Skip to main content

Advanced NAno-Structured TApeS for electrotechnical high power Insulating Applications

Final Report Summary - ANASTASIA (Advanced NAno-Structured TApeS for electrotechnical high power Insulating Applications)


Executive Summary:

The objective is to develop radically innovative electrical insulating tapes and process to improve the energy conversion efficiency of electrotechnical systems. It mainly addresses the electric power generation issue.Today the energy conversion efficiency of generators is restricted by thermal and electrical strength limitations due to the electrical insulator tapes themselves. The concepts of these multifunctional tapes are far behind the electrical insulating state of the art. The project aims to develop a new process chain leading to a drastic improvement of insulating tape structure homogeneity. The today’s limitations of tape come from its heterogeneous multilayer structure bringing together very different materials like glass fibre fabric, mica flakes and polymers. Enabling this homogenisation requires higher performance materials, which will be obtained by adjunct of inorganic nanofillers according to two proposed approaches: nanodielectrics polymer or inorganic polymers (sol-gel). This will lead to a more robust process chain with a better productivity (+50%) and an insulating tape with enhanced performances like a higher field strength (+40%), a better thermal conduction (+60%). At the end, a much thinner tape (-30%) enabling the design of more compact generators is expected. This project can strongly impact the energy production field. It will also affect other very large markets like the industrial motor field using similar insulation tapes.

The work realized during theses 3years has been cut into 3 routes:

The first one is the top down approach: This is approach in the sense that the nanofillers are generally prepared from grinding processing of raw material or synthesised, then dispersed in the epoxy resin as adjunct. As previously described, the current study has shown that adding nanoclays (e.g. montmorillonite) in a micro quartz /epoxy composite system can lead to a significant improvement of the resistance of polymer surface to corona discharge, an improvement in AC electrical breakdown performance and promotion of the decay of space charge in the material. This multi-dimensional composite concept will be adopted in this project as a conservative low risk approach.

- Route1A will address the impregnation resin. The nanoclays/microcomposite system will be improved by advanced dispersion techniques of dedicated nanofillers (nanowires and nanotubes)
- Route1B will address the dielectric barrier optimisation by introducing nanoflakes.

The second route is the bottom up approach for advanced nanostructuration of epoxy. An important conclusion of previous research work on nanodielectrics was that, in the case where nanoparticles are added, it is generally found that the pre-treatment of these particles by using a proper coupling agent plays a determining role in the property outcome. It means that to warranty the expected enhanced electrical property, the characteristic of the interface (nanoparticle/matrix) is certainly a major determining parameter to be closely controlled. This is this conclusion that precisely motivates the present route.

In Route 2, precursors for the in-situ production of nanoparticles will be used, instead of preformed ones. This bottom-up approach is particularly appealing due to the elimination of the intermediate process of nanofiller dispersion, in which the (often) strong cohesive forces between nanoparticles are to be overwhelmed in order to obtain proper nanostructuration.

The third route is the quasiinorganic approach. In this approach, highly advanced solutions are envisaged, exploiting the expertise acquired and best results obtained in Routes 1 and 2. Two main directions will be explored:

a) Quasi-inorganic insulating material, with extreme thermal resistance and dielectric properties.
b) Multifunctional layered tapes, through 2D structuration for the strict control of properties along tape thickness

All results are presented in the file attachment (final report and end user presentation power point)

Project Context and Objectives:

Energy saving is a major societal issue that concerns a large field of technical applications, in particular electric power generation. The basic statement motivating the project is that today the energy conversion efficiency of generators is restricted by thermal as well as electrical strength limitations due, in particular, to the electrical insulation itself. The key feature of the generator field is that it deals with very long lifetime systems (40 years!).

Consequently, it involves very traditional insulating materials and well established tape structures whose electrical properties and concepts are in fact far behind the current insulating material state of the art. As a result, high voltage insulating tapes are rather thick, poor thermal conductors and require energy consuming cooling systems.

Today, high voltage insulating tape technology presents a very significant scope for progress that must be absolutely exploited to increase the generator conversion efficiency.

That is why the ANASTASIA project intends to introduce into the generator industry advanced insulating tapes based on nanostructured material scenarios to address the energy saving concern. Even for a small efficiency increase (+0,2%) the expected benefit represents a very substantial energy saving.

At the European scale, a +0.2% gain in generator conversion efficiency could save the equivalent of one nuclear power plant of 1000 MW (1.5 billions €) or nearly 10 fossil fuel power plants and related reduction in CO2. The main characteristic of the conventional insulating tape used today is the strong heterogeneity of its structure since there is a glass fibre carrier tape co-laminated with a mica paper (dielectric barrier) and impregnated by polymer resin.

This multilayer tape structure by multiplying the number of interfaces (glass-polymer, mica-polymer and mica-glass) is conceptually in contradiction with the targeted electrical and thermal properties. Indeed, each interface behaves as an additional thermal barrier and/or a potential source of high voltage partial discharges. In practice, to overcome these drawbacks a large quantity of mica is required (typically, 1 mm mica paper thick) to provide an efficient partial discharge protection thanks to the “brick wall” effect.

This results in the tape exhibiting very poor mechanical stability and a high sensitivity to environmental factors, notably water. To offer new perspectives of insulating tape improvement, a technological breakthrough is necessary. The ANASTASIA tape concept needs to develop a new process chain leading to a drastic improvement of tape structure homogeneity. To facilitate this homogenisation higher performance materials are required, which will be obtained by the incorporation (or in-situ synthesis) of inorganic nanofillers.

This will lead to:

- A robust and more homogeneous process flow with a better productivity (+50%).
- A radical change in the filler to polymer interface providing a multifunctional insulating tape with enhanced electrical, thermal and mechanical performances

The majors objectives of Anastasia are:

High power generators manufacturers consider today that the upgrading potential of the air cooled systems can be realized by a combination of insulators having, a higher dielectric field strength, thermal conductivity and thermal resistance together with a smaller thickness allowing a better filling ratio. Alstom considers that the target values of the following table could open the way to a new class of generators offering up to 0.2 % of additional yield.

These targets will be the project quantitative objectives. To meet these objectives, the ANASTASIA project will develop three complementary tape nano-structuration approaches:

- A polymer route based on in situ growth of dielectric nanoparticles
- A sol-gel inorganic based matrix for a quasi inorganic concept.
- A novel mica paper structure involving dielectric nano-flakes.

Once validated, these new tape architectures will be implemented on a prototype tape manufacturing equipment to produce a tape length suitable to wrap a stator bar. By this way, both tape production process and tape implementation onto a rotating machine will be validated.

Finally and for the specific case of quasi inorganic matrix, a specific development will be carried out to develop optical methods to perform in line control of nanofillers homogeneity during the tape production process.

The improvements expected for the tape are the following:

- With mica nanostructuration, a reduction of potential leakage current paths (see figure 2) and partial discharge risks leading to higher field strengths (improvement: up to +40%)
- Better thermal conduction (up to +60%) in line with a reduced tape thickness (-30%). This will facilitate the design of more compact generators.
- A new class of dielectric material of high thermal resistance fitting with Class H requirements (180 °C working temperature)

Scientific approach of ANASTASIA:

As previously described, the benefit brought by the nanostructuration of insulating material has no longer to be demonstrated. However, to better highlight the relevance of this concept, it is necessary to go back to a more fundamental comprehension level:

Basically, any insulating material, organic or inorganic, exhibits a certain amount of unavoidable imperfections which are responsible for the presence of electrical charges inside the material (so called “space charges”). These imperfections are chemical impurities, electronic defects (doping elements), physical interfaces, voids, etc…..

Their concentration in an insulator is dependent on its manufacturing history, namely its preparation technique, thermal post-treatment, electrical pre-conditioning, state of mechanical stress and so on. In operation, when the insulator experiences electrical and temperature stresses, the space charge concentration within the material tends to increase with time.

This accumulation of space charges is generally non-uniform, creating local electrical field enhancements. These local fields will actively contribute to the occurrence of partial discharges inside the insulator. In the long term, this can lead to insulator degradation and ultimate failure.

In practice, to avoid the occurrence of failure in a power generator, the system operates at relatively low field strength, 2-3kV/mm, far lower than the potential electrical breakdown strength (>100 kV/mm). Increasing the design field strength experienced by the insulation in a rotating machine, as proposed in the ANASTASIA project, requires insulating tapes to operate at higher fields without increasing space charge accumulation within the dielectric.

In that sense, the nanostructuration of insulating tape is a particularly relevant and promising solution.

Today, the nanodielectrics concept is mature enough to be brought out of the research laboratories provided that careful material developments are done to control the dielectric nanostructuration in view of a better management of space charge distribution within the tape.

According to the partners already involved in the nanodielectrics field (partners 3, 5, 7 and 8), introducing nanoparticles in an insulator precisely reduces the space charge accumulation by promoting electrical charge evacuation.

A number of possible explanations can be proposed to account for this. First, Lewis and all suggested that the presence of nanoparticles may result in the formation of diffuse regions of interfacial charges, which assist charge transport through the material. Tanaka and all has also suggested that it is the nature of interfaces that are responsible but, in this case, the mechanism involves the proliferation of localised electronic interface states, coupled with an increase in the concentration of mobile ionic species.

Finally, the evacuation process could simply be ascribed to electrical field enhancement in the proximity of nanoparticles, which resembles that which is observed at a tip surface. Locally, the electric field at the nanoparticle/polymer interface is increased, enhancing charge transport through the system and, consequently, evacuation of charge and/or bipolar recombination.

However, all three mechanisms rely critically on the form of the nanofiller. In the Lewis and Tanaka models, this will influence percolation paths; field enhancement depends on both the radius of curvature of the surface and also on the change in dielectric constant at the interface.

As a result, whatever the dominant underlying cause, space charge decay depends strongly upon the nanoparticle size, shape and nature. Furthermore, it is obvious that the nanoparticle concentration and dispersion uniformity inside the polymer matrix will also be key parameter.

This is the reason why, in ANASTASIA project, the material/tape development is focused on the control of these parameters. For that purpose, different ways of nanostructuration are proposed.

ANASTASIA routes for the development of a new insulating tape:

Completely suppressing insulator imperfections is not realistic and would require a total change of technology, which is a challenge going far beyond the scope of this project. The strategy of ANASTASIA is to modify the tape in order to mitigate the harmful electrical effects of these imperfections: this role is devoted to the nanofillers adjunct provided that this adjunct is well controlled.

With respect to the electrical insulating tape technology, the strategy of ANASTASIA is to gradually introduce innovation by developing the three following complementary routes:

- A top-down approach (Route 1) consisting in replacing conventional fillers by well controlled in size and aspect ratio nanofillers dispersed by an innovative technology to maximize the uniformity and dielectric performances of both impregnation resin and dielectric barrier.
- A bottom-up approach (Route 2) targeting in-situ generation of dielectric nanoparticles by introducing inorganic precursors into the epoxy monomer to change radically the nanofiller/matrix interface.
- A quasi-inorganic approach (Route 3) involving sol-gel inorganic/organic hybrid matrix in place of the conventional resin to reach outstanding thermal and dielectric performances.

The low risk Route 1 will generate short or medium term innovations but will serve mainly to assess the dielectric performances/nanofiller structure relationship by a scientific approach.

Developed on the basis of Route 1 results, the medium risk Route 2 and high risk Route 3 target outstanding breakthroughs and have the potential to support long term innovation.

Project Results:

The objectives of the project are as follows:

- Implement a new technology with respect to the electrical insulating tape technology, the strategy of ANASTASIA is to gradually introduce innovation by developing the following three complementary routes:
- The bottom-up approach (Route 2) aims to generate dielectric nanoparticles in-situ by introducing inorganic precursors into the epoxy monomer to change radically the nanofiller/matrix interface. (WP2)
- The quasi-inorganic approach (Route 3) involving the synthesis of a sol-gel inorganic/organic hybrid matrix in place of the conventional resin to give outstanding thermal and dielectric performance. (WP2)

Developed from the Route 1 results, the medium risk Route 2 and high risk Route 3 will target major breakthroughs in technology with the potential to initiate long term innovation. The various strategies will be compared.

• Characterization resin and tape test samples, perform electrical, thermal and mechanical characterization and evaluate the benefits of nanostructuring with respect to key properties. Control of size and surface density particles. (WP3).
• Define and develop the protocols for appropriate electrical characterization. (WP3)
• Consider basics ageing effects, to ensure the stability of samples developed from optimal resin/nanofiller systems. (WP3).
• Demonstrate the technological viability of the new tape manufacturing processes. (WP4)
• Establish the tape process development and define the material scenario for the new tape, which will be basis for the final material selection. (WP4)
• Make a demonstrator development (WP4).
• Demonstration activities that will be performed at two points within the supply chain: the tape manufacturing step; and fabrication of a generator conductive bar. The impact of these innovations on generator design will be analysed. (WP5)
• Ensure appropriate application of the knowledge that has been developed during the project, manage IPR, and broadly communicated appropriate innovations and their benefits. (WP6)

All these objectives are linked together and will enable optimized demonstrators (designs, structure and characterization) to be produced and mechanically and electrically characterized within the final project period.

1- Bottom-up approach: Synthesis of nanofillers - Actions and results:

The bottom-up approach concerns the in-situ synthesis of nanofillers in the epoxy resin. The method which has been developed by the CEA/LCSN is novel and innovative.

Indeed, the formation of interpenetrate network between organic polymer, silica-epoxy hybrid polymer and silica grid through a sol-gel process by using coupling agent has been studied in the past. Also the introductions of pre-hydrolyzed silica alkoxide in epoxy matrix to obtain nanostructure composite after curing have been screening. However, no relevant studies have showed the investigation on the growth of silica nanoparticles in resins which are suitable for electrical insulation systems. Although the sol-gel process based on the hydrolysis and condensation of silica precursor is well known to synthesise silica particles, no literature has been describing this method as suitable for mass production as well as for dielectrically performing materials.

The first goal of this work is to draw up and apply to industrial application an experimental protocol to synthesis particles inside an organic resin. From experimental data, a tailor-made resin is built. One of the parameter of this synthesis is the independent reactivity of the components, which will form particles, and the resin. The second limiting parameter was the final viscosity of the doped resin.

The second goal of this part of the project is to assess the efficiency of an “in-situ” doped resin compared to a classical introduction by mixing particles into resin.

1.1 Materials and methods:

Sol-gel method introducing nanoparticles into Damisol 3309 (polyesterimide resin, Von Roll) and Harzkomb EPE19A (polyester/epoxy resin, Walter Mäder AG), has been successfully achieved. The reactions were designed to be solvent-free and the nanoparticles were directly synthesized in the polymer matrix by reacting tetraethoxysilane (TEOS, d = 0.934 M = 208.33 gmol-1, Sigma Aldrich) using basic conditions at room temperature under inert gas.

1.2 Experimental set up

• First set up of up-scaling:

Experiments were conducted in a 10 L double-glass cylindrical reactor where a thermofluid is circulated and linked to a constant temperature bath. The reactor content is homogenized with a high shear mechanical stirrer placed at the top of the reactor.

• Second set-up of up-scaling:

The second experimental set up used to obtain a solvent-free synthesis of nanoparticles directly in the polymer matrix, was the same set-up as the previous one but completed with a lid and controlled flow of inert gas. The lid has three openings which support three devices: an inert gas injector and temperature sensor, a condenser and a Soxhlet and a motor. The set of Soxhlet-condenser is use to receive the by-products and the water realised during the synthesis. A valve at the bottom of the reactor let us the possibility to take a sample of the batch during the synthesis.

1.3 Experimental procedure

The stirred reactor is filled little by little with the different components of the reaction, in this order: water, ammonia, resin and silicium’s alkoxide. A period of 20–120 min is needed between each addition to stabilize the emulsion, the starting time (t = 0 min) of the reactions corresponds to the TEOS addition. The viscosity is increasing during the hydrolysis and condensation of the silicium’s alkoxide and the color of the batch became white. When the reaction is completed those both parameters became again in the initial range. For a complete reaction, one day of synthesis is needed.

With the second set-up, the first step of the experiments was consisting in the preparation of an oxygen free reactive and temperate at 25°C. So, all the chemicals were degassed under inert gas for 1h before using them.

1.4 Experimental characterization

The progress of the reaction were followed by taking sample during the experiments and by measuring the viscosity at 23°C with Brookfield viscometer DVDII Pro and the polycondensation of TEOS by Raman, FTIR spectroscopies.

The morphology of the silica nanoparticles synthesized in the Damisol and EPE19 matrix has been analyzed by Transmission Electron Microscopy (TEM) and Scanning Electron Microscopy (SEM) studies. For this, after the hydrolysis and in-situ condensation of alkoxisilane, the resulting polymeric solutions were dissolved in solvent and the silica nanoparticles produced in-situ were sedimented by centrifugation.

The thermal degradation of the cured samples was followed by thermogravimetry using an TGA instrument (Setaram 92 ®, TA instrument®) at a heating rate of 10 °C min−1, from room temperature up to 800 °C, in dynamic nitrogen flow and then air atmosphere flow. Also their thermal diffusivity was easily measured by flash method using LFA 447 NanoFlash® at 23°C.

The electrical properties of the novel nanostructured materials were determined by dielectric spectroscopy at 50Hz and breakdown spectroscopy at room temperature. Space charge measurements were done under 5kV and 10 kV pooling first at room temperature then at 60°C.

1.5 Results and discussion

In order to obtain the growth of silica nanoparticles in the organic resin several options were reviewed, during the project.

First a classical reverse emulsion process (w/o) has been explored to design porous spherical nanoparticles of silica in which epoxy resin had replaced partially the inorganic solvent.

The results were promising because the resin’s behaviour was not altered by the introduction of chemicals (surfactant, co-surfactant...) however the total amount of inorganic nanoparticles was limited (0,2-0,6% in weight) and the epoxy resin was not a realistic resin use in electrical industry. In the second step, the maximum loading of nanoparticles versus the viscosity of industrial resin so as to optimize the proportion of silica content for the in-situ synthesis was determined. 5% in weight was targeted as the maximum loading in order to minimize the impact on the viscosity.

Finally, tailor-made syntheses were performed by mixing optimized amount of water, aqueous base, industrial resin and silicium alkoxide at suitable regulate temperature. Porous spherical nanoparticles of silica (<100 nm) were synthesised directly in the industrial resin without any solvent.

The industrial resin used were Damisol 3309 (polyester-imides, Figure 5) and EPE19 (bisphenol A- epoxy, Figure 4) supplied respectively by Von Roll and Alstom.

Several synthesis were developped to play with the silica content from 1.5% to 3.2%. Thermal tests as well as space charge measurements were done in order to select the best formultation for un-scaling. According to those data, a tailor-made synthesis with 1,8% of silica particles was up-scale from gram to kilogram as it was detailed in the materials and methods part.

Furthermore, a new process has been investigated to decrease the size of the nanoparticles down to 10-15 nm, accordingly to the consortium discussions on impregnation efficiency and on the potential benefit of smaller size particles of dielectrical performances.

The newly developed synthesis has been applied to 5L batches to allow prototype testing. In the second part of the project, 50 kg of EPE19 doped resins as well as 30 kg of doped Damisol resins were produced and were sent to Alstom for bars impregnation and Von Roll for laminates and prototyping. In the same time, the production of 1 mm thick samples for breakdown, dielectric spectroscopy and corona resistance were carried on in order to complete the characterisation requirements.

The weibull distribution measured on the samples showed the introduction of nanoparticles improved the voltage breakdown strength. The samples from both batches Exp 37 and Exp 43, which contain 1.8% of silica nanoparticles, were reproducible in terms of breakdown spectroscopy. The homogeneity as well as the higher resistivity of those samples remained significant compared to the neat resin.

On another hand, the permittivity of the tailor made resin was a little bit higher than the neat resin. As laser ablation of samples is a very good indicator of corona resistance, carbon dioxide laser (Synrad) at wavelength 10.5-10.6 µm was used to irradiate samples. The mass loss was measured. It appeared that the silica protect the resin from laser ablation.

The promising performances of the resin developed were confirmed by the VET tests done on the impregnated bars in Alstom. However, the instability of resin during transport and its oxidation were pointed out. As well, the aggregations of silica were observed in several batches:

In order to fix those problems, propositions of set-up optimisation were discussed. It had been decided to try new synthesis under inert gas. The second set-up described previously was designed after several synthesis tests.

Indeed, the first trial with closed reactor has highlighted the problems of by-products. The IRTF measurement revealed the presence of water and non-reacted components. It appears that the condensation was not occurred in this trial. The Soxhlet and condenser were added to the set-up in order to remove the water during the synthesis and move the reaction from hydrolysis to condensation. Several trials were performed with stirring and temperature of the process optimizing until the viscosity and the IRTF measurements showed the complete water removing.

1.6 The difficulties encountered during the project

The major difficulty encountered during the last month was to find the best way to produce the samples for the several characterisations.

Indeed, in order to obtain flat plate without thickness variation, moulds were designed and processes were developed. The high faculties of the resin to stick to the mould were a problem. Several releasing agent were tested and finally a Teflon-based overcoat was find to be the best solution for removal the sample from the mould. Also this alternative permit to work without any chemicals added on the mould that could be transferred to the plate.

The versatile dimensions of the samples for the several tests gave us lot of work. For each thickness, moulds should have been design and for the thickness below 0.5mm the samples were too brittle to achieve those dimensions.

The important delay of the characterisations results was also problematic for an efficiency formulation work. The adaption to the planning of all partners has been done as good as possible. However, quite long periods remained between sending the samples and the characterisations. Thus compiling all the data and determine the best synthesis took much more time that predicted.

