Forschungs- & Entwicklungsinformationsdienst der Gemeinschaft - CORDIS

FP7

PVCROPS Report Summary

Project reference: 308468
Funded under: FP7-ENERGY

Final Report Summary - PVCROPS (PhotoVoltaic Cost r€duction, Reliability, Operational performance, Prediction and Simulation)

Executive Summary:
The main objectives of PVCROPS are:
- Reduction of 30% of the LCoE of PV to achieve 0.14-0.07€/kWh by 2020 and 0.20-0.09€/kWh by 2015 and an increase of 9% in the performance ration of PV systems.
- Enhancement of the grid integration of PV by mitigation PV power fluctuations and integrating energy management and storage to allow 30% of PV penetration by 2020.
For the first objective, PVCROPS has developed two sets of results: a set of design tools composed of technical specifications – simulation tool – quality control procedures oriented to the bankability of PV systems and to assign responsibilities in case of underperformance, and toolboxes for the automatic detection of performance failures for both, BIPV and PV plants.
For the second objective, PVCROPS has developed toolboxes for the prediction of both PV production and PV power fluctuations, and also energy management strategies and the corresponding toolboxes for the integration of batteries in PV systems.
In particular, the PVCROPS results are:
Design tools:
1- A document with technical specifications for grid-connected PV systems ready to include in contractual frameworks (R1).
2- A manual of good and bad practices to improve the quality and reduce the cost of PV systems (R2).
3- A free and open-source simulation toolbox, called SISIFO, for robust design of PV systems (R3).
4- Built-in learning tools, included in SISIFO, destined for professionals, researchers and students in solar energy (R4).
5- Testing kits for the quality control of BIPV and PV plants (R18).
6- A document detailing the quality control procedures to be included in contractual agreements (R19).
Tools for the automatic detection of performance failures:
7- A toolbox allowing detecting operational problems of PV plants from SCADA monitoring systems (R14).
8- A toolbox allowing the detection of performance failures of BIPV through webservices (R15).
9- A database as observatory of real production and performance of thousands of PV systems in Europe (R16).
10- A tool, called SOWEDA, for the free access to real-time irradiance of tilted surfaces with higher accuracy than satellites data (R17).
Prediction tools:
11- Two toolboxes to predict PV energy production on an hourly basis to optimize electricity market sale: one based on parametric methods and one based on non-parametric methods (called PROPHET production) (R5).
12- One toolbox to predict fluctuations of a PV fleet connected to a grid node (R6).
13- Two toolboxes to predict PV power fluctuations, one based on parametric methods and another one based on non-parametric methods (called PROPHET fluctuation), and six energy management strategies to mitigate them (R7).
14- Guidelines for the integration of PV plants into the grid (R8).
15- A document estimating the PV power that can be integrated into the EU electrical networks (R9).
Tools for the integration of batteries in PV systems:
16- Equipment for the energy management and storage control in PV plants (R10).
17- A toolbox for the design and simulation of the energy management strategies in PV plants (R11).
18- An energy management strategy to maximize the economic output of a PV plant (R12).
19- Equipment for the energy management and storage control in BIPV (R13).
20- A toolbox for the design and simulation if the energy management strategies in BIPV.

Thirteen of these 20 results are marketable and exploitations plans have been produced.
Other remarkable results are:
- The creation of two spin-offs (WebPV and SunWings)
- 57 publications in international scientific journals and international conferences.
- 10 patents.
- Seminars and technical visits to the demonstrators have been organized for the market uptake of the solutions and for the general public advice.

The analysis of the performance ration of more than 30,000 PV systems in Europe has allowed to evaluate the impact of our results. Our set of design tools can improve the PR in 6.81% and our tools for the automatic detection of performance failures in 3.13%, totalizing an increase of 9.94%. As the application of these tools does not imply increase of cost, this increase is directly translated into an equivalent reduction of the LCoE.
Furthermore, the tools for the prediction of PV power and the ones for the integration of batteries have covered all of the technical challenges identified by EPIA and the project PVGRID for the so-called “paradigm shift scenario”. This way, from a technical point of view, all of the technical barriers to achieve 12% of EU electricity demand by 2020 and 25% in 2030 have been removed, that means more than 30% of PV penetration in Southern EU countries. These tools have also impact in the LCoE. The prediction tools allow reducing the whole electrical system costs in terms of reducing spinning reserves and of reducing new infrastructures, estimated in an equivalent reduction of LCoE of PV systems of 10%. The integration of batteries in PV systems allows them to collaborate in the management of the whole grid, offering ancillary services that can represent another 10% in the reduction of LCoE.

Project Context and Objectives:
The PV CROPS project responds to the call for proposals identified as “Energy 2012.2.1.1: Reliable, cost-effective, highly performing PV systems”. It addresses the three key objectives of the call relative to solar photovoltaics (PV), at the system level:
1) Improvement in the performance, reliability and lifetime of PV systems
2) Reduction in the cost of PV systems
3) Enhancement of the integration of PV into the grid

The two first objectives are directly related to one important Key Performance Indicator (KPI) established by the Solar Europe Industrial Initiative of the SET-PLAN (SEII) which is the reduction of the Levelized Cost of Energy (LCoE) of PV generation. The LCoE (in €/kWh) depends mainly on four parameters related to a PV system:
• its initial cost (in €/Wp);
• its energy production (in kWh/kWp), which is related to its performance;
• its operation and maintenance costs (in €/year), which are related to its reliability;
• and its lifetime and that of one of its main components.

Therefore the minimization of LCoE is a multiparameter optimisation exercise in cost, performance, reliability and lifetime. We can thus reformulate these three main objectives of the call advantageously into two more general targets:
1) Reduction in the LCoE of PV generation
2) Enhancement of the integration of PV into the grid

Therefore, we can advance the PV CROPS main objectives that correspond to these call targets:
Objective 1: Reduction of 30% of the LCoE of PV generation to achieve 0.14 – 0.07 €/kWh by 2020 and 0.20 – 0.09 €/kWh by 2015 and an increase of 9% in the performance ratio of PV systems.
Objective 2: Enhancement of the grid integration of PV by mitigating PV power fluctuations and integrating energy management and storage to allow 30% of PV penetration by 2020.

The sub-objectives related with objective 1 are the following:
SO1.1- Development of toolboxes for PV system design optimization, robust modelling, advanced energy losses scenarios and advanced simulation.
SO1.2- Development of technical specifications, ready to be used in contractual frameworks, to foster excellence in the design and installation of PV systems.
SO1.3- Development of diagnosis toolboxes for the detection of hidden problems affecting energy performance and PV system lifetime.
SO1.4- Development of quality control testing procedures to be included in the contractual agreements allowing PV systems performance, reliability and lifetime to be optimized.
SO1.5- Development of software and hardware toolbox solutions for the testing and commissioning of PV systems.

The sub-objectives related with objective 2 are the following:
SO2.1- Development of optimized tools allowing operators of PV plants to predict PV electricity production on an hourly basis and with at least three days notice, in order to minimize the deviations from predictions to production and maximize the price for the sale of electricity.
SO2.2- Development of tools allowing grid operators to predict PV power fluctuations on an hourly basis and with three days notice at every node in the grid, and to activate protection mechanisms (also developed by PV CROPS).
SO2.3- Development of procedures to mitigate PV power fluctuations to increase PV penetration into the grid.
SO2.4- Development of guidelines to determine how and where to connect new PV plants in the topology of the grid to preserve grid stability.
SO2.5- Incorporation of energy accumulation in PV plants to increase the PV penetration into the grid.
SO2.6- Development of software and hardware solutions to manage and optimize the energy flows in PV systems with energy accumulation.
SO2.7- Adaptation of different energy accumulation technologies for their use in PV systems, including the use second-hand batteries coming from the electric vehicle market to reduce costs.

The results will be lead to the following quantified impacts:
• 30% LCoE reduction by 2020.
• 9% Performance Ratio increase.
• 30% of PV penetration in the grid by 2020.

