Power Converter

Power Converter


Based on several usage scenarios developed by analyzing the market for energy storage and flow batteries, initial power converter system analysis was carried out. Within this task, numerous studies have been carried out in order to integrate the battery stack with the power converter. Several options are under evaluation and have been studied together. The main objective of the topology proposed by GPT is to obtain a rapid and cheap system. Therefore, system proposed is able to scale up according to grid and installation requirements. Highlighting scalability and flexibility together with operation safety and security guarantee have been the aims for the proposed developed. All of that is meaningless without solid cost justification; hence, after a deep cost analyzed, the topology is also reasonable compared to project requirements.

A scheme of the final topology proposed is presented in figures 16 & 17. Construction as well as test, simulation and results achieved along the different tasks will be based on this final proposal.

Figure 16: Final topology

Figure 17: Galvanic isolated DC/DC topology

Regarding the technical, economic and environmental analysis two options were deeply studied –modular and hybrid technologies– with different structure as shown next. At the beginning, the cells provided with energy storage capability needed a DC/DC converter. The aim of the converter were to boost the voltage provided by the battery up to the voltage of the DC link. A transformer was designed to be included in order to provide the required galvanic isolation.

The DC/DC converter contains a LC filter on the secondary winding of the transformer. The reason was that the second converter of the cell is in charge of providing a sinusoidal wave to the transformer and the battery DC/AC converter. The topology for this DC/DC is shown in Figure . This topology can be divided into three sections. From the battery side to the output inverter, the parts are a DC/AC before the transformer, a LC filter and a controlled rectifier just before the DC Link of the output inverter.

The first inverter, named as “battery DC/AC”, is intended to control the current in the primary winding of the transformer. By controlling this current, it is possible to control the current on the battery. This converter needs a stable sinusoidal voltage on its AC side. The AC voltage is provided by the “DC link rectifier”, whose AC side voltage is filtered and passed through the transformer to the primary winding. The output inverter is the interface between the continuous voltage of the DC Link and the grid. Its main purpose in the demonstrator is the power exchange between batteries and the grid.

Figure  shows a single output inverter. It is a standard half bridge single-phase inverter, a slight variation of a half-bridge VSI, where two large capacitors are required to provide a neutral point N, thus each capacitor maintains a constant voltage.

18: Single output inverter

When previous inverter topologies (Figure  and Figure ) are replicated and placed as shown on Figure , the intended system is achieved. These structures are easily scalable, so the line to line voltage needed in any application can be reached without problems.

Finally, because of economic and market reasons, the “Galvanic isolated DC/DC topology” was rejected. Hence, the “single output inverter” was chosen, although higher voltages batteries are required. Series association of batteries together with a battery management system could solve this problem in current commercial storage systems.

GPT/GTE has worked on the design, construction and preliminary tests of the final converter which was later sent to KEMA for final testing. The demonstrator is a scaled low power converter. It has been built in order to prove the proper performance of the selected topology.

Figure 19: Whole system installation with “Galvanic isolated DC/DC topology”

GPT has built a compact power converter prototype including the six IPM (individual power module) shown below in Figure . The aim was to build an easy to transport power converter in order to validate it in KEMA facilities.

Figure 20: Mounted POWAIR converter

Apart from the power converter, another objective was the design and test of the Hot-swapping characteristic. The application of the “Hot Swapping” concept provides safety operation in case of an unexpected event or failure, implying a longer lifetime of the developed equipment.

Design and implementation of control algorithms (GTE)

Two mode of operation were design:

  1. POWAIR working as a STATCOM; thus, reactive power is controlled and grid stabilization processes are applied.
  2. POWAIR in ESS mode; in this case, the installed system is able to control both active and reactive power. The extreme case of just one module connected to an ESS is validated through simulations and with different test in the power converter shown previously.

Design and build of the control hardware for real time control based DSP technology (GPT, GTE)

As can be observed in the Mounted POWAIR converter, differences were made between local boards which correspond to every BACC and local control boards in charge of the global control algorithms.

  • Central control
  • Local control

Design and implementation of global energy management of the power system

Design and implementation of global energy management is required in order to achieve an efficient energy global management. In addition improvement of basic behaviors from current battery systems, which are not able to adapt to customers and changing services demand, need to be considered. The system proposed by GPT which is modular and scalable, there could be some problems of power demand and unbalanced flow energy system, which is very critic to solve. Because of that the control system should synchronize every module in order to get the desired AC output, as well as monitoring almost every part of the system. Whether the system needs to achieve all those features, the control system receives data from each BACC (basic AC cell of the modular power converter) through sensors properly placed. Another important function is to display all this information for further analysis, such as communication state or BACC and battery stack data.