2 Bottom up approach / Synthesis of a sol-gel inorganic / organic hybrid matrix

2.1 Introduction: Quasi inorganic approach

Route 3A has the objective to develop homogeneous hybrid sol (organic/inorganic) tape structures with outstanding thermal and dielectric properties. Completely avoiding polymer phase inside the tape is not realistic since the tape must remain flexible enough to be wound. However using this Sol-Gel method to impregnate the tape carrier could bring significant advantages, namely:

• A silica or zircon based hybrid matrix chemically compatible with both tape carrier and dielectric barrier materials (mica or silica flakes).
• Materials promoting higher thermal conductivity.
• The possibility of quite easily nanostructuring this silica based hybrid matrix
• Very good sol impregnability (possibility of very low viscosity solution).
• A soft and more environmentally friendly chemistry.
• The possibility to use optical diagnostics to control the nanofiller content inside the tape.

Route 3B addresses the development of a multilayer process for obtaining multifunctional tapes: to this aim, the control and optimization of the properties can be ensured by depositing a layer by layer (LbL) assembly on the tape.

2.2 Route 3A of the project

In this part, the strategy was to select each precursor with best performance in thermal conductivity and dielectric strength. The advantage of this technology is the wide choice on precursors and the composition of the material.

So a complete state of the art was done the select these precursors and an experience plan to find the ideal composition. Before include any particles in our material all the precursors and matrices must be chosen for compatibility reasons, and material properties (mechanical, ageing, thermal and compatibility reasons).

The first step was to select the appropriate elements for the matrix composition and secondly for the fillers introduction. All elements have a particular task:

• PDMS for the mechanical flexibility, thermal resistance and ageing.
• Zirconium like cross linkage element to realize a 3 dimensional network with thermal resistance too.
• Fillers: 2 routes can be explored with first the growing in the matrix of nanoparticles and then the introduction of nanoparticles in the bulk material during reaction step.

The precise Sol Gel reaction is described below, the first step is the chelation of the metal alcoxide with a ligand to form a stable complex and avoid precipitation during hydrolyses. The coordination number is increased by a nucleophilic substitution with an alcool, acid, base or chelating agent.

2.3 Experience plan:

So a ternary diagram was done to investigate the different compositions of materials in order to reach the ideal properties researched. This method of organisation allows a real control and reproductivity of our samples. All compositions can be explored and exactly adjusted to obtain the required properties.

2.4 Mechanical characterization:

In accordance to the objectives and the experimental plan mechanical tests have been made to measure the tensile strength of our elastomeric material. So Young’s modulus, tensile strength, elongation and percent elongation at break are observed. The ability of the Sol-Gel material to deform is determined by the mobility of its molecules, characterized by specific molecular motions and relaxation mechanisms that are accelerated by temperature and stress. Since these relaxation mechanisms are material specific and depend on the molecular structure, they are used here to establish the desired link with the intrinsic deformation behaviour.

Test pieces used for this characterization must have an accurate known dimension. After stirring and step of reaction the solution is poured in a Petri dish in Teflon to avoid any grip to the mould. After curing time thickness of the sample is measured and shape of the test piece is made with a pastry cutter. For our application these properties of elastomeric material are useful for the implementation of the tape. Indeed, we have a low viscosity to impregnate the tape and high flexibility to assembly the elements of motors and turbines. Moreover, thanks to this figure 7 we show that mechanical properties are not a limiting factor.

This characterisation will be also done for ageing characterisations after high temperature cycles and expositions. All these results correspond to the optical characterisations (Delivrable D2.5) of the Fourier Transform Infrared Spectrometry. Indeed, the increase of strength at break is the illustration of the bonds represented in the material. The addition of Zirconium stiffens the structure of the material because of cross linking and mobility of polymer chains is decreased. There is the same for the addition of SiO2 charges which reinforces the material.

2.5 Thermal characterization:

Thermal characterization contains 2 properties that materials must reach:

• Thermal conductivity (SOA:0.3W/m.C°) (T) which have to be higher than 0.5 W/m.C°
• Thermal class (SOA:155°C class F) must increase to class H 180°C

2.6 Thermal Conductivity:

This measure represents the heat quantity transferred per surface unit and per time unit under a thermal gradient of 1°C/m.

At this point, only the diffusivity and density values are obtained. Calorific capacity is being measured to have the thermal conductivity of the different samples. Nevertheless, these thermal diffusivities are higher than an epoxy resin (0.06mm²/s). All these results will be explained on following part and related with mechanical values. Concerning experimental procedures samples have accurate dimensions (10mm of diameter and between 1 and 2 mm of thickness). It’s a thermal analyse based on Infra Red radiations: the source is an IR lazer and the detector measure IR radiations. So, a graphite deposit is done on samples to allow the material to absorb and emit all Infra Red radiations. As mechanical tests the sample are sized with a pastry cutter.

2.7 Results correlation:

In this part an experience plan is done to find the better compromise between all the results. The goal is to limit the number of experiences by mathematic simulation in order to reach the ideal material composition. A 3Dimensionnal graphic is shown on figure 12 to illustrate the results by a correlation between all the results previously presented. System resolution and plots have been done with software MATLAB. At the highest crossing of these two plans the best characteristics of our materials are shown. The results given by this simulation have been analysed and the couple of properties is validated. We have selected the coordinates of the crossing point with the value of thermal diffusivity which was 0.15 mm²/s. This composition was synthesised and the value of thermal diffusivity was 0.14 mm²/s. So we can conclude that our experience plan is validated and electrical tests can be implemented.

2.8 Electrical characterization:

According the, some thermal measures have been over the objectives and electrical tests have been done on the previous measures. Different methods have been applied like Space charge measurements, breakdown strength to reach the electrical objective. One of the key feature of this material is the development of systems that exhibit enhanced breakdown strength and which, consequently, are able to withstand higher fields in service without undergoes electrical degradation. The Breakdown tests were performed using a ‘high’ HV AC breakdown kit which has been modified to run at a ramp rate of 50 V/s with a sphere to sphere electrode system and the final breakdown voltage was recorded. Several remarks can be done on theses results.

First of all, the β (shape parameter) low value is due to the variance in thickness of the samples because of the pouring process which have to be improved. Then, the breakdown strength is a real value despite of the thickness which distorts the material homogeneity. And the highest breakdown strength value is obtained for the same that thermal test. The best composition identified for all the compositions tested is 1:4 Zr:PDMS with 8% of silica particles. So the introduction of silica particles has a real influence and improvement on this material for breakdown resistance.

2.9 Space charge measurements:

The thermal Step Method (TSM) developed at the University of Montpellier 2, France, completes and explains the previous measures of breakdown strength. Indeed, the presence of space charges in materials seems directly linked to the deterioration of their physical properties since charge trapping is mainly associated with structure defects.These defects may have several origins, namely manufacturing process, interfaces, and stresses applied to the material. In the nanocomposite materials, the interfaces are numerous but in the vicinity of nanoparticles with the electric field can be very high and initiate volume conduction phenomena which can encourage the flow of space charges when the material is submitted to high applied electric field. So, this technique reveals the space charges contained in the insulation volume.

This is a non-destructive measurement technique for both the material and contained space charges (they are not evacuated during the measurement). This is a major aspect of the TSM, as it allows following the evolution of residual electric field for different applied electrical stresses. Space charges measurements in short circuit conditions vs. applied electric field were performed on 2 samples with and without silica nanofillers.

Two measures were performed before and after an electrical poling up to 8 kV/mm at 60°C during 2 hours. The maxima currents (TSM signal) on the cathode and anode are represented on the figure 7 and show almost the same amount of space charge in each material before any electric exposure du to the defects expressed previously. The presence of sub-micron particles of SiO2 has weak impacts on the electric behaviour of this vacuous material with regard to the sample without particles of SiO2.

Signals of same amplitude on anode and cathode denote symmetry in the distribution of space charges in the thickness of the sample. It is mainly bound to the manufacturing process of samples and essentially their cooling which was to be almost identical on two faces. Then, the electric conditioning of samples seems to impact more on the sample without SiO2 than on the sample with particles. The presence of nanoparticles tends to limit the accumulation of space charges. A mathematical treatment of the signals would allow to obtain the distribution of space charges in the thickness of samples and to confirm these hypotheses.

2.10 Corona and erosion resistance:

The testing procedures described above all focus on bulk properties. In addition, it is also desirable to assess the resistance of the novel systems being developed here to surface electrical activity. Exposure to these specimens to corona using conventional electrode arrangements led to little sample damage and, therefore, an alternative approach was used to simulate surface discharge activity. Arcing on insulator surface results in the deposition of energy which, in turn, results in local degradation. Since it is difficult to quantify the energetic of this using electrical means, local degradation was induced using a 5s pulse from a 28W CO2 laser set to a 50% power output. The results indicated only a slight mass loss. In this route, the control of tape properties has to be ensured by a multilayer coating on tape. A good method for controlling the tape nanostructuration spacing or periodicity along the electrical field direction is the layer-by-layer repeated deposition of thin film and curing, to produce a stacking of different layers. The multilayer approach allows combining different functional layers taking different nature, size and surface properties of particles for each layer. Some tests have been done on this part but we don't have enough time and results to put in this report. Nevertheless, we will continue this route because some results are encouraging.

2.11 Conclusion of this route

The ANASTASIA consortium set out to push to the limit of the present insulation tape technology for HV rotating machinery. Several routes have been explored for the nanostructuration and nanodielectrics syntheses. The Sol-Gel route presented in this report is the high risk route where all the components of insulating material have been changed. Flexible nanodielectrics with inorganic cross-linking agent and inorganic fillers rise above thermal and dielectric properties of several epoxy resins in a certain proportion (see table). Some compositions with Zr/PDMS with silica particles were selected and give the best results of all samples tested with others cross-linking elements and fillers. This study reveals all the advantages of this Sol-Gel synthesis which made this material a promising candidate for high power insulating

3 Synthesis of the bottom-up approach (route 2A) / Quasi-inorganic approach (route 3B).

3.1Technical approach

POLITO has been involved in the bottom-up approach (route 2A) and in the quasi-inorganic approach (route 3B). Route 2A concerns the in-situ synthesis process of hybrid organic/inorganic composites. Route 3B addresses the development of a multilayer process for obtaining multifunctional tapes: to this aim, the control and optimization of the properties can be ensured by depositing a layer by layer (LbL) assembly on the tape. Furthermore, POLITO has been involved in the thermal characterization of the materials prepared by some of the other partners of Anastasia Project (namely, NTU, CEA, Von Roll and Alstom). For this purpose, POLITO has performed differential scanning calorimetry (DSC) and thermo gravimetric analyses (TGA); these latter have been carried out on the basis of a standard procedure set by POLITO and described in the appendix of this report.

• Route 2A

POLITO has prepared coatings by using a dual-cure process that combines a polymerization induced by UV-light with a subsequent thermal treatment for promoting sol-gel reactions. The coatings formulation and the dual-cure conditions have been studied and optimized. In particular, the liquid UV-curable formulations were exposed to UV light and subsequently thermally treated at 80°C for 12 hrs at 100% RH. Indeed, humidity favors the hydrolysis of the alkoxysilanes and the formation of silica nanoparticles embedded in the epoxy matrix. The dual cure process is schematized in Figure 1.The first set of samples was based on the following materials:

o A cycloaliphatic epoxy resin, namely 3,4-epoxycyclohexylmethyl 3’,4’-epoxycyclohexanecarboxylate (CE),
o A coupling agent, 3-glycidoxypropyltrimethoxysilane (GPTS),
o A flexibilizing comonomer, namely hexanedioldiglycidyl ether (HDGE),
o A Poly(dimethylsiloxane), diglycidyl ether terminated (GTPDMS; Mn≈ 980)
o A cationic photionitiator, triarylsulfonium hexafluoroantimonate diluted in 50% propylene carbonate solution
o Different alkoxysilanes (tetraethoxysilane, TEOS or tetramethoxysilane TMOS).

3.2 Route 3B

The adopted layer-by-layer deposition technique consists in alternatively dipping the substrate in oppositely charged polyelectrolyte solutions or nanoparticle dispersions so that a coating of positively and negatively charged layers piled up on the substrate surface is formed.

3.3 Best results obtained

- Route 2A

The first set of samples was characterized by means of termogravimetric analysis in air, differential scanning calorimetry, space charge and dielectric permittivity measurements. In particular, in order to perform TGA analyses, a standard procedure was agreed among the Anastasia partners. The procedure is reported in the appendix of this report.

The impregnation of HDGE/CE UV-cured systems with TMOS determines the formation of irregular silica microparticles. The same behavior (formation of irregular silica nanoparticles together with low silica contents) was found when GTPDMS was copolymerized with CE or CE/HDGE formulations and the obtained films impregnated with TMOS. On the other hand, the addition of the coupling agent (GPTS) to the UV-curable epoxy mixtures (i.e. to the epoxy co-monomers) promotes the formation of quite a high silica residue (up to 10-12 wt. %) and leads to spherical silica micro- and submicro-particles well distributed within the polymer. On the basis of the obtained results, the dual-cure processes exploited for the formation of silica phases within the UV-cured epoxy matrix (Route 2A) were not easy to control, as far as the morphology, distribution and content of silica are concerned. The last results achieved in the presence of GPTS were found quite promising for the prosecution of the research activity; their feasibility to the project requirements could depend on their electrical properties. Small stripes of Von Roll Mica tape (length = 90-100 mm) were cut and placed on a glass substrate; the UV-curable mixture was poured on the tape and was kept in contact for 2 min in order to favor the impregnation step; then the excess of the liquid mixture was removed by using a 10 m wire wound applicator.

The impregnated tapes were placed under a static UV-lamp (Helios Italquartz, Milano, Italy) for performing the photo polymerization process. The UV radiation intensity on the sample surface was measured by means of a UV-meter and was found equal to 55 mW/cm2. Each surface of the impregnated mica was exposed to UV-light for 30 s.

SEM microscopy (LEO-1450VP apparatus) was exploited aiming to assess the impregnation level of the mica tape by the different UV-curable formulations.

- Route 3B

In parallel to the preparation of silica/epoxy nanocomposites by the sol-gel method, POLITO has also prepared and characterized layer by layer coated samples by using different plastic substrates, namely UV-cured epoxy systems (having a composition similar to that described in Route 2A), polycarbonate sheets (PC, Makrolon® ET3127, MFI: 7 g/10min at 300°C/1.2 kg), epoxy resin-rich mica tapes (supplied by Von Roll). To this aim, a robot for LbL was used; some attempts were also performed for LbL treating the tapes with a pilot roll-to-roll plant. Regardless of the used substrate, the deposition was made by using two nanoparticle aqueous dispersions, bearing an opposite electrical charge: the positive dispersion is made of Ludox CL alumina-coated silica nanoparticles (1 wt. % concentration) and the negative consists of Ludox TM40 silica nanoparticles (1 wt. % concentration). Prior to the deposition, the epoxy and polycarbonate substrates were activated and cleaned by a cold plasma treatment (using a 40 kHz Pico 1.1.2 semi-automatic controlled system - Diener Electronic GmbH -), performed at 50 W for 5 min, in the presence of Argon and Oxygen (flux of each gas: 10 cm3/min). The depositions were carried out obtaining 10 and 20 bi layer assemblies (Table 4). Figure 9 collects some typical SEM pictures of the layer by layer-treated UV-cured epoxy systems: it is worthy to note that the LbL coating is quite homogeneously deposited on the surface of the substrate. These samples have been sent to IREQ and UM2 for the dielectric characterization. Since the layer-by-layer deposition seems to be useful for finishing treatments performed on the final impregnated tape regardless of its composition and structure, the LbL technique has been further investigated.

In particular, prior to apply the LbL to resin impregnated mica tapes, POLITO has chosen polycarbonate sheets (2 mm or 200 micron thick) as a model substrate, on which the LbL coatings have been assembled. Figure 10 presents some typical SEM pictures of the layer by layer-treated PC sheets: again, the coverage of the polymer surface is quite homogeneous; furthermore, the thickness of the coating (in the sub-micronic range) strongly increases by moving from 10 to 20 BL, as indicated by the EDS signal related to Si element.

In order to further characterize the obtained LbL assemblies, Attenuated Total Reflectance (ATR) spectroscopy and wettability measurements have been carried out. The former allows qualitatively evaluating the composition of the surface of the coating deposited on PC sheets. The deposited coating grows as a function of BL number and, at 20 BL, it becomes thick enough so that only the silica signals (at 1150 cm-1) are detectable. Wettability measurements have been carried out by measuring the static contact angle with water of PC before and after the LbL treatment. The LbL deposition of a coating made of silica nanoparticles is capable of significantly reducing the water contact angle, i.e. increasing its hydrophobicity: this finding can be ascribed to both the hydrophilic nature of the deposited nanoparticles and to the increase in surface roughness occurred after the deposition, as revealed by SEM observations. The LBL treated PC sheets have been sent to US and UM2 for the dielectric characterization.

The silica/silica LbL treatments have been performed on Von Roll resin-rich mica tapes (Samicatherm® 366.28-02). The depositions were made by using the already mentioned two nanoparticle aqueous dispersions, bearing an opposite electrical charge. Prior to the deposition, the resin-rich mica tapes were activated by a cold plasma treatment (using a 40 kHz Pico 1.1.2 semi-automatic controlled system - Diener Electronic GmbH -), performed at 50 W for 5 min, in the presence of Argon and Oxygen (flux of each gas: 10 cm3/min). The depositions were carried out obtaining 10 and 20 bilayer assemblies. Then the samples were examined by SEM microscopy and subsequently subjected to a thermal curing at 170°C for 18 hrs. The morphology of the coating after the curing process changes quite drastically: indeed, the LbL coating turns out to be partially removed from the mica tape.. This suggests that the adhesion of the silica/silica assembly cannot be preserved after having thermally treated the mica tapes. Therefore, a new series of 8 square silica/silica LbL-treated samples was prepared, piled up and sent to Von Roll for the thermal curing (under pressure, simulating the real processing conditions). The obtained LbL-treated laminates have been sent to UM2 and US for the dielectric characterization.

4 Resin and tape characterization - Results.

The WP3 work comes to an end. WP3 entitled "Resin and Tape characterization" mainly concerns electrical and thermal characterization of materials developed in WP2. Three design routes were carried out in WP2 to achieve a total of about 150 different materials. The different characterization techniques proposed by the partners involved in this WP resulted in a big quantity of experimental results, which allowed highlighting potential candidates for the prototype development, fulfilling the targets of the project. In general, the different samples issued from the three routes proposed in WP2 and the « laminate » materials issued from the WP4 have already been reported in several intermediate deliverables done within this WP.

The deliverables, which will be described in this report, are D3.2 D3.3 and D3.4. The various sample designs were defined in deliverable D3.1 based on needs identified by several characterization techniques.

4.1 Samples investigated

These materials mainly based on the epoxy nanostructuration come from different manufacturing processes issued from the routes proposed in WP2. Several processes of implementation have been used:

• Incorporation of particles within the matrix (Route 1- Top-down approach).
• In-situ nanostructuration of the resin (Route 2 – Bottom-up approach).
• Nanoscale assemblies through the Layer by Layer method (Route 3 – More inorganic approach).

The nanostructured resins are derived from the following components:

• Different types of polymer matrices (DER332, Damisol, Epoxy Resin, ...)
• Different types of particles (Clays, SiO2, BN, CNT, Al2O3, Aln, ...)
• Different shapes (spherical, flakes, wires, ...)
• Different sizes (nano, micro, sub-micro, ...)
• Different surface treatments (washing, drying, functionalization, ...)
• Different content and micro to nano ratio

In the « basic » characterizations, the most encouraging results were used for the implementation of "tapes". These "tapes" are produced by using a mica paper, a glass fabric and a resin impregnation. The tapes and laminates (made of several stacked tapes) were produced under WP4.

4.2 The sample characterizations

This WP was organized according to several steps to identify candidate resin systems for the development of prototypes. Their characteristics should correspond to the targets defined in the specifications of the Anastasia project. The project's targets are mainly related to an improvement of electrical and thermal properties, namely, improved breakdown voltage insulation for electrical aspects and an enhancement of thermal conductivity and thermal classification. To achieve this aim, several techniques for characterizing the thermal and electrical behavior have been proposed by the partners involved in this WP.

Electrical characterizations were performed by assessing the dielectric strength, space charge, dielectric permittivity and erosion resistance.The measurement of glass transition temperature by differential scanning calorimetry, mass losses by Thermo gravimetric Analysis, diffusivity and/or thermal conductivity and viscosity were exploited for determining the thermal properties of the materials.

4.3 Table of results

A summary of the different materials produced by each partner involved in WP2 and WP4 will be provided before each table. For more detailed information about the samples, it is advisable to consult the WP2 and WP4 deliverables.

4.3.1 Route 1: Top-down approach

- NTU samples

NTU has produced several series of nanocomposite samples based on a commercial resin, namely DER "PHR 34" (composition: 50g of DER332 and 17g of PPG_NH2) essentially incorporating nanoclays (C10A and C30B clays). Some samples are based on industrial Damisol resin mixed with C10A clay. The samples differ for the amount of nanoparticles (2, 3 and 4 wt %), their treatment (washing, drying at different temperatures), and the polymerization conditions. US samples

The samples made by US, are also based commercial DER epoxy resin (DER332 + hardener Jeffamine D-230), in which nanoparticles (Boron Nitrade and SiO2) having different sizes (nano, micro and submicro) and different contents (2.5 5 and 10% wt) were added.

Before mixing the particles, several resin/hardener ratios, curing time and temperatures were initially explored to obtain the correct process parameters, as shown in the graph below:

- CEA--LCPE samples

LCPE group from CEA has been working to produce samples from industrial resins (namely, Damisol 3309 and EPE19) by adding various nanofillers (SiO2, Clays, CNT, Aln, Al2O3) with different contents (from 0.01% to 1% by wt depending on the filler).

- POLITO samples

The samples made by POLITO in this route refer to the in-situ synthesis of hybrid organic/inorganic composites, consisting of a mixture of epoxy monomers (HDGE and CE) in the presence of a silica precursor (tetraethoxysilane). Furthermore, glycidoxypropyltrimethoxy silane (GPTS) was used in order to couple the two phases. It consists of a dual-cure process that combines a polymerization induced by UV-light with a subsequent thermal treatment for promoting sol-gel reactions. Humidity favours the hydrolysis of the silicon alkoxides and the formation of silica nanoparticles embedded in the epoxy matrix.

- CEA-LCSN samples

The CEA-LCSN group has manufactured resins based on industrial Damisol or EPE19, which were in-situ nanostructured. Indeed, these resins allow silica nanoparticles growing thanks to a sol-gel process based on the hydrolysis and condensation of silica precursors. Thus, porous spherical nanoparticles having a diameter in the 20 to 100 nm range have been synthesized directly in industrial resins without solvents for obtaining samples containing 1 to 5% by wt of SiO2. The characterization results obtained on these materials are presented hereafter.

- Von Roll samples

"Laminates" samples made by Von Roll consist of a stack of 8 tape layers. Several tapes were used: a reference industrial tape "Samicapor 366.58" or a modified tape by the addition of a submicron layer of Boron Nitride an/or the addition of a nanoSiO2 layer. These laminates were then impregnated with reference industrial resins (Damisol or EPE19) or with in-situ nanostructured resins provided by the CEA-LCPE group. Further details are given in deliverable D3.4 however a summary of experimental results is given in the table below.

4.4 Candidates for prototype

Characterization campaigns carried out in the frame of WP3 have allowed highlighting potential candidate materials for the prototypes developments.

First, these materials must be optimized and homogeneous (perfect control of the manufacturing process) and then have to possess both good electrical and thermal properties.

4.5 Conclusions:

All electrical and thermal characterizations have allowed ascertaining the particles that lead to a significant improvement of electrical and thermal properties of the studied materials.

Thus, it was possible to select potential candidate materials for the prototypes development in the frame of WP5, based on:

- The manufacturing process (depending on the route adopted in WP2)
- The resin (priority use of industrial resins commonly used by Von Roll and Alstom Companies)
- The particles (strong consideration of their nature, shape, size, content, processing and treatment)

The following systems used singularly or in combination can be suggested as potential candidates:

- Boron nitride sub micro particles
- Nanosilica filled epoxy resins
- Sol-Gel silica resin in situ growth
- Clay

5 Demonstrator stator bars
5.1ALSTOM’S activities
ALSTOM’s role in ANASTASIA project was to manufacture standard demonstrator stator bars with the selected material scenarios. Additionally, the impact of the advanced insulation systems on the generator design was provided.

The following deliverables were made by ALSTOM:

• D4.1 Report on material scenario and tape specification
• D5.2 Demonstrator stator bar tested
• D5.3 Report on the impact of the new insulation on rotating machines

Deliverable 4.1 presents an overview of the state of art of manufacturing processes for high-voltage insulation of stator windings of electrical machines and the requirements for the insulation tapes and resins used for such an insulation system.

Deliverable 5.2 summarises the activities and results concerning demonstrator stator bars, which were manufactured using the newly developed nanostructured mica tapes and resins introduced in the course of the ANASTASIA project. The electrical performance of the demonstrator stator bars with modified main wall insulation was assessed.

Deliverable 5.3 demonstrates the influence of the enhancement as defined in ANASTASIA targets on the design of a complete machine. Performance of existing machines was recalculated with using new values for electrical field strength and thermal conductivity.

5.2 Material scenario and tape specification

According to the state of art for today’s manufacturing process for the manufacturing of high voltage stator bars there are two possible material scenarios. The first one is for the Vacuum Pressure Impregnation system (VPI) and the second one is for the Resin Rich Technique (RRT).For Anastasia both tape specifications were applied. In addition the specification for a fully cured insulation describes the requirements for final stator bars or coils.The tape specifications for the two systems have different material properties. A VPI tape needs to be different from a RRT tape. The processing of both systems is different and requires different properties.

5.3 Demonstrator stator bars

Demonstrator stator bars were manufactured using vacuum pressure impregnation technology (VPI), where a mica tape with very low resin content is wrapped around a copper conductor and impregnated with low viscous impregnation resin under vacuum. At last, the bars are cured in order to reach the dimensional stability as well as desired mechanical, thermal and electrical properties. In this work, full copper profiles were used as conductor bars.

5.3.1 Selected material scenario

Based on the results from laboratory scale investigations with other ANASTASIA partners, the following material scenario was chosen for demonstrator stator bar manufacturing.

Impregnation resins tested

- Standard Resin
- Standard impregnation resin of Alstom
- Used as a reference
- Nanosilica Resin
- Tailor made synthesis of colloidal silica added to the standard impregnation resin
- Nanoclay Resin
- Organoclay dispersed into the hardener of the standard impregnation resin
- Batch 1

Demonstrator stator bars were manufactured using Standard Tape and Standard Resin as well as Nanosilica Resins Exp37 and Exp43. Two impregnations with Nanosilica Resins were made, but only one of them was successful.

- Batch 2

Demonstrator stator bars were taped using Standard Tape, Nanosilica Tape (B12) and Nanosilica & Boron Nitride Tape (B13) and impregnated with Standard Resin and Nanosilica Resins (Exp69 & Exp71). Due to the increased thickness of tape B13, fewer layers could be applied into the insulation. Also the tape B13 was rather stiff due to its increased thickness and the additional BN layer. Application of nanosilica layer onto the glass side of the basic tape did not seem to affect the flexibility or thickness of the tape significantly.

- Batch 3

Demonstrator stator bars were taped with Standard Tape, Nanosilica Tape (B12) and New Nanosilica & Boron Nitride Tape (B15) and impregnated with Standard Resin, Nanosilica Resin (Exp93) and Nanoclay Resin. The tape B15 was an optimised version of the tape B13, resulting in reduced tape thickness. This allowed the mica amount and number of layers of the final insulation to be kept the same compared to the reference. The mica of B15 was based on the same mica paper, but the glass fabric was thinner. Nanosilica was added into the binder resin between mica and glass instead of being applied as a kiss coating on the glass side, which was the case in tape B13.

The Nanosilica Resin (Exp93) was made with adjusted synthesis under controlled atmosphere in order to avoid agglomeration of silica nanoparticles. The agglomeration of the silica nanoparticles into white large clumps as seen in case of Exp71 for instance, could be avoided, but small amount of brownish agglomeration happened and therefore the resin was filtered after synthesis.

Impregnation with the Nanoclay Resin was also tried, but due to the strong shear thinning viscosity, the bars were not fully impregnated and therefore no voltage endurance test was performed on those bars.

5.4 Impregnation quality

The following measurements were performed to check the impregnation quality of the impregnated bars:

• Dissipation factor measurement
• Partial discharge offline measurement
• Organic content
• Glass transition temperature
• Microscopic analysis

The quality of the bars was proven to be good by dissipation factor measurement, except for nanoclay impregnated bars. PD Offline measurement could not detect partial discharges. Insulation made with the New Nanosilica & Boron Nitride Tape had significantly lower resin content, only 18 wt% compared to 22-23 wt% of the reference. Glass transition temperature was found out to be lower for bars impregnated with Nanosilica Resin.

Potential Impact:

Anastasia has shown that some very interesting performance improvements can be achieved in dielectric strength, in space charge accumualtion and thermal activities. The addition of particles in the tape as well in the VPI resin is positive in terms of performance improvement, however we need to investigate deeper the influence and contribution of the different parameters. The result and values are too scattered for finalizing the definition of an optimal tape design.

Also, materials systems relevant to technological deployment in rotating machine application have been developed in ANASTASIA. These materials have been used to construct demonstrator bars and improved characteristics have been shown in ANASTASIA. This is not the end, experience of these materials is currently limited and we believe that further refinements will lead to greater benefits. The ANASTASIA concept has been proven.

ALSTOM’s role in ANASTASIA project was to manufacture standard demonstrator stator bars with the selected material scenarios. Additionally, the impact of the advanced insulation systems on the generator design was provided.

Deliverable 4.1 presents an overview of the state of art of manufacturing processes for high-voltage insulation of stator windings of electrical machines and the requirements for the insulation tapes and resins used for such an insulation system.

Deliverable 5.2 summarises the activities and results concerning demonstrator stator bars, which were manufactured using the newly developed nanostructured mica tapes and resins introduced in the course of the ANASTASIA project. The electrical performance of the demonstrator stator bars with modified main wall insulation was assessed.

Deliverable 5.3 demonstrates the influence of the enhancement as defined in ANASTASIA targets on the design of a complete machine. Performance of existing machines was recalculated with using new values for electrical field strength and thermal conductivity.

1. Impregnations

During the project, the demonstrator stator bars were manufactured in three batches:

Batch 1: Bars 2011.09.01-32
Batch 2: Bars 2012.03.01-36
Batch 3: Bars 2012.05.01-30

- Batch 1

Demonstrator stator bars were manufactured using Standard Tape and Standard Resin as well as Nanosilica Resins Exp37 and Exp43. Two impregnations with Nanosilica Resins were made, but only one of them was successful. Demonstrator stator bars of batch 1. Bars taped with Standard Tape and impregnated with Nanosilica Resin (Exp43)

- Batch 2

Demonstrator stator bars were taped using Standard Tape, Nanosilica Tape (B12) and Nanosilica & Boron Nitride Tape (B13) and impregnated with Standard Resin and Nanosilica Resins (Exp69 & Exp71).

Due to the increased thickness of tape B13, fewer layers could be applied into the insulation. Also the tape B13 was rather stiff due to its increased thickness and the additional BN layer. Application of nanosilica layer onto the glass side of the basic tape did not seem to affect the flexibility or thickness of the tape significantly. Demonstrator stator bars of batch 2. Bars taped with three different tapes and impregnated with Standard Resin

- Batch 3

Demonstrator stator bars were taped with Standard Tape, Nanosilica Tape (B12) and New Nanosilica & Boron Nitride Tape (B15) and impregnated with Standard Resin, Nanosilica Resin (Exp93) and Nanoclay Resin. The tape B15 was an optimised version of the tape B13, resulting in reduced tape thickness. This allowed the mica amount and number of layers of the final insulation to be kept the same compared to the reference. The mica of B15 was based on the same mica paper, but the glass fabric was thinner. Nanosilica was added into the binder resin between mica and glass instead of being applied as a kiss coating on the glass side, which was the case in tape B13.

The Nanosilica Resin (Exp93) was made with adjusted synthesis under controlled atmosphere in order to avoid agglomeration of silica nanoparticles. The agglomeration of the silica nanoparticles into white large clumps as seen in case of Exp71 for instance, could be avoided, but small amount of brownish agglomeration happened and therefore the resin was filtered after synthesis.

Impregnation with the Nanoclay Resin was also tried, but due to the strong shear thinning viscosity, the bars were not fully impregnated and therefore no voltage endurance test was performed on those bars.

2. Voltage endurance test results

A voltage endurance test (VET) was carried out at 38 kV and 27.5 kV. These test voltages correspond to an average field strength of 19 kV/mm and 13.75 kV/mm for 2 mm insulation thickness. Some of the bars were only tested at one test voltage.

The test was conducted at ambient temperature. For each batch of impregnations, a reference system with standard resin and standard tape was also made and tested simultaneously. Statistical analysis was done with the Weibull distribution with 90% confidence limits and by comparing the 63.2 %-percentiles.

3. Influence Nanosilica Resin

The interesting idea of creating silica nanoparticles directly into the impregnation resin has shown potential, as an improvement up to factor 2 in voltage endurance could be achieved with this type of modified resin. However, due to challenges with up scaling the synthesis, this approach would need further development work in order to better understand the mechanism affecting the dispersion and stability of the Nanosilica Resin.

4. Influence of Nanoclay Resin

Due to high viscosity of the Nanoclay Resin, demonstrator stator bars could not be fully impregnated. However, nanoclay has shown high potential in improving dielectric properties of the plate samples at the earlier stage of the project.

5. Influence of Nanosilica Tape

Nanosilica Tape has given contradicting results. In batch 2, the tape seems to have a negative impact on the lifetime. In batch 3, small improvement in lifetime was seen. All in all, the Nanosilica Tape did not have a significant influence on the lifetime, because the differences were rather small and the confidence intervals were strongly overlapping.

6. Influence of Nanosilica & Boron Nitride Tape

With the first version of the Nanosilica & Boron Nitride Tape, the electrical lifetime was significantly reduced. The reason for this was most probably the lower number of tape layers in the insulation resulting from increased tape thickness compared to the reference tape.

7. Influence of Nanosilica Resin in combination with Modified Tape

Nanosilica Resin slightly increases the lifetime also with modified tapes, but the difference is negligible with highly overlapping confidence intervals.

8.Influence of New Nanosilica & Boron Nitride Tape

Lifetime using the New Nanosilica & Boron Nitride Insulation could be improved by factor of 4.5 to 5.8 which is a significant improvement. However, this result should be considered only as an indication, due to the fact that the lifetime was calculated using estimated lifetime exponents, because the test was accelerated by using higher voltage.

9. Thermal conductivity test results

Silica particles applied into the tape or resin, did not influence the thermal conductivity of the insulation. However, thermal conductivity could be significantly improved by boron nitride. With the optimised design of the New Nanosilica & Boron Nitride Tape, the thermal conductivity resulted in improvement of ca. 43%, increasing from 0.30 W/mK to 0.43 W/mK.

10. Conclusions on demonstrator test bar

The tape BM104B15 has shown significant potential towards reaching the original targets of the Anastasia project. With this so called “New Nanosilica & Boron Nitride Tape”, the most important properties could be significantly improved – electrical lifetime by a factor of 4.5-5.8 and thermal conductivity by 43%.

For Alstom as an industrial partner, the improvements achieved with the modified tape are very interesting and seem to be more promising than nanostructuring of the impregnation resin. Additionally, the modified tapes would be easier to implement in the existing manufacturing set up than the resins, and they could be used for dedicated designs.

- Material systems relevant to technological deployment in rotating machine applications have been developed in ANASTASIA
- These materials have been used to construct demonstrator bars and improved characteristics have been shown in ANASTASIA

THE ANASTASIA CONCEPT HAS BEEN PROVEN

- This is not the end: experience of these materials is currently limited and we believe that further refinements will lead to further to greater benefits

List of Websites:

http://www.anastasia.eu/