Project Results:
INTRODUCTION

PVCROPS has developed two sets of results: one oriented to the increase of performance of PV Systems, and another one for the high penetration of PV systems into the European electrical networks. Both of them contribute to the reduction of LCoE.

The first one is composed of a set of design tools composed of technical specifications – simulation tool – quality control procedures oriented to the bankability of PV systems and to assign responsibilities in case of underperformance, and toolboxes for the automatic detection of performance failures for both, BIPV and PV plants.

The second one is composed of toolboxes for the prediction of both PV production and PV power fluctuations, and also energy management strategies and the corresponding toolboxes for the integration of batteries in PV systems.

In particular, the PVCROPS results are:
Design tools:
1- A document with technical specifications for grid-connected PV systems ready to include in contractual frameworks (R1).
2- A manual of good and bad practices to improve the quality and reduce the cost of PV systems (R2).
3- A free and open-source simulation toolbox, called SISIFO, for robust design of PV systems (R3).
4- SISIFO also includes built-in learning tools destined for professionals, researchers and students in solar energy (R4).
5- Testing kits for the quality control of BIPV and PV plants (R18).
6- A document detailing the quality control procedures to be included in contractual agreements (R19).

Tools for the automatic detection of performance failures:
7- A toolbox allowing operational problems of PV plants from SCADA monitoring systems (R14).
8- A toolbox allowing the detection of performance failures of BIPV through webservices (R15).
9- A database as observatory of real production and performance of thousands of PV systems in Europe (R16).
10- A tool, called SOWEDA, for the free access to real-time irradiance of tilted surfaces with higher accuracy that satellites data (R17).

Prediction tools:
11- Two toolboxes to predict PV energy production on an hourly basis to optimize electricity market sale: one based on parametric methods and one based on non-parametric methods (called PROPHET production) (R5).
12- One toolbox to predict fluctuations of a PV fleet connected to a grid node (R6).
13- Two toolboxes to predict PV power fluctuations, one based on parametric methods and another one based on non-parametric methods (called PROPHET fluctuation), and six energy management strategies to mitigate them (R7).
14- Guidelines for the integration of PV plants into the grid (R8).
15- A document estimating the PV power that can be integrated into the EU electrical networks (R9).

Tools for the integration of batteries in PV systems:
16- Equipment for the energy management and storage control in PV plants (R10).
17- A toolbox for the design and simulation of the energy management strategies in PV plants (R11).
18- An energy management strategy to maximize the economic output of a PV plant (R12).
19- Equipment for the energy management and storage control in BIPV (R13).
20- A toolbox for the design and simulation of the energy management strategies in BIPV.

Thirteen of these 20 results are marketable and exploitations plans have been produced.

Other remarkable results are:
- The creation of two spin-offs (WebPV and SunWings)
- 57 publications in international scientific journals and international conferences.
- 10 patents.
- Seminars and technical visits to the demonstrators have been organized for the market uptake of the solutions and for the general public advice.

In the following sections, we will detail these results following the above-mentioned structure.

DESIGN TOOLS

Simulation tool

This simulation tool is in the set of design tools to improve the performance and reduce the LCoE of PV systems.
Due to the bank financing is associated to the percentile 90 of the estimation of the energy production of the PV system, the simulation tool must allow reducing the uncertainty of the estimation. And due to the contractual frameworks, the simulation tool must assign responsibilities if the actual performance of the PV systems is under the expectations established in the contract.
The result of the work is a free-software simulator of photovoltaic (PV) systems, called SISIFO, which is available at www.sisifo.info , that uses models that require as inputs just information guaranteed by the manufacturers. Therefore, the objective of assigning responsibilities in case of under performance is covered. Furthermore, the design of the SISIFO tool is coordinated with the technical specifications, which allows evaluating and reducing the uncertainty of the estimations, increasing the financing and the bankability.
SISIFO allows the simulation of different types of grid-connected PV systems, such as large grid-connected plants and building-integrated installations, which is by far the main application of PV. SISIFO is based on models and energy scenarios whose suitability has been validated in the commissioning of several tens of mega-watt size PV plants installed in Spain, Portugal, France and Italy, whose aggregated capacity is nearly 300MW.
Besides, the tool has been developed taking into consideration the recommendations and necessities from different actors of the PV community, such as promoters, manufacturers, installers, professional associations and end-users. In particular, independent PV experts have reviewed several aspects of the tool (usability, technical documentation, modelling, simulation options, etc.) and they have suggested modifications and improvements.
SISIFO may be used for professional or educational purposes. On the professional side, there are obvious applications, such as the realisation of feasibility studies, the comparison of performance for different technical designs, or the analysis of uncertainty on the energy production when different solar resource databases are used. But, as said before, the main feature of SISIFO, which is not usually addressed by other software simulation packages, is the possibility of using it in the frame of quality control assurance procedures. Indeed, implemented system models are uniquely based on parameters that can be extracted from product datasheets, which could be guaranteed by manufacturers. Next, these parameters can be verified experimentally after PV system installation by means of on-site ad-hoc testing procedures, which have been also developed in PVCROPS (WP9). The experimental verification of these parameters may be used not only to check contractual warranties but also for tuning the simulations and reducing the uncertainty associated to PV energy yields predictions.
On the educational side, SISIFO may be used, for example, in academic or training courses on PV systems. For this purpose, a detailed technical reference manual has been developed, which can be used as textbook. Besides, the following learning tools have been integrated in SISIFO:
- Interactive recommendations, which inform the user on the best design practices when some input parameters are selected.
- Video-tutorials on how to use the tool.
- Forum, which allows the communication between users community and receive feedback concerning bugs, improvements, etc.
- Virtual laboratory. A set of simulation exercises has been created, which can be used in PV engineering simulation courses or just as self-learning material.
SISIFO is online and free-software, it can be redistributed and/or modified under the terms of the GNU Affero General Public License version 3 (AGPLv3). This license allows the users the access to the source code and the freedom to run, copy, distribute, study, change and improve the software. This allows the PV community to contribute to its evolution. As example of the latter, the developer of the 3D design tool Skelion for thermal and photovoltaic systems (www.skelion.com) will integrate the same electric models developed in SISIFO.
PVCROPS has also integrated this tool with OpenDSS tool, allowing simulating not only the behaviour of a PV system but also its impact in the electrical network. In order to show its potential, SISFO has been used to simulate the productivity of a 150MW PV plant in Morocco.

Technical specifications

In order to achieve a reduction of the LCoE, one of the most important actions is to have an initial design of the installation optimized to avoid undesired mistakes, failures and underperformance of the different PV system components. So, it is needed to develop technical specifications that take into account all of the problems encountered in the PV systems already installed to optimize the future performance PV systems, to obtain its maximum profitability and to foster excellence in the design and installation of PV systems.
PVCROPS has opted for developing a set of technical specifications with the mentioned objective but also specially focused on the bankability of PV systems and on their direct application to contractual frameworks.
In the one hand, bankability means accurate energy yield estimation. So a simulation tool is required. But bankability also means financing, usually linked to the percentile 90 of the previous estimation. So, low uncertainty is also required.
In the other hand, contractual frameworks mean technical specifications to be included in the contract or in the tender calls. And the companies involved will be very thankful if the procedure to check the fulfillment of the specifications, that is to say, the quality control tests, is also included. But contractual frameworks also means assign responsibilities in case of troubles. So, technical specifications and simulation tool must be based just in parameters guaranteed by manufacturers.
So, Technical Specifications and Quality Control Procedures plus low uncertainty have been the design criteria for our Technical Specifications for contracts and bankability. And Simulation tool plus Guaranteed Parameters has produced our SISIFO simulation tool.
The problems identified during the first period of the project, together with the requirements of bankability and of the contractual frameworks, have been the base for the development of our technical specifications for grid connected PV systems (corresponding to the result R1 of the contract “Technical specification for grid-connected PV systems ready to include in contractual frameworks”).
The final first version has 92 pages and includes not only the technical specifications that should meet a general PV system to assure a good energy production, but also the quality control procedures (described later in this report) which define the tests that must be done to check if the final implementation of the PV installation fulfils these technical specifications. The document also includes two annexes. The first one shows a comparison between different PV energy performances models, including the one implemented in the PVCROPS project. The second one shows the experimental work done to establish criteria to determine when an overheating in a PV module can be considered as defect (hot-spot) and should be rejected.
We have opted for a single document in order to be easily applicable to any PV installation and to any contractual framework or tender call. In order to allow the uncertainty in the estimation of the productivity (related with the bankability), this document proposes several means such as the installation of reference modules as sensors of irradiance and solar cell temperature. In order to assign responsibilities, this document proposes a simple model to calculate the productivity of the PV systems that uses just parameters guaranteed by manufacturers.
In order to illustrate its usefulness, they have been adapted to the case of a 150 MW PV plant to be installed in Morocco. In fact, as this is a project that is going to be implemented in a close future, this report can be useful as a part of the technical requirements of the tender call. This PV plant is going to be constructed at Tafilalt, which is located in the province of Zagora, in the South of Morocco. It is going to be divided in two different generators: one of them, of 75 MW, should be mounted in a static structure faced South and tilted between 25º and 35º; the other one, of 75 MW too, should be mounted in a one-axis North-South horizontal solar tracker.

Quality control procedures and testing kits

Most of the PV installations use to be financed as “Added-Value Projects”; that is, the bank requires the electricity production of the installation to guarantee the recovery of the investment. So, the bank ask for quality control procedures known as “Due Diligence” in order to check if the PV installation has been built as it had been specified in the contract. These quality control procedures have to be designed to evaluate the PV installation in different stages, during design phase, construction phase and operational lifetime. The final objective of these procedures is to ensure that the quality of the devices installed is the declared previously by the manufacturers so that the performance, reliability and lifetime of the installation are as good as possible, resulting in an improvement of these parameters, as well as in a decrease of the LCoE.
PVCROPS has developed quality control procedures for the commissioning stage (Provisional Acceptance Certificate, PAC), for the first year of operation (Final Acceptance Certificate, FAC) and for the continuous operation of the PV plant. The first version of these procedures was reviewed by experts and several actors of the PV industry.
The final version of the quality control procedures (Result 19 of the contract) has 92 pages and includes the quality control procedures that are useful to check if the final implementation of the PV installation fulfils the technical specifications that are also included in the report.
In order to execute the quality control procedures, the PVCROPS team has developed and implemented testing kits that allow performing all the tests previously defined. Some of them are self-made devices to do specific tests (cheaper than the ones available in the market and with added capabilities) and other ones are the combination of commercial devices that have been specifically prepared to check the performance of different parts of the PV system. These testing kits are:

• Reference PV modules, which allow reducing the uncertainty in the measurement of the on-plane irradiance and PV module temperature.
• I-V curve tracers, which allow on-site testing of the performance of individual PV modules or whole PV arrays up to 2 MW.
• Climatic box, which allow testing on-site or at laboratory of the parameters guaranteed by PV manufacturers.
• PID tester, to test on-site or at laboratory if the PV modules are prone to/affected by Potential Induced Degradation (PID).
• Hot-spot tester, to detect hot-spots inside PV modules, automatically or manually.
• Full PV system tester, which allows executing a detailed performance analysis of the whole PV system.

These testing kits are described in a single report of 81 pages which: explains what are the different testing kits; presents the possible alternatives for each testing kit, if there are several implementations (for example, one oriented to PV plants and another one oriented to BiPV; or one automatic and another one manual; or one more expensive and another one cheaper; etc.); informs about what kind of PV devices can be checked with each testing kit (PV modules, PV inverters, whole PV system, etc.); and shows the result of measurements obtained in real installations.
In order to exploit the testing kits and the service that can be offered with them to the PV community, a spin-off company called SunWings has been implemented and registered. SunWings offers services for local and international PV service providers.
Finally, PVCROPS has contacted with the EU project SOPHIA in order to facilitate the knowledge of what are the facilities of this project able to perform the tests proposed in our project.

TOOLS FOR THE AUTOMATIC DETECTION OF PERFORMANCE FAILURES

PVCROPS has developed tools for the automatic detection of performance failures in both BIPV and PV plants. These tools are key means to achieve the goal of increase the performance and reduce the LCoE of PV systems.

Automatic detection of performance failures in BIPV systems

Regarding BIPV, PVCROPS has developed a method that is capable of detecting operational problems on the sole basis of the identification of abnormal variations of a performance indicator that is characteristic of the proper functioning of the installation, and which construction does not require the knowledge of the operating conditions of the BIPV systems. One particular requirement for this indicator is therefore its stability during the normal functioning of the installation. The indicator that we have developed has been designated by Performance To Peers (P2P), because it is based on comparisons between neighboring and similar installations. The fault detection threshold separates two different P2P populations: correct functioning and failures.
This method has been developed on the data of more than 5,500 BIPV systems located in Belgium and monitored by Rtone from 2011 with a temporal resolution of 1 data every 10 minutes. A toolbox has been developed for the automatic fault detection by P2P and that is being commercialized by a spin-off (WebPV) created in the framework of the project.
This toolbox has been validated through the automatic fault detection of 5,550 BIPV installations. Faults were detected on more than 500 of these installations from January to October 2015. A validation campaign was set up from June to October 2015 to verify whether the faults that were detected by the toolbox were real, and whether the toolbox was able to detect all the major faults hindering the performance of the PV installations. Our results confirm that it is possible to carry out automatic fault detection procedures without solar irradiation data. P2P proves to be more stable than PR most of the time, and thus constitutes a more reliable performance indicator for fault detection procedures.

Automatic detection of performance failures in PV plants

Regarding PV plants, PVCROPS has developed a toolbox with different procedures for the performance analysis of PV plants, identifying problems associated to SCADA alarms, possible hidden problems, degradation, anomalous trends etc. This toolbox is based on the following steps:
1- Calculation of the PV plant power at Standard Test Conditions (STC)
2- Calculation of the theoretical production
3- Analysis of the technical availability
4- Comparison with the actual production and management of the information
The good performance of this tool has been validated, mainly checking the good behaviour of the theoretical energy calculation and the STC power calculation. The result of the theoretical energy validation is that, in the 82% of the events, the error (difference between theoretical and real production) is less than ±1.5% and the maximum error committed (±4.5%.) represents just 4% of the events. The PSTC annual mean have a similar value (0.925±0.001) and the maximum deviation found during a year is ±1%. Moreover, when three years are aggregated the maximum deviation is ±2%. Furthermore, the toolbox has been implemented at Amareleja PV plant (46MW) showing its functionality and good performance, and currently the toolbox is being exploited by ACCIONA.

EU map of real-time irradiance on tilted surface
PVCROPS has also developed other tools related with the performance of PV systems. The first one is a website with the EU map of real-time irradiance on tilted surface (SOWEDA), data commonly used to analyze the performance of PV systems. It is based on an alternative method to provide free or low-cost hourly solar irradiation data for the whole European territory through a web platform. This allowed us to provide hourly solar irradiation data with a spatial resolution of 1 data/(12x12km2). The method that we have developed makes use of two completely different and independent sources of information to estimate the solar radiation. First, the PV systems are used as sensors that measure the tilted solar radiation from the power output that is read at their energy meters. Second, the combination of clear-sky simulations and weather conditions data is used to generate hourly solar radiation. This novel method allows obtaining hourly global solar radiation data which accuracy is at least as good as satellite-based solar radiation data, or even better. It also allows obtaining the data at a very low cost, and for the whole European continent. In the presence of PV systems equipped with monitoring hardware, it also allows obtaining data with very high spatial resolution, depending on the distance separating the location where the solar radiation is desired from the nearest BIPV systems. SOWEDA tool is available at www.soweda.net . Advanced services related with this tool are offered by the spin-off of the project WebPV at www.webpv.net

Analysis of the state of the art of PV systems in Europe

The second one is an analysis of the state of the art of PV systems in Europe. PVCROPS has analysed the data from more than 31,000 PV installations in Europe. The mean Energy Yield of the PV systems located in each of the four reference countries is 1115 kWh/kWp for France, 898 kWh/kWp for the UK, 908 kWh/kWp for Belgium, 1450 kWh/kWp for the PV plants in Spain mounted on a static structure, and 2127 kWh/kWp for those mounted on a solar tracker in Spain.
We have shown that the distribution of the yearly integrated PR can be modelled well using a Weibull distribution for PR values ranging from 0.6 to 0.9. This range of values represents the majority of the PV systems, and we suggest that they are representative of the state-of-the-art for PV systems in Europe. We suggest that the typical PR value for the PV systems installed in Europe over the last few years is 0.79, and that the value for the PV systems installed in 2015 is 0.81. The corresponding mean values are respectively 0.76 and 0.78. The relationship of PR with different PV module technologies and inverters has been also analysed.
The wide disparity in yearly integrated performance ratio, between 0.6 and 0.9, implies that there is a difference of some 30% between the best and the worst performers. Ideally, the PV sector should aim at reaching PR values over 0.85 for most of the PV systems to be installed in the future. More quality controls and further improvement in the state of the art are therefore a very promising option towards a leap in overall performance, which could lead to an average value of PR over 0.85, representing an improvement in performance around 10%, and a corresponding reduction in LCoE of the same order of magnitude. The details of this analysis can be found in the deliverable 8.5 of the project.

PREDICTION TOOLS FOR THE HIGH PENETRATION OF PVSYSTEMS IN THE GRID

Toolboxes for the prediction of PV production

PVCROPS has explored two main types of modelling to predict the AC production given the required inputs:
• The parametric model, which conceives the PV system as a white box where each subsystem can be modelled using a collection of parameters.
• The nonparametric model, which conceives the PV system as a black box. This approach does not presume any knowledge of internal characteristics and processes of the system. Instead, it is a data-driven model that estimates the behavior of the system from a historical time series of inputs and outputs.

Both approaches have been implemented and validated in order to develop two different toolboxes for the prediction of PV energy production.

The parametric model uses predicted ambient temperature and predicted horizontal irradiance as input variables to predict the hourly AC power of the PV plant. This approach includes a detailed modelling of PV plant components that are mainly based on standard information provided by manufacturers or promoters, which can also be verified experimentally by on-site quality control testing procedures. The toolbox is able to simulate PV energy production for any location just introducing the parameters (geographical data, PV system configuration and simulation option).
From the validation of the toolbox based on the parametric model, it can be highlighted that:
• Considering the whole model, the source that introduces the greatest deviation is the prediction of Ta and G0 (i.e., the inputs).
• If the model is fed with accurate meteorological variables, the error of the model is less than ± 5%.
• The higher the daily clearness index is, the more accurate the model is. This fact is relevant from the energy production perspective.
• The first part of the model, that converts global horizontal irradiance and ambient temperature in in-plane effective irradiance and cell temperature, introduces more error than the second part which simulates the losses in each element of the PV installation.
This toolbox is being exploited currently by ACCIONA.

The non-parametric model uses forecasts of several meteorological variables (produced by a Numerical Weather Model), and spatial indexes (estimated from the NWP variables) as inputs to predict the hourly AC power of the PV plant. The model is constructed upon a machine learning tool, the Quantile Regression Forests, which must be trained with past series of both NWP forecasts and AC power measurements. This tool is able to produce both a central forecast (median) and a confidence interval, which is useful to assess the reliability of the forecast. It has been implemented under R environment. Two packages that resumes the methodology, meteoForecast and PVF, have been published under a GNU-GPL licence. An online graphical toolbox (PROPHET – PV production) is freely available at http://vps156.cesvima.upm.es:3838/predictPac/.
From the results of validation, it can be said that:
• The training set length has no significant impact on the model performance, especially with time series longer than 15 days.
• The presence of the variable of irradiance forecasts and/or calculated extraterrestrial irradiance (Bo0) in the training set leads to better predictions.
• There is no need for prediction of lots of meteorological variables, as long as irradiance data is available.
• The best skill scores range 0.324 to 0.361. These results compare satisfactorily with those reported in Bacher.Madsen.ea2009 with a set of forecast methods of AC power for next day horizons. These authors published skill scores up to 0.36 for aggregated forecasts corresponding to the average power of a set of 21 different PV systems in a region. In contrast, our proposal is focused on the forecast of different PV plants as separated entities.

A comparison between both methodologies was made in terms of daily energy error using the Mean Bias Error (MBE) and the Mean Absolute Error (MAE), both normalized respect to the energy generated during the same day the statistic was computed (cvMBE and cvMAE, respectively). The cvMAE measures the goodness of the predictions for applications requiring hourly predictions during a period of a day, whereas the cvMBE is an index of the goodness of the total daily energy production. The following table summarizes the results of the validation:

Statistic Method 0 ≤ KT < 0.531 0.531 ≤ KT < 0.687 0.687 ≤ KT ≤ 1
cvMBE Parametric 2.9% 4.7% 3.4%
Non-parametric 1.2% 0.1% 0.1%
cvMAE Parametric 9.3% 9.0% 6.1%
Non-parametric 8.7% 6.5% 2.2%
Table. Comparison between models in terms of the error in the daily energy prediction according to different clearness indexes.

Toolboxes for the prediction of PV power fluctuations

Over the last few years, the considerable increase in the number of multi-MW PV plants has led system operators (TSO and DSO) to express concern over potential PV short-term power fluctuations caused by transient clouds. For intervals of less than 10 minutes, these fluctuations are directly absorbed by the electricity system in the form of frequency variations, yet without the utility operator having the response capacity to correct the imbalances. If these frequency deviations exceed the permitted limits, then there is a risk of a power system failure. This variation not only may cause extra operating costs for committing costly reserve units but also poses one of the great challenges for grid operators who are in charge of its integration into conventional networks.

PVCROPS has developed technical solutions and two toolboxes in order to make PV generation friendlier to the grid stability minimizing the potential risk of PV power fluctuations without increasing the costs. These novel results allow a high penetration of PV systems into the EU electrical system in safety and, second allow reducing the levelized cost of PV electricity. Ultimately, all the procedures and models developed in the frame of this work package has been reveled site-independent, which then can be applied worldwide safely.

In order to contribute to solve these problems, PVCROPS has developed the following solutions:
1. Two toolboxes, according to two different methodologies, to predict PV power fluctuation.
2. Procedures to mitigate PV power fluctuations into the grid.
3. Guidelines to determine how and where to connect new PV plants in the topology of the network.
4. Protection mechanisms for the risky days when dangerous PV fluctuations are unavoidable.

To develop these solutions, PVCROPS has developed exhaustive monitoring campaigns in several real PV plants, producing a database that comprises more than 75 MWp monitorized with a time sampling up to 1 second and remains in operation. Without doubt, has no parallel in the state of the art and it is an exceptional observatory to the analysis of PV power fluctuations.

PVCROPS has created the first toolboxes in the state of the art for predicting PV power fluctuations. Two different approaches, leading to two different toolboxes, are presented: based on uncertainty and non-parametric models.

The first toolbox is based, first, on the detailed modeling of the PV plant (parametric model), and second, on forecasting uncertainty as a way of predicting the risk of fluctuations. In principle, those days with a low uncertainty in the prediction corresponds to very clear or very cloudy, which are relatively easy to be predicted. On the other hand, those days with considerable fluctuations, the uncertainty in the prediction increases. The uncertainty in irradiance forecasting is also transferred through the PV plant parametric model to power output uncertainty (parameter F). This make possible to predict days with either a low, medium or high fluctuation risk (parameter T). In order to add a plant parametric model, temperature forecasting is also required as input data. One year validation of this method leads to a successful result that only 1.6% risky days were not detected.

This methodology has been implemented in an easy-to-use toolbox in Matlab®, which allows determining the fluctuation risk from basic meteorological prediction data. This toolbox can be easy translated to C# language and implemented in a PV plant or TSO/DSO SCADA in order to predict the risk of PV power fluctuations and take the consequential actions. This tool is being exploited currently by ACCIONA.

The second toolbox does not presume any knowledge of internal characteristics and processes of the PV system. Instead, it is a data-driven model that estimates the fluctuations of the system from a historical time series of inputs and outputs. This approach uses machine learning methods that learn from data and external variables (provided by a Numerical Weather Model) to produce forecasts. The machine learning model (Quantile Regression Forest) is trained with the train time series. A prediction of the median (quantile 5) and a confidence interval (quantiles 1 and 9) are produced with the test time series, providing the fluctuation magnitude. After validating the prediction of fluctuations during 600 days, the mean daily error of the prediction is very good (lower than -5%). The free open-source toolbox, named as “PROPHET – PV power fluctuations” has been built in a web page. Current url of the toolbox is http://vps156.cesvima.upm.es:3838/predictRamps/. The full code implementing the procedure is freely available from the repository https://github.com/iesiee/PVF.

Toolbox to predict fluctuations of a PV fleet connected to a grid node

Furthermore, the frequency domain analysis of PV production data of several dispersed plants in Navarra has revealed a simple and effective model in order to simulate the fluctuations generated by a PV fleet. The model developed only needs as an input the irradiance measurement in any point inside the perimeter of the PV fleet, the number of PV plants N and the average surface of the PV plants S, that is, so well-known parameters. This model has been simulated for one year 1-s data irradiance data for 6 PV plants and validated at the light of two important parameters from an electric system operator point of view: daily maximum fluctuation and daily spinning reserves needed to smooth fluctuations, reaching RMSE<=2%. The details can be found in the deliverable 5.2 of the project.

Protection schemes against unavoidable fluctuations

Regarding the protection schemes against unavoidable fluctuations, PVCROPS have developed solutions based on natural smoothing mechanisms (due to the PV plant area and due to geographical dispersion of PV plants), on prediction of fluctuations, on power curtailment, and on Energy Storage Systems with bidirectional charger at the point of the common coupling of the PV plant.

The proposed procedures can be divided in two categories, depending on the use or not of any kind of energy buffer in order to filter the PV output variability.

a) Protection schemes without storage
There are two natural factors known to smooth short-term irradiance fluctuations in relation to PV power fluctuations. The first is smoothing due to the PV plant size, and the second is due to the geographical dispersion of a PV fleet. PVCROPS has developed empirical equations to calculate the value of the peak power fluctuation depending of the size and the number of PV plants. The smoothing due to dispersion is greater than that due to plant size. In summary, PVCROPS has identified two natural mechanism in order to smooth PV power fluctuations based only on how PV plants are installed:
• The power fluctuations occurring in a PV plant are smoothed in relation to the irradiance variations. The smoothing factor is the PV plant size, and is relevant for times of just a few seconds, disappearing as the sampling time increases. Likewise, an empirical equation is proposed, and which provides the value of the peak power fluctuation in relation to the PV plant size.
• The geographical dispersion of the PV plants is a highly effective way of smoothing the power fluctuations, even for ten minute sampling intervals. It is sufficient to locate two PV plants at a distance of 6 kms, one from the other, to ensure that the fluctuations over 10 minute intervals are independent of each other and are smoothed out when combined. Furthermore, an empirical equation is proposed, that will give the value of the peak power fluctuation, based on the number of plants grouped together and the PV plant size.

The results obtained support the idea that PV plants power fluctuations are attenuated more by the number of plants grouped than by their size for the same rated PV power installed. This result is very valuable when planning the distribution of PV systems in the topology of the grid. Taking into consideration this topology, mitigation of PV power fluctuations without storage can be achieved.

PV power prediction is another way of limiting the effect of fluctuations. Consider a scenery where the penetration of PV energy into the electricity mix does not compromise the grid stability. In those circumstances, the traditional generator can compensate the maximum PV power fluctuation that can take place without any risk for the system. However, energy reservoirs are necessary in order to compensate the fluctuations, specially during very fluctuating days, influencing in a significant way the final electricity price. Logically, in the absence of any kind of PV fluctuations methods, these reserves must be prepared either during clear o cloudy days, with the additional costs. Although these fluctuations can achieve high values, the frequency occurrence is relatively low. Hence, it has sense to develop a method in in order to classificate a day-ahead the risk of PV power fluctuations. These will allow the TSO to efficiently schedule the spinning reserves, be more efficient and then, decrease the final energy price. PVCROPS has respond to this need, creating a toolbox for predicting PV power fluctuations, as explained in the previous section.

Finally, an effective scheme in order to ensure that PV power fluctuations are not going to cause a significant impact into the grid stability is to curtail the maximum power that the PV plant can inject into the grid. Logically, this strategy involves production losses, hence, this protection scheme must be consider as a last resort, a similar view than EPIA. PVCROPS has quantified these losses. Regarding the limitation rules, two possibilities have been identified. The first possibility is to do a complete curtailment of the PV power and inject constant power to the grid, in order to avoid any possibility of PV power fluctuation. The second possibility is to limit the only positive fluctuations to the maximum allowable ramp via the inverter MPP, because for the negative ones any kind of storage is needed. This strategy is coherent in a scenario where the regulation capability of the TSO it is only limited in the upwards fluctuation (similar to a negative change in the demand).

b) Protection schemes based on the energy storage system at PV plant
Taking advantage of the energy management strategies developed in the framework of PVCROPS and that will be explained later, we have developed protection schemes against PV power fluctuations.

The direct solution is the installation of an ESS with bidirectional charger at point of the common coupling (PCC) of the PV plant. Here, two different options are possible depending on how the PV+ESS system is controlled: exclusively by the bidirectional charger or involving the PV inverters. This last option also can consider if the PV inverters have communication with the bidirectional charger or not. The protection schemes proposed have been simulated for 5-sec power outputs over the course of one year (2013) at the 38.5 MW Moura PV plant for a maximum allowable ramp-rate limitation of rMAX=2%/min, showing excellent results. The details can be found in the deliverable 5.4 of the project.

Guidelines for the integration of PV plants into the grid

PVCROPS has elaborated some guidelines in order to correctly distribute the PV plants and minimize the effect of PV power fluctuations. The guidelines are based on the general rule of taking advantage of natural smoothing mechanisms. Making use of the size and dispersion smoothing factors, a significant reduction in power fluctuations can be achieved without increasing the levelized cost of PV energy.

The methodology followed moves from general to particular considerations, always seeking that a harmful PV fluctuation cannot take place. The summary of the guidelines is the following:
1. Determining the maximum PV penetration index: maximum PV power that can be installed in a particular power system without risk for the power system stability. Two conditions must be met. First it is necessary to guarantee that the maximum possible fluctuation of the PV power installed at noon must be less than the value which typically produces the maximum deviation in the frequency. Secondly, decorrelation between fluctuations must be ensured (PV plants spaced sufficiently far apart from each other).
2. Determining the number and size of the PV plants. Once the maximum PV power that can be installed is known, next step is to define the number and size of PV plants (clustering degree) in order to delimit the magnitude of the fluctuations of the PV fleet.
3. Determining the worst fluctuation that can take place. The magnitude of the worst fluctuation can be calculated once the layout and the particular shape of the future PV planned plants are established. If the magnitude of the fluctuation is beyond what should be permitted, a rethinking of the PV fleet distribution is needed.
4. Installation of centralized ESS. Finally, and as a last option, there can be sceneries (current or future) where, even though all the previous guidelines have been attempted to apply, the fluctuations can be harmful for the grid. For those situations, the installation of an ESS in a grid node is the most cost effective solution.
As a general conclusion, all the guidelines contained here present two great advantages over the state of the art. Firstly, they are based on empirical equations and models obtained from an unparalleled database (up to real 75 MW of PV plants monitored, every second). And as an added benefit, the analysis of these models reflects that all proposals can be applied worldwide, because they do not depend on geographical/climatic parameters. The details can be found in the deliverable 5.3 of the project.

Document estimating the PV power that can be integrated into the EU electrical networks

PVCROPS has estimated the PV power that can be integrated in current EU networks if PVCROPS solutions were implemented. The analysis is based on previous European studies (EPIA and PVGRID). The methodology has been, first, analysing the technical challenges identified by these previous initiatives; second, to connect these challenges with the solutions provided by PVCROPS and evaluating the level of fulfilment; and third, to evaluate if the PV power foreseen by these initiatives could be implemented with the technical solutions of PVCROPS. The details can be found in the deliverable 5.5 of the project.

INTEGRATION OF BATTERIES FOR THE HIGH PENETRATION OF PVSYSTEMS IN THE GRID

The incorporation of accumulation in PV generation plants allows further mitigating power fluctuations, allowing the power injection to be shifted from its generation by the PV modules, the dispatchability of PV plants, and adding other functionalities like ancillary services. PVCROPS has achieved to integrate batteries into PV systems (Li-ion and Vanadium Redox batteries), developing hardware and software to manage the energy flows, and to optimize control strategies.

Toolbox for the design and simulation of the energy management strategies in PV plants - An energy management strategy to maximize the economic output of a PV plant - Equipment for the energy management and storage control in PV plants

More and more there exists codes related with the maximum allowable ramp rate that a PV plant can inject into the grid. Without fulfilling these codes, a PV plant cannot be connected to the grid. PVCROPS has developed an effective method (the “Worst fluctuation model”) to calculate, for any PV plant size and maximum allowable ramp-rate (rMAX), the maximum power and the minimum energy storage requirements alike.
Once Energy Storage System (ESS) requirements to mitigate the possible worst fluctuation event have been given, the size of the storage system will depend on the control strategy implemented. The strategies developed, aiming at reducing the required size of the battery, are the following:
• Ramp rate control
• Moving average
• Step control
• Power ramp-rate control using the PV inverter
• Power ram-rate control based on the PV power plant model.

The details of these energy management strategies, that are currently being exploited by ACCIONA, have been given in deliverable 6.5.

It makes sense also to think about alternative uses of the ESS in order to maximize its economic output. Especially when it is necessary to install an ESS oversized in energy. One of these alternatives is the generation shifting. PVCROPS has analyzed the economic viability of the generation shifting scenario as is expected to be one of the first widespread applications of ESS to generate additional revenues in a PV project.

A toolbox that integrates all the strategies presented has been developed. Essentially, the toolbox calculates, for any PV plant size and maximum allowable ramp-rate, and the six energy management strategies presented, the maximum power and the minimum energy storage requirements. In addition, depending on the chosen strategy, the toolbox gives the main results regarding to the energy storage system (ESS) use. The toolbox provides a pdf document with these results.

In addition to the development of these new strategies, PVCROPS has validated them through two demonstrators. A lithium-ion battery energy storage installed at Tudela PV plant and a VRB installed at Évora (Portugal) university. The validation of the strategies under real test conditions let us to learn about the problems that have arisen during the validation process carried out from January 2015 to October 2015 in both demonstrators, Li-ion and VRB. After solving all the issues mainly due to delays in the communication systems, the main results of the validation are shown in the following table. This table summarizes the results in % of fulfillment of the maximum ramp-rate variation imposed of 10%/min during the validation process. It can be seen how all the strategies achieve to fulfill the imposed requirement with a very high rate of compliment.

STRATEGY % FULLFILMENT
Ramp-rate control Lithium ion battery: 96.1
VRB: 96.6
Step-rate control Lithium ion battery: 92.1
Moving-average control Lithium ion battery: 92.9
Ramp-rate control using the PV plant model Lithium ion battery: 100
Table: Fulfillment (%) achieved for each strategy during the validation process.

Finally, the required hardware to implement the energy management strategies and to control accordingly the PV plant, has been developed, mainly the DC-DC charger required to control the battery power and the control unit where the energy management strategies are implemented.

Toolbox for the design and simulation if the energy management strategies in BIPV - Equipment for the energy management and storage control in BIPV

Building-integrated photovoltaic installations have a good share in the destabilisation of the grid in the scenario where the photovoltaic production is not regulated. In residential areas, with a high ratio roof surface pro energy consumer, the photovoltaic local penetration can become a critical issue. At certain moments at day, most installations could be producing much more power than they consume thus using great part of the local grid capacity in the opposite way. A grid designed to distribute energy towards the consumers is calculated to offset the voltage drop due to losses in the grid. So, a significant amount of power travelling in the reverse way will cause a dangerous situation. On the other hand, we have the opportunity to turn such massive menace into a kind of distributed control for the grid stabilization by giving every little installation the ability to manage the power it produces and consumes. PVCROPS has addressed the key spots in order to transform the initial situation aiming the best integration of photovoltaic technology into the grid.

Two demonstrators were built to have a field on which to achieve this objective. They were developed on Heredade da Mitra, in Évora, at the Universidade de Évora facilities: one based on a Vanadium Redox Flow battery (VRB) and another one on a Li-ion battery.

Specific hardware has been developed: many improvements and modifications were performed on the converters’ hardware making it much more reliable and robust and better adapted to the requirements of the operation into a BIPV system. One of these main requirements is that the power output must react as fast as possible in order to avoid grid distortions. Furthermore, the validation process has led to the re-design and re-engineering for the next generation of VRFB system which is now in production.

Another essential piece of the system is the controller. A new device has been designed and developed for the task of making decisions for controlling the BIPV system. These decisions respond to the strategy running which will be one or other depending on the requirements of the installation. Different strategies have been developed and tested as explained later on this text. EMS Manager is the business name for this controller device whose industrialisation and commercialisation is already fully completed.

A tool for the monitoring of BIPV systems has also been developed. Its name is EMS Tools and it allows to graphically configure the layout of the BIPV system and the strategy parameters, to pause and resume the control, to reboot the system, to view real-time power flows of every device, to draw graphs of the evolution of every relevant measurement, and to export data to other formats such as MS Excel®. It is a useful tool for quick and simply tracing the installation’s operation.

All this equipment is being exploited by INGETEAM.

A key point in the project is the development of the strategies supported by the products described above. Different strategies intend to resolve the problems arisen from different scenarios. In some situations the challenge is to sustain a as smoothest as possible fluctuation in the power exchanged with the grid, other times it is required to have the minimum consumption from the public grid, or in other cases it will be needed to ensure a consumption below a given value.

A simple scheme in which:
- every time the generation is greater than the consumption, the surplus is stored into the battery until it is full, and
- when the loads consume more than the photovoltaic can produce, the shortfall is supplied by the storage system until it is depleted, will result in the maximum possible exploitation of the renewable resource or, in other words, in the minimum possible consumption from the outside and so the maximal ratio for own consumption. Such procedure for managing the energy flows represents a strategy that, inside the PVCROPS, we call Self-Consumption Optimisation.

Another strategy studied and developed is the Constant Power Control. It is intended to maintain the storage system at a specific state of charge so that it can deliver or take up the needed amount of power that keeps the interaction with the grid almost invariable.

When the user wishes to reduce the power purchased from the utility, the Peak Shaving is the strategy to run. This algorithm makes the storage system deliver the amount of power needed to keep the grid power below a given value. Whenever possible, power is taken up from the grid to ensure the enough stored energy that avoids shortages at any time.

Together with the previous strategies, variations including manageable loads were analysed and implemented. Usual loads suiting best these strategies are water heaters and A/C systems.

A toolbox was created for automatic sizing of BIPV systems running different strategies and battery technologies. It allows fine-tuning the control parameters of the energy management strategy. This toolbox processes a simulation of one year of operation of the system and ends up with the creation of a comprehensive report showing all the simulation results.

Finally, the most interesting and representative Energy Management Strategies were tested in the demonstrators, which are the Self-Consumption Optimization and the Constant Power Control. The validation test campaign took place from February 2015 till October 2015.

As an example, the following table summarizes the results of the validation of one of the strategies:

Merit factors Lithium-ion VRFB
Maximum Power from grid (P+) 0.011 kW (98 %) 0.049 kW (99 %)
Maximum Power to grid (P-) 1.266 kW (6 %) 2.506 kW (42 %)
Maximum Power Derivative (MPD) 699 W/h (27 %) 1772 W/h (30 %)
Average Power Derivative (APD) 31 W/h (52 %) 99 W/h (63 %)
Table. Merit factors for both demonstrators performing the SCO strategy. Absolute values, and reduction relative to the net power.

Potential Impact:
IMPACT OF THE RESULTS

The PV CROPS results include toolboxes, databases, technology developments and technical documents that are related to the expected impacts of the topic and the KPIs of SEII – SET PLAN 2020:
• Reduction of LCoE (€/kWh)
• Increase grid integration
• Increase of Performance Ratio (PR)
• Increase of the ratio self-consumption/grid export
• Increase of performance stability

The commitment of PVCROPS was based on the quantification of the impact for the two main sectors (PV plants and BIPV systems) and for the three main KPIs: LCoE reduction, Performance Ratio increase and grid integration increase:
- 30% LCoE reduction: LCoE of PV plants in Southern Europe to be reduced to about 0.09 €/kWh by the end of 2015 (the corresponding KPI fixed by the EU is to attain 0.07 €/kWh by 2020, therefore this proposal is in line with it), and LCoE of BIPV in mid-latitude Europe to be reduced to about 0.20 €/kWh by the end of 2015 (the corresponding KPI fixed by the EU is to attain 0.14 €/kWh by 2020 for BIPV).
- 9% Performance Ratio increase: Performance Ratios to be achieved for PV plants in Southern Europe of about 81% by the end of 2015 (this corresponds to annual energy losses in the PV system of some 19%) and Performance Ratios to be achieved for BIPV in mid-latitude Europe of about 86% by the end of 2015. .
- Reduction of PV power fluctuation to less than 10% in 10 min to allow 30% of PV penetration in the grid: achievement of the SET Plan goal of 12% of PV in the EU grid and PV penetration of 30% in the sunniest European regions.

In order to evaluate the impact of our solutions, we have analysed the state of the art of more than 30,000 current PV systems in Europe regarding their performance and their potential improvement.

We have concluded that the state of the art of the PV systems in Europe corresponds to a range of yearly integrated PR values between 0.6 and 0.9, with average values typically between 0.75 and 0.8. This represents a difference of some 30% between the best and the worst performers. Ideally, the PV sector should aim at reaching PR values around 0.9 for most of the PV systems to be installed in the future. In practice, the highest PR values might not lead to the lowest Levelized Cost of Energy (LCoE) for the PV systems, because the best performers could be more expensive than the other systems. If the ultimate goal is to minimize the LCoE of the PV systems in Europe, the optimum could therefore correspond to PR values that are somewhat lower than 0.9. Nevertheless, we have not observed a clear and systematic correlation between the performance of the PV system components and their overall performance, which leads to think that the optimum could correspond to PR values that are in any case higher than 0.84, and that many of the low PR values are not justified by a lower cost of installation. So, the tools developed in the framework of PVCROPS related with an optimised design and automatic detection of performance failures are therefore a very promising option towards a leap in overall performance, which could lead to an average value of PR over 0.84, representing an improvement in performance around 10%, and a corresponding reduction in LCoE of the same order of magnitude.

In order to establish a relationship between improvement of PR and our tools, we have compared the Weibull fit to all the yearly PR data, and a gaussian fit to the central part of the distribution. The PR values for which the Weibull distribution shows higher frequencies than the gaussian distribution (which happens for PR values below 0.72) probably indicates the presence of failures on the PV systems. The PR values in the Gaussian distribution under the P90 =83.8% indicate that performance improvement due to a better design and selection of PV components could be achieved.

So, as the mean PR value is now 76.5% and that could be improved to 83.8 without any increase of cost, the increase of PR would be 9.94%: 3.13% due to our tools for automatic failure detection and 6.81% due to a better design and components selection. This last impact is covered by the set technical specifications, quality control procedures and simulation tool.

Furthermore, the automatic detection of performance failures can reduce the maintenance costs and therefore a reduction of LCoE between 5% and 10%, for PV plants and BIPV respectively; the reduction of the uncertainty in the estimation of PV plant production reduces the risk and improve the conditions of financing, estimating a reduction of LCoE of 2%; the technical specifications and the manual of good practices will allow reducing the installation costs with an impact of 3% in the installation costs; and the energy management strategies, including demand side management, will allow offering ancillary services and maximizing the PV self-consumption, reducing the LCoE at least 10%. Finally, the prediction tools will allow reducing the whole cost of the electrical system as the spinning reserves to back the PV systems can be reduced. Although it is difficult to evaluate this impact, it could be translated into a reduction of LCoE of at least another 10%. This represents a reduction of LCoE of more than 30%, without taking into account the reduction due to the evolution of technology, such as the reduction of the price of PV modules, that is out of the scope of PVCROPS.

Regarding the high penetration of PV systems into the EU electrical networks, the tools for the prediction of PV power and the ones for the integration of batteries, that limit the PV power ramp rates to the thresholds established by the National regulations, have covered all of the technical challenges identified by EPIA and the project PVGRID for the so-called “paradigm shift scenario”. This way, from a technical point of view, all of the technical barriers to achieve 12% of EU electricity demand by 2020 and 25% in 2030 have been removed, that means more than 30% of PV penetration in Southern EU countries. So, if this goal is achieved in 2030 or before will depend on political barriers but not on technical ones as PVCROPS solutions allow to solve them.

PV CROPS has also contributed to broader societal objectives for Europe, from three different perspectives:
• Energy policy of Europe, reducing the overall cost of electricity generation, PV energy payback time and CO2 emissions, as a consequence of the reduction of the LCoE and the high penetration of PV systems into the EU electrical networks. As PVCROPS allows 30% of PV penetration, devoted to substitute conventional energies, it has a similar impact in the reduction of CO2 emissions. Furthermore, the analysis of the use of second-hand batteries in PV installations opens the door to a second life for this equipment, reducing the associated environmental problem.
• Competitiveness of the European PV industry in the world, through the maximum exploitation of local sources of value creation, employment, and geostrategic advantages that constitute strong barriers against the entrance of outside competitors. PVCROPS results allow the leadership of EU companies in the sector of PV systems, even though the manufacturing of PV modules is in other regions of the world: by achieving the performance gains attainable in the PV plants and BIPV areas through PV CROPS, the EU PV industry will gain technological leadership positions in the global PV industry. The huge database of 1-second PV production, the developed energy management strategies for PV systems with batteries, the prediction tools and the automatic detection of performance failures give to the companies of the consortium, and to other EU companies interested in these results, a leading technological position in the market, with an important potential impact in the employment. In fact, companies from the sector of batteries and companies promoting PV plants and BIPV, have contacted the consortium to exploit the results.
• Strengthening the cooperation in energy policy between EU and Northern Africa through the inclusion of a Moroccan partner in the consortium ONE, country where an exponential growth of PV market to export PV electricity to Europe is foreseen, supposing an excellent opportunity for the EU companies.

Finally, it is worth to underline that PVCROPS has implemented two spin-offs that are working with SMEs, one of the most important sources of employment in Europe. The two spin-offs are:
- WebPV: focused on the exploitation of the tools for the automatic detection of performance failures.
- SunWings: focused on offering advanced services of quality control of PV systems and on the commercialization of the testing kits.

DISSEMINATION ACTIVITIES AND EXPLOITATION OF THE RESULTS

The steps to reach the desired impact are those covering the distance between the RTD and the adoption of the solutions by the PV community through dissemination and, overall, commercial exploitation.

After the RTD work to develop the solutions, a set of documents and toolboxes have been delivered online and free to use as a way for a more extensive knowledge of the marketable solutions. These open results have been the following:
R1: Technical specification for grid-connected PV systems ready to include in contractual frameworks
R2: Manual of good and bad practices to improve the quality and reduce the cost of PV systems
R3: Toolbox for robust design of PV systems: SISIFO (www.sisifo.info )
R4: Built-in learning tools destined for professionals, researchers and students in solar energy (integrated in SISIFO tool).
R9: Document displaying the PV generation that can be integrated into the European grid
R19: Document detailing the quality control procedures to be included in contractual agreements.

Although they were not foreseen as open-source tools, finally we have developed two different versions of the tools for the prediction of PV power, one open-source and free and another one for commercialization. In a similar way, we have produced two different versions of the toolbox for providing real-time irradiance of tilted surfaces, one for free and the other one for commercialization:
R5 open: Toolbox predicting PV energy production on an hourly basis with non-parametric methods
R7 open: Toolbox predicting PV power fluctuations with non-parametric models
R17 open: Website with real-time irradiance data on tilted surfaces: SOWEDA (www.soweda.net )

The rest of the results are devoted to commercialization and PV CROPS has developed detailed exploitation plans for the thirteen marketable results.

The main part of dissemination has been oriented to reaching the targets and potential clients identified in the exploitation plans. This strategy has been denominated in the project as “Common Exploitation and Dissemination Platform”. Means such as website, electronic magazines, trade journals, technical visits to demonstrators and seminars and workshops have been used. In fact, the spin-offs and the industrial partners of PVCROPS are already exploiting the results:
ACCIONA: R5, R7,R11, R12, R14, R18
INGETEAM: R10, R13
WebPV: R15, R17
SunWings: R18, R19

Common Exploitation and Dissemination Platform

The common exploitation and dissemination platform has contributed to orientate a great part of dissemination to the exploitation of the commercial results. PVCROPS has orientated the dissemination into two main axes: diffusion of the results to the PV community and key stakeholders and the commercial exploitation of the results. The dissemination part related with the exploitation plans and the commercialization of results has been in charge of DIT. The dissemination to the PV community and key stakeholders, whose first goal is to disseminate the progress beyond the state of the art, has been led by APERE.

The following actions have been carried out in the framework of the Common Exploitation and Dissemination Platform:

a) Seminars and Workshops

A total of two seminars oriented to the exploitation of the results, three seminars devoted to the dissemination of the results (to the Transmission System Operators, TSOs, Distribution System Operators, DSOs, and to the PV community) and three specific workshops targeting TSOs-DSOs and the PV community.

Seminars oriented to the exploitation of results: two half day exploitation seminars were organized. The first seminar was held at InterSolar, Munich in June 2014, the conference and exhibition of reference of the EU PV industry. Its title was “Developing innovative solutions for improving PV penetration within electrical networks”. The second exploitation seminar was held at Solar Energy UK, Birmingham in October 2015, a conference and exhibition event emerging in Europe thanks to the growing PV market in UK where the most important EU companies are present. Its title was “PVCROPS at UK Solar: Commercial opportunities in Solar PV”.

Seminars devoted to the dissemination of the results: three deep seminars were organized. The first one took place at ELIA headquarters (2nd April 2014), the Belgian TSO and was oriented to disseminate PVCROPS results among the TSOs and DSOs. It was entitled “PhotoVoltaic Cost r€duction, Reliability, Operational performance, Prediction and Simulation”. The second one was organized as a parallel event in the European Photovoltaic Solar Energy Conference 2014 (EUPVSEC 2014) in Amsterdam, Holland (24th September 2014). The title of the half-day event was “Grid-connected PV systems: field testing, performance monitoring, and energy storage”. And the third one was carried out again as a parallel event in the EUPVSEC 2015 in Hamburg, Germany (17th September 2015). Its title was: “PVCROPS: Novel solutions for a high PV penetration in EU electrical networks with lower LCoE”.

Specific workshops: three different workshops were organized. The first one took place at Innogrid2020+ 2015, the Transmission and Distribution R&D Conference (31st March – 1st April 2015) in Brussels, Belgium. It was devoted to the TSOs and DSOs companies and its title was “Management of PV plant fluctuations by the use of forecast and a properly battery sizing: some results of PVCROPS project”. The second one was organized at Smartgreens 2015, also oriented to TSOs and DSOs 21st May 2015), in Lisbon, Portugal. Its title was “Design of batteries and monitoring for BIPV systems”. And the third one was organized in Brussels, Belgium (6th October 2015), in the framework of a joint seminar with the project PerformancePlus, in order to maximize the impact in the PV community. Its title was “PV performance and Reliability: Results and Recommendations”.

b) Communication through internet: web, electronic magazines, trade journals and PV forums

After targeting more than twenty trade journals, DIT published ten trade journal articles covering the main results of PVCROPS. The social media communications has been carried out trough news in PVCROPS facebook page. Furthermore, DIT has continued to promote the commercial deliverables via a Twitter and Linkedln accounts. News in the main PV forums in Europe have been also disseminated: www.ForumPhotovoltaique.fr (France), www.SolarWeb.net (Spain), and www.Apere.org (Belgium).
The specific section of PVCROPS website (www.pvcrops.eu ) “Exploitation and Dissemination seminar” has been the main tool to disseminate the project results, both the deliverables for the general public and the description of the marketable results for the PV industry.

c) Publications in International scientific journals and in International conferences

In the one hand, 15 papers in international scientific journals of the highest impact factor have been published. Furthermore, five more papers are now being finalized to be sent to other international journals.

In the other hand, 42 publications and presentations in international conferences have been carried out. Most of the publications has been in the different EUPVSEC from 2013 to 2015, assuring the widest dissemination of the project results among the PV community.

d) Technical visits to demonstrators

PVCROPS has implemented three demonstrators: one in Tudela PV plant (Spain) showing the energy management strategies for PV plants in a 100kW Li-ion demonstrator; one in Evora to show the energy management strategies for BIPV systems in a 10kW Li-ion demonstrator; and the third one also in Evora (Portugal) to show the energy management strategies for PV plants and BIPV systems in a VRB demonstrator of 10kW.

The consortium managed to organise three technical visits to demonstrators to disseminate the PVCROPS results in a practical way.

First technical visit: it was organized at Evora University, Portugal, the 7th October 2014. Its title was “Photovoltaics: Engineering and grid interfaces (centralized production, BIPV, other applications) – Technical workshop.

Second technical visit: it was took place at Tudela (Spain) the 16th April 2015 under the title “Photovoltaic plant and energy storage: Technical workshop and demonstrator”.

Third technical visit: linked to the audience of the Smartgreen conference in Libon, a visit from Lisbon to Evora was organized to show the two demonstrators with their corresponding energy management strategies. It was the 22th May 2015 with the title “Energy management: self-consumption and storage control in BIPV: Technical workshop”.

e) Patents

A total of eight patents were filed over the course of the PVCROPS project by work package leaders with the support of DIT Hothouse. Moreover, two open GNU-GPL licenses have been also filed.

f) Exploitation plans

A total of eight documents have been produced containing exploitation plans for the 13 commercial products. Some of the results have links between them, so a single document has been prepared for them. The exploitation plans support the commercialization of the marketable results, providing key information for potential investors.

The content of the exploitation plans is the following:
- Executive summary
- Description of the result
- Market opportunities
- Market size
- Competitive landscape
- Financial analysis

List of Websites:
Website of PVCROPS project: www.pvcrops.eu
Contact of the Project Coordinator:
Luis Narvarte
email: navarte@ies-def.upm.es

Related information

Contact

Prieto, Roberto (Vice-rector for Research)
Tel.: +34 913366048
Fax: +34 913365974
E-mail
Record Number: 182517 / Last updated on: 2016-05-12
Information source: SESAM