow rate during the three skimmed milk concentration runs was 3.0 2.0 and 3.5 L/min respectively. The skimmed milk flow rate was not constant during the experiment but decreased with increasing TS. At the beginning of the experiment the flow rate was 3.0 L/min (0.25 m/s), at the end of the third day 1.5 L/min (0.12 m/s).
The energy consumption is the energy use of the heat exchanger at the top (between condenser and membrane). The energy consumption is calculated multiplying flow, specific heat capacity of milk and temperature difference over the heat exchanger. It is expressed in MJ per m3 distillate.
The milk was concentrated to 31% DW. From the experiments it seems that in the first section the milk was concentrated to 28% DW, in the second section the concentration was limited and in the third section no further concentration was realized.
After the selection of the membrane and the first skim milk concentration tests, work was done to investigate how membrane fouling affects the performance. Experiments were done at different temperatures, flow rates and concentration.
Experimental results show that concentrations of 50% total solids can be achieved for skimmed milk, which makes MD a competitor for evaporation. For skimmed milk with 25% DW at 50°C, an initial flux of 12 L/m2.h was achieved. Due to gradual fouling the flux declined to 6 L/m2.h after 13 h. At concentration, the flux declined from 12 to 3 L/m2.h. Increasing velocity and temperature of the skimmed milk had a positive effect on the distillate flux, with the best results obtained at the highest tested velocity (0.3 m/s). The concentration has a larger effect on the flux than the operation time. It should be noted that these results were obtained for high temperature differences (~45ºC).
The next step in the development of the MD process was testing on bench scale. For this purpose a bench scale unit for testing MD modules was built. To work under sanitary conditions and avoiding the use of sodium azide, periodical pasteurization of the unit was studied prior to the skimmed milk concentration experiments.
Skimmed milk concentration was done with various MD modules. In the cooled tank the temperature is below 10°C. The modules have 2.6-5.8 m2 membrane surface and are constructed with food-tolerant materials. Water removal capacity is around 10-30 L/h. Spiral wound membranes are used and the channels with the streams with the highest temperature are located at the inside of the module to ensure minimum heat loss. The unit can be operated batch and semi continuous. The two main streams in the condenser and the membrane flow counter-current. The distillate is in counter current with the condenser stream and current with the membrane stream. The distillate is recovered at the cold side of the module
Form the experiments it was proven that the application of periodical pasteurization is technically possible. However, it still has to be proven on pilot/demo scale. This would overcome a major hurdle towards large scale MD for milk concentration. The clean water flux was restored for 89-90%. It seemed that after 70 h of operation, cleaning with water was sufficient.
The results presented in D6.1 showed that the total process (RO + MD + MC-assisted closed loop dryer) can be capable of achieving 49% reduction of heat (steam) consumption compared to the benchmark, besides a full recovery of water. Power consumption will be comparable, or slightly less than that of the benchmark. In D6.1 an extensive process model based on energy and mass balances was made. For re-evaluation of the benchmark and scenarios based on the experimental results, it was decided to work in this task with a more comprehensive and concise model.
The physical process of both the Membrane distillation as well as the membrane contactor have been modelled. The models were used to determine the mass and energy balances for the two membrane technologies.
Overall the energy reduction of the envisaged RO + MD + MC-assisted closed loop dryer is estimated at 20%, which is much lower than the expectation at the start of the project (50%). This is caused by the lower performance ratio and the required cooling energy of the MD, which was previously not included. To improve the performance ratio, the mass and heat transfer in the MD should be improved. A technical solution must be found for the high cooling energy of the MD, which could be found in the operating conditions of the MD and alternative cooling systems.
WP7: Demonstration at industrial scale pilot facilities
Enzymatic cleaning
For the enzymatic cleaning no specific adjustments had to be made to the pilot facility.
Indirect heating system
For both the RF-heating as well as the solar thermal system an modification to the indirect heating system was required. The most important modification and improvements are:
- The primary design of direct/indirect heating system (DIHS) was made with a flow divers panel that allowed for choosing different treatment combinations.
- Retention time of direct, indirect and RF treatment was adjusted
o The retention times in the system were too long and therefore adjustments were made to reduce length of the system, pipe diameters were adjusted and sections were eliminated
- Balance tank improvements
- Sterile condensate to homogenizer
o Additional condensers installed to reach the required flow rate
o Additional flowmeter to control output of the extraction hood
o Differential pressure sensors, for correct and balanced elimination of steam in the extraction hood
o Adjustment of the vacuum pump, to prevent milk flow towards the vacuum pump.
Radio Frequency system
The developed Radio frequency prototype system was installed at the pilot facility in Spain. Before being able to operate it some improvements and adjustments had to be made.
- Working pressure inside RF applicator
o During operation it is important that the flow is pressurized inside the system to prevent evaporation. Evaporation will increase the pressure and can cause an explosion and breaking of the tubes.
- The pressure regulating valve of the DIHS and RF is positioned before the laminar flow bench in order prevent pressure drop when collecting an sample
Solar thermal system
The solar thermal system was installed on the roof of the pilot facility. The main adjusted to have been done are:
- The solar thermal field was position on the roof optimally with respect to the buffer tank and beside the solar tube that are in the roof to provided sunlight to the facility on the inside.
- The pipe diameters of the system were altered to improve the hot and cold overheated water flow between the collectors and the buffer tank.
- All pipes and connectors were made leak free
- A proportional control valve was installed for regulation of the tertiary circuit of the solar thermal system
- A heat exchange cascade control system was installed to optimising the operation of the different heat exchangers
Monodisperse atomiser
The original design for the adjustment of the pilot facility in the Netherlands also include the membrane technology and the inline monitoring. Only the mono-disperse atomising has been partially installed in the pilot facility.
Benchmark trial were run in the pilot facility with skimmed milk to compare to the final validation experiments with the mono-disperse atomiser on energy and mass flows and powder properties.
Due to the limited time within the project it was not possible to extensively test the print stability and direction of the multi-nozzle print head. Therefore the firm belief that despite our best efforts safety could not be guaranteed when droplets would not enter the tower but instead would remain in the air distributor in the top section. The risk of fouling and slow smouldering or burning in the air distributor was too high and could not easily be solved. The preferred routing was to engineer and develop a completely new tower air insert, in which more space could be available for positioning of the nozzle system and which could create the opportunity for cooling the multi-nozzle print head system. This route would provide droplets entering the tower and also drying of the droplets while still in mono-disperse state.
To be able to safely demonstrate the nozzle an alternative trial plan was developed. A smaller spraydryer had become available in the pilot facility with lower water evaporation capacity but with easier access into the feed inlet. Again a new insert was developed and engineered to be able to introduce the nozzle into the existing SD setup.
The feed section of the Anhydro SD has a much simpler setup and can easily be replaced. The atomizer section (wheel and rotor/motor) could simply be removed from the tower and a new insert was created. The new insert was developed to incorporate the nozzle.
The insert was designed to mimick the exact space that normally also the wheel atomizer takes up. This creates no influence or differences in the airflow pattern inside the tower. No other modifications were required for testing the nozzle as a retro-fit inside the Bodec small SD. With this new setup drying trials were performed.
Membrane contactor
For the membrane contactor unit an investigation was made on the possible implementation in the pilot facility. The amount of original outlet airflow was too large and therefore a bypass pass considered with an extra support fan to compensate for any effect on the stable operating pressure of the spray dyer introduced by the membrane system.
- A tie-in to the original main system was foreseen including an extra tube to recycle the dehumidified air back into the outlet tubing.
- A H2O outlet was taken into consideration
- Fan for covering the pressure drop
- Hot water supply is required
- 13 kW Cooling capacity for Brine regeneration
Membrane distallation
The membrane distillation unit (for concentrating the skimmed milk) can be positioned as a standalone unit. No tie-inns were required since the product can easily be transferred to the SD feed tank. Pre-concentration of skimmed milk can either be performed via RO membranes or via evaporation. Tanks and utilities are available as well as a three stage evaporator for benchmarking.
Capacity of the add on system to the spray-drying operation as well as limitations (flow range, temperature, relative humidity), the effects on product quality and possible safety issues of the final setup were discussed and addressed between the involved parties. The membrane distillation setup would require a temperature source which helps operate it (currently a brine solution is suggested which enables the re-use of heat from the membrane contactor).
Plant modifications The membrane distillation unit can be operated as a stand-alone unit. The only required modfications to the setup are the hose-connections:
- from the inlet tank containing skimmed milk towards the membrane distillation unit
- From the membrane distillation unit towards the spraydryer feed-tank (from which the concentrated milk will be used as feed for the nozzle systems)
Utility specifics for electricals, steam and cooling water need to be provided
Integration at pilot facility in Netherlands
The foreseen layout in the pilot facility is developed. Within the Spanish pilot facility extensive testing and validation has been performed on the enzymatic cleaning. These experiments included the validation of open and closed surfaces and different types of fouling. The performance was analysed by comparing the cleaning with enzymatic formulation towards a two chemical cleaning protocols. The results were analysed and this is described in the deliverable of wp2 and the task progress of task 2.5 in this periodic report.
The radio frequency heating was validated in the Spanish pilot facility. Here different treatment combinations of pasteurization, sterilisation and heating were compared at different operating settings. The results were compared on the following parameters: Microbial analysis, Micronutrients, Physical characteristics, Sensory analysis.
The solar thermal system was made operational and testing during multiple months to see the effectiveness of the system in providing heat and energy for the pre-treatment processes.
As explained in task 7.2 the inline monitoring was transferred to the OVGU pilot facility and the membrane technologies were performed on bench scale testing.
Based on the runs performed in the pilot facility an energy picture for the process has be calculated in which both the conventional technology as well as the new technologies will be compared in order to assess possible improvements: Powder properties, Energy savings –if these can already be calculated-, Opportunities for re-using energy in the system
Prior to the trials on nozzle reference cases were determined for making the comparison on process and powder quality once the modifications were tested and could be included in the pilot plant setup. All mass and energy balances for skimmed milk drying were determined for reference. These results were also delivered as input to WP1 (process system engineering), WP5 (modelling of the dryer) and WP9 (LCA).
Referece trials were run on both the large spray dryer as well as on the later involved smaller spray dryer. For the trial runs were made with water, maltodextrin and skimmed milk, in order to create increasing complexity.
The multi-nozzle print head was mounted inside the small spray dryer and operated with water and maltodextrin. The tests were used to analyse the fouling inside the spray dryer and the powder was collected from the spray dryer and analysed.
WP8: Development of a food quality and safety concept for novel products
Within ENTHALPY project, a comprehensive food quality and safety concept based on the principles of HACCP has been developed in order to assure the food quality and safety of the products derived from the processes established.
The most important achievements for task 8.1 could be considered the following:
• Identification and evaluation of all potential hazards associated with ENTHALPY’s Skimmed Milk Powder production.
• Detailed description of the individual process steps for the production of ENTHALPY’s Skimmed Milk Powder
• Identification of critical control points
• Identification of control measures
• Establishment of an HACCP plan for the production of ENTHALPY’s Skimmed Milk Powder

Within the periodic report of period one, the development of a quality assurance programme was already reported. Thus, the second 18 months period included the programme application in order to evaluate the products obtained within the project such as the flavoured milk drink formulations (spray dried / agglomerated) developed within Task 8.5.
Within the analysis, protein denaturation, microstructure, particle size, sensorial and microbial properties were tested. Hence, protein denaturation of (i) spray dried and (ii) fluid bed agglomerated flavoured skimmed milk powder, clearly showed that heat treatment promotes denaturation. The magnitude of which depends on the temperature/moisture-time trajectory of the product during processing. Of the samples examined, BZN has the least concentration of each of the three protein components. The responses of each of the protein components to the same treatment are evidently not the same. However, factors such as added sugars within the recipe, varying HPLC sample preparation times and slightly-varying added HCl volumes could also have influenced the results. Besides, microstructure analysis using X-ray MCT revealed that the structure of the spray dried flavoured skimmed milk is similar to the control whereas other components such as sugars may cause differences in the overall structure. Using the agglomeration process, the cavities of the spray dried skimmed milk become thicker and thus more stable against mechanical treatment. The agglomeration became much more visible by analysing the surface morphology using a microscope. Finally the particle size revealed that the size of the agglomerated powder is almost four times bigger than the spray dried flavoured skimmed milk powder as well as the control.
The overall conclusion derived from the sensorial analysis is that both samples were accepted by the potential consumers. However, more effort and investigation is needed in both cases in order to improve the touch of the final product. In terms of flavour and taste, the strawberry flavoured sample showed the greatest performance with the non-flavoured also scoring high in criteria values. Finally, the strong advantage of non-flavoured skimmed powder milk against to the flavoured one is the appearance characteristics with most important the colour acceptance, transparency, visible purity and homogeneity. The flavoured sample appeared a white-yellowish hue and also a phase separation in the surface of the sample, both factors are deterrent to the potential consumers.
Microbial analyses were successfully performed, all samples being within the permitted EU food legislation limits. All detailed methods and results are summarised within deliverable 8.3 (Report on quality assurance analysis)
In order to conduct shelf life tests and meeting the requirements defined in Task 8.2 accelerated protocols were successfully established and applied for the flavoured milk drink products in combination with sensorial and microbial analysis.

Considering that a key criterion for the success of the powder based products is their shelf life and storage stability, within task 8.4 microbiological changes during storage of skimmed milk has been effectively examined. A particular focus was set on yeasts, moulds and aerobic or anaerobic germs. Moreover, specific parameter that could influence the shelf life of the products such as temperature and humidity were also considered. For that purpose and, as powders commonly possess a long shelf-life, accelerated shelf-life test protocols have been successfully applied.
All detailed methods and results are summarised within deliverable 8.4 (Report on shelf life tests performed).

During Task 8.5 the main focus was put on developing microencapsulation procedures for the encapsulation flavours as well as protective native milk ingredients.

Task 8.5 was initiated by developing the strategy to be followed in order to develop microencapsulation processes for the encapsulation of flavours as well as protective native milk ingredients as shown in the figure below.

As starting point, a suitable recipe for the ENTHALPY flavoured milk drink has been successfully developed by BZN. For this purpose, different ingredients / concentrations were tested and appropriate flavour formulations were identified.
Further, two different processes and their related protocols have been effectively established and evaluated:
I. Process for the agglomeration of skimmed milk powder with the selected flavour formulations using fluid bed dryer and
II. Process for the encapsulation of skimmed milk powder with the selected flavours formulations using spray drying.
Thereby, while the recipe development was performed, multiple tests using fluid bed dryer (lab scale, Glatt GPCG-3) for encapsulation of skim milk and flavour were carried out. As results, all necessary parameters such as appropriate agglomeration temperature (product temperature), time and speed of spraying of the main powder component (skim milk) were successfully determined. Further the preliminary recipe was used to verify the process but also to optimise the recipe mainly in terms of flavour content.
In parallel, protocols for spray drying flavours as well as native milk ingredients have been evaluated and successfully established.
Overall, the results of both processes validation revealed that for the final development of the ENTHALPY milk drink, the fluidized bed process developed represented the most suitable option for the microencapsulation / agglomeration of the ingredients.
Beside this, during the processes development, multiple analyses (e.g. particle size, bulk density, etc.) have been performed, results being described in Deliverable 8.3 “Report on quality assurance analysis”.
WP9: Life cycle analysis
Inventory analysis is a task that already started in period 1 by gathering all the LCA data of the conventional dairy chain. In the second period the remaining activities in this task involved the data concerning the innovative technologies developed within the ENTHALPY project.
The inventory analysis is the LCA phase that involves the compilation and qualitative - quantitative identification of inputs and outputs for a given product system throughout its life cycle or for a single process. This implies an inventory of all materials and energy inputs and all emissions to air, water and soil. An inventory is a listing of how much energy or material is used in each process during a product's manufacture and how much solid, liquid or gaseous waste is generated during its manufacture, use and eventual disposal. The inventory analysis includes iterative data collection and the compilation of the data in a Life Cycle Inventory (LCI) table.
Data collection is the basis of the Inventory Analysis. The next step after data collection is the calculation of the inventory results with the use of the collected data. The inputs for all the operations are used to calculate the mass balance linking all the subsystems in the system and estimate the outputs of each subsystem and of the overall system as well.
In general, data collection and manipulation included:
• directly measured data by ENTHALPY consortium partners, through completion of data sheet questionnaires;
• data from simulation tools of parallel project actions in WP1, which forms a credible model of industrial situation;
• specific data for milk industry from Food database of Gabi 6;
• literature data;
• calculations based on specific formulas taken from literature and from simulation of WP1–this procedure mostly refers to data conversion;
LCI analysis has been performed for the processes involved in the SMP production value chain and aggregated results were reported for: Transport for Milk (Farm to Processing Plant), Bulk Storage/Mixing, Separation, Pasteurization, Reverse Osmosis, Pre-heater, Evaporator, Spray Dryer, Fluidized Bed Dryer, Cyclone/Bag Filter, Cleaning In Place (CIP), Wastewater Treatment, Pasteurization Radio Frequency Heating (RF), Membrane Distillation (after RO), Monodisperse Dryer with Membrane Contractor, Monodisperse Dryer with Zeolite
Overall, LCI analysis of this study is restricted only to reporting of results and not perform comparison among them. This will be performed in a later stage, where data from innovative technologies will be finalized and derived from the demonstration plant.
The Task of LCIA has also already started in period 1 and is continued in period 2. The Life Cycle Impact Assessment (LCIA) identifies and evaluates the amount and significance of the potential environmental impacts arising from the LCI. Inputs and outputs are assigned to impact categories and their potential impacts quantified according to characterization factors. An essential step is the selection of so-called impact categories. Some of the impact categories like global warming and ozone depletion have a global impact, while other impact categories such as acidification, eutrophication and human toxicity have regional and local impacts. The choice of the impact categories is based on the recommendations of the Product Environmental Footprint and on the scope of this study.
In the LCA analysis a comparative analysis is performed of the process chain with the innovative Enthalpy technologies compared to the conventional processing chain.
Beside the environmental analysis also the economic analysis has been performed with the same framework. Even though an optimal route is selected in energy, water, economic and environmental sound terms it makes sense to the industry to evaluate the absorption of the innovative technologies fragment by fragment. Thus, the methodology for this is based on a fragmental change (one innovative technology implementation each time) in the conventional dairy value chain and the optimal tool selected to evaluate the risks (internal and external) is the SWOT analysis. In a SWOT analysis all the factors that pose potential risks to the introduction of each innovative technology to the conventional practices are examined.
The results of the SWOT analysis can be handled by the interested parties as a manual with all the data in hand on what it takes to bring the innovative technologies to the market. In some cases, the introduction of more than one innovative technologies seems to be beneficial for the whole value chain. Thus, the interested stakeholders could have access to the SWOT analysis and the respective data from the technology developers and examine the upscale and introduction of the technologies to the market.

Potential Impact:
Dissemination and Communication tools and activities
Within the lifetime of ENTHALPY various communication and dissemination activities promoted the project and facilitated the transfer of project foreground to different target audiences, to aid exploitation. The full list of all activities is provided in chapter 2 (Tables A1 + A2), the following description highlights only the most relevant activities in this regard.

New media, social media and interactive tools
The project website (www.enthalpy-fp7.eu/) was the central hub of project information. Besides the news items, and an event list it contains links to the social media accounts, information about the consortium, scientific methodology, but also a media corner to make available other material in a digital form. An RSS feed enables users to be actively updated on website changes. The website was constantly updated and adjusted to reflect progress in the project, for example with an additional menu button for the final conference.

Social media channels: LinkedIn and Twitter were filled with content on a regular basis, providing information about upcoming events, public deliverables and other news. While Twitter was little followed, more than 100 people connected to LinkedIn.

ENTHALPY whiteboard animation (https://www.youtube.com/watch?v=GDz5gLzs2i4): To reach the general public a novel approach was designed in the form of a whiteboard animation. This short video demonstrates the production of milk powder and displays conventional processes in parallel to innovative methods developed in ENTHALPY and the according savings in energy consumption. So far it received 350 views from all over the world on YouTube. The project YouTube channel also contains a playlist with demos for the ENTHALPY simulation tool GUI developed by Wageningen University and video by dissemination partner SUSMILK (https://www.youtube.com/playlist?list=PLmUlVJcIQkiq0g4pipZ7jzaTdaXMG8tRL)

Print material and publications
A project flyer and scientific poster were designed, both for print and also as online versions. For the final event and participation with a booth at the EFFoST 2016 also a roll-up was designed. To reflect the progress of the project, but also to promote the final event with stakeholders a newsletter was written. All work packages presented first results and raised interest for the ENTHALPY final conference. The newsletter was spread via project channels, is available on the website, but was also sent directly by partners to stakeholders and also to EnReMilk and SUSMILK.

Publications: Academic partners published their research in several scientific, peer reviewed journals and master and PhD thesis. OVGU published two papers in Drying Technology, and a third one is currently under revision and three more are being prepared. In addition one PhD thesis and one master thesis focus on results from ENTHALPY. Wageningen University published their results in Trends in Food Science and Technology and two master theses were finished. LBORO submitted a paper to the International Diary Journal and is preparing two more publications. Student from both UAB and IRTA are preparing their PhD theses. There were also several papers in conference proceedings.
ENTHALPY Executive Summary: In an effort to support post-project exploitation of project results this summary of commercially relevant results was prepared. This high quality brochure sums up the main achievements of the project, with a special focus on commercially exploitable results. The document is not only being distributed as a print version but is also available for download on the project website.

To announce the launch of the project, two press releases were produced: a general press release and a technical press release. At the end of the project a final press release was prepared, based on the successful ENTHALPY final conference in Monells, Spain, and project results from all work packages.

Events and conferences:
ENTHALPY final conference, 26.10.2016 Monells, Spain, at IRTA facilities. To maximise the impact of ENTHALPY results and foster post-project uptake of commercially relevant results, a final demonstration event was organized. The location the IRTA facilities in Spain provided the opportunity to combine presentations with a tour of the live-working demonstration facilities. In addition to several speakers from the ENTHALPY consortium, both our sister FP7 projects EnReMilk and SUSMILK were represented by speakers. The program attracted 55 participants from all over Europe. A local TV station conducted interviews and reported about the event, as did several regional newspapers.

The 30th EFFoST 2016, Vienna, Austria, 28-30.11.2016. ENTHALPY was present with a booth to promote the final results and distribute the ENTHALPY Executive Summary. In addition to RTDS, UAB and TNO were also participating in this activity. Wageningen University prepared a handout and a video on their simulation tool and TNO exhibited the monodisperse atomizer.

Highlights of participation in conferences: Consortium members presented ENTHALPY and their work within the project at a variety of scientific conferences, fairs and events during Period 2. The most relevant ones were the following: oral presentations at the 1st Nordic Balitc Drying Conference, Gdansk, Poland, 17-19.06.2015 (OVGU). Oral presentations and publications in conference proceedings from OVGU and WU at EFFoST 2015, Athens, Greece, 10-12.11.2015. IDS 2016, Oral presentations and publications in conference proceedings of International Drying Symposium, Gifu, Japan, 07-10.08.2016 (OVGO and WU). Posters at IDF World Dairy Summit, Rotterdam, Netherlands, 17-21.10.2016 (NTUA and LBORO). Oral presentation at SUSMILK final conference, Santiago de Compostela, Spain, 22-23.09.2016 (IRTA). Two talks at 8th International Conference on Food Factory, Laval, France, 19-21.10.2016 (IRTA).

Industrial Advisory Board (IAB)

The project had an industrial advisory board which consisted of big industrial companies to steer the project in the right direction with industrial relevant input. The members were:
- Capsa
- Danone
- DSM
- Fonterra
- Friesland Campina
- GEA Niro

This companies were informed on the progress of the project and based on the progress they provided feedback on the industrial relevance of the topics and provided input on relevant data and numbers for the research. Throughout the project also bilateral meetings were organized with individual members of Enthalpy and interested companies from the IAB to facilitate potential follow-up projects on the Enthalpy technologies.

Dissemination collaborations: are a useful tool to increase the audience size and reach a larger number of stakeholders. Collaboration was established in with the two other KBBE project funded under the same topic as ENTHALPY, EnReMilk (www.enremilk.eu) and SUSMILK (www.susmilk.com). Each other’s news and events were posted and ENTHALPY participated in the final SUSMILK conference. Also ENTHALPY was invited to a session organised by SUSMILK at the Euro Food 2016. Both SUSMILK and EnReMilk contributed with presentations to the ENTHALPY final conference in October 2016.

Exploitation of results

The exploitable foreground in ENTHALPY refers to results with commercial and industrial applications in the milk powder production. ENTHALPY successfully improved or developed a number of technologies despite setbacks in project implementation (bankruptcy of a partner). Most of the exploitation plans are currently under discussion among the developers and their corresponding technology transfer departments and attorneys. The main exploitable results are listed below as follows:
• An enzymatic cleaning process including a new enzymatic formula to clean CIP facilities (TRL 7) – this system prototype was demonstrated in the demo facilities of IRTA in 2016 and presented to interested stakeholders during the final ENTHALPY conference on 26 October 2016. This innovative process and product are covered under the patent application No. PCT P28012EP00 (Pending). Nonetheless, controlled dissemination activities have been developed in order to promote and facilitate further research and development plans and commercialization of this result. This result was promoted at the EFFoST conference at the Austrian Economic Chamber of Vienna on 28-30 November 2016. Also two related papers are expected for 2017 including related knowledge covered in the aforementioned patent application. This innovative solution can be use in food and beverage production lines outside dairy. In this sense, its developers UAB and ITRAM have agreed to the creation of a spin-off company to manage the production, commercialization and further research of this result.
• Simulation tool Graphical User Interface (GUI) (ENTHALPY App) is application developed by Biobased Chemistry and Technology (BCT) group of Wageningen University & Research (WUR). This group is currently discussing with its technology transfer department in order to set up a business plan for further research and commercialization of this innovative tool that can be used to perform quick simulation of different milk powder production processes including the analysis of energy consumption, costs and environmental impact while comparing different stored scenarios. This simulation tool is a software that be run on a PC (TRL8).
• A pilot plant was developed in WP3 for the validation and demonstration of a thermal solar system working with pressured water, connected to a food processing equipment at IRTA (Monells, Spain). IRTA will exploit this pilot facility by renting it for testing and validation of new food products and processing technologies. In addition IRTA is strongly interested in the dissemination of the test results through publications and communication activities to attract further research projects.
• An innovative atomizer was developed by TNO in order to facilitate the drying process in the milk powder production. TNO will continue the further research and development activities under its B2B projects according the needs of its clients in the food processing sector. The atomizer tests led to the development of an optimized drying tower and know-how that will be exploited by TNO commercially and through further research.
• Measurement equipment for in-line monitoring of particle size distribution and know-how was generated by OVGU in order to understand the behaviour of liquids inside the drying towers. Also AVA developed an In-line monitoring system for particle moisture content inside a pilot plant spray drying tower. AVA will use the experience gained in ENTHALPY to improve their current equipment while OVGU will discuss the exploitation plans with the technology transfer department of that university.
Further research will be needed to validate the use of on membrane distillation and membrane contactor in the food production sector beyond milk powder production.

Impact
The background of the call and the Enthalpy project was the underlying ambition to reduce the usage of energy and water. The prescribed target at the start of the project was 62% of energy reduction. Throughout the project the technologies have been validated by models and in the last year of the project they have been validated at a pilot scale facility. The resulting energy reduction for the combined technologies is 45%, which is lower than the theoretical ambition at the start of the project but is a huge reduction of energy and therefore could form an important contribution to the ambition for a more sustainable food chain. Within cleaning an water reduction of 30-50% was obtained in the pilot trial depending on which equipment was cleaned. So also on water reduction the Enthalpy project shows great potential. Therefore the Enthalpy project has an substantial sustainability impact for the dairy processing.

Of course one of the key points from industry (IAB) was that the sustainability aspects are very important, but that they should always be considered in combination with the product quality. Since a reduction in energy and water is not economically feasible if it is achieved with a reduction of the product quality. Therefore the powders created were analysed and the results were similar to that of conventional powders and in the sensorial analysis there was a positive response to the products. So the quality level has not been altered while a significant sustainability effect has been achieved making the technologies a potential success for future implementation.

This sustainability impact was further investigated in an LCA analysis, which takes into account the energy and water usage, the environmental burden of the equipment and the economic cost and this analysis resulted in a positive analysis for the combined technologies and also a positive analysis for integrating separate units into an existing processing plant.

List of Websites:
Enthalpy website:
http://www.enthalpy-fp7.eu/

Coordinator:
Pieter Debrauwer TNO
email: [e-mail]
phone: [phone]

dissemination manager
Stephen Webb, RTD Services
email: [e-mail]
phone: [phone]

Below are the contact details of each individual partner

Pieter Debrauwer (TNO)
De rondom 1,
5612 AP Eindhoven
PO box 6235
5600 HE Eindhoven
Netherlands

Antonis Politis (NTUA)
National Technical University of Athens
Zographou Campus
School of Mining and Metallurgical Engineering
Office Number 2.16
Iroon Politexneiou 9, GR15780 Athens
Greece

Andrew Stapley (LBORO)
Loughborough University
Department of Chemical Engineering
Loughborough
LE11 3TU
United Kingdom

Evangelos Tsotsas (OVGU)
Otto-von-Guericke-Universität Magdeburg
Institut für Verfahrenstechnik
Lehrstuhl für Thermische Verfahrenstechnik
PSF 4120
39016 Magdeburg
Germany

Ton van Boxtel (WU)
Wageningen University
Biobased Chemistry and Technology
dept. Agrotechnology and Food Sciences
PO Box 17, 6700 AA Wageningen
Bornse Weilanden 9,
6708 WG Wageningen
Netherlands

Paul Deckers (Bode)
Bodec bv
Scheepsboulevard 3
5705 KZ Helmond,
Netherlands

Alexandru Rusu (Biozoon)
Biozoon
Fischkai 1
27572 Bremerhaven
Germany

Oscar Xiques (ITRAM)
ITRAM HIGIENE
c. Miramarges 7, 3-3
08500 Vic Barcelona
Spain

Luciano Falqui (ODC)
Officine di Cartigliano
Via San Giuseppe 2
36050 Cartigliano
Italy

Manuel Gimeno (INOXPA)
Inoxpa SA
C/Telers 57
17820 Banyoles
Spain

Markus Henneberg (AVA)
AVA- Anhaltinische Verfahrens- und Anlagentechnik GmbH
Mittagstrasse 16P
39124 Magdeburg
Germany

Kaushal Kothari (PLC)
PLC Ingredients Limited
Limerick food centre nit 12
Limerick
Ireland

Caroline Hennigs (NATUR)
Naturstoff-Technik GmbH
Marie-Curie-Str. 11
27711 Osterholz-Scharmbeck
Germany

Jan Henk Hanemaaijer (i3)
I3 Innovative technologies
Utrechtseweg 27
6962 AB Oosterbeek
Netherlands