Converter modelling and optimization have allowed further improvements in size, efficiency, and usable power range. Improvements in converter efficiency and noise generation may be possible using a resonant or other converter approach.

Grid connection algorithms for different scenarios

The aim of this task is to validate the correct performance of the power system proposed. The complete system needs high performance grid integration control algorithms depending on the final application of the complete system. Therefore, a complete study of the possible applications of the proposed system has been carried out including applications such as:

  • Active power management in wind power systems
  • Fault ride through capability
  • UPS
  • Parallel active filter, STATCOM
  • Frequency control

E.ON has provided the University of Seville a couple of models. Control strategy and algorithms have been discussed for these applications.

The following simulations have been specified to University of Seville to demonstrate how the battery and power electronic converter control system can support the grid. All simulations have to be performed with and without the battery system to show the improved performance

The global model of the power system has been integrated in the grid model to test the control algorithms and to accurately predict the storage necessity and the overall performance of the energy storage system.

Modular power converters integration (GPT, GTE, EON)

Based on the list developed together with KEMA, next a summary of the test carried out to the POWAIR converter is shown. Three type are differentiated:

  • Routine tests required every time the POWAIR converter is installed in order to ensure secure and safety operation
  • FAT tests, which were performed in the USE lab with the objective to validate a proper operation without major failure. As a consequence, most of the found problems could be solved
  • SAT tests carried out in KEMA facilities, verifying the operation modes of the POWAIR converter as well as obtaining the characteristics of the equipment under test

Integration & commissioning of battery stack, power converter and control system

GPT and the University of Sevilla were responsible for planning and carrying out the complete FAT on module and system level. The FAT tests were performed in the laboratory of the University of Sevilla with the advice of KEMA / DNV GL. The FAT tests on a single module performed from 19-23 May 2014.

  • The FAT tests on at least 6 modules performed from 14-18 July 2014.
  • EUT commissioning, 15 September
  • Test in KEMA facilities: 22 September – 3 October
  • Packaging: 6 October-13 October

The FAT tests on at least 6 modules performed:


With the configuration shown in Figure 21, the aims are reactive power and grid stabilization control. The system will be able to provide the necessary reactive power according to an external output reference (sent from the HMI) or by means of control of voltage in the point of connection. The STATCOM commissioning process entail several tests, first for low voltages and then at the desired voltages and reactive power.

Figure 21: STATCOM configuration


Both, active and reactive power together with voltage and frequency control are the main characteristics of this working process. The objective was to demonstrate the operation of the POWAIR converter in the most adverse situation. Thus, the proposed scenario assumes that only one of the modules is connected to the used batteries or energy storage system as can be observed in the Figure . Usually, batteries are installed in every set of modules, implying a more efficient and reliable operation. However, the planned installation with a novel control algorithm was able to show that a modular implementation was possible with very good results. The best option would have been to show both scenarios, however, because of time restriction, it was decided to go for the most complex one.

The tested configuration could be considered as two different devices. On the one hand, the upper part of Block 2 (Figure ) will be acting like a STATCOM converter. On the other hand, the lower part will inject or absorb active and reactive power.

The first part of the commissioning test in the ESS mode focused on the control of the voltage of each capacitor and the reactive power produced by the upper part. Once both the voltages and the reactive power were controlled, the test proved the control of the lower part. This part of the EUT is in charge of the active power injection. Finally, both parts of the EUT were connected together.

  • STATCOM part of ESS mode
  • This part controls the reactive power the lower part injects to/absorbs from the grid
  • ESS part (lower modules) of the ESS mode
  • The control of this part of the EUT consists of three different functions: First, exchanging active power with the grid; the second one is controlling reactive power injected or absorbed to or from the grid; finally, the dc voltages on the capacitors not connected to the grid should be kept stable
  • ESS Complete Block
  • After the two part of the ESS mode were proved, the next step was the integration and testing, verifying the POWAIR was able to work according to the references provided.

Figure 22: ESS configuration

During the testing process of the POWAIR converter prototype, different failures and troubles were found and solved.

Testing of grid functionality and defining grid connection standards and requirements

  • STATCOM test

These tests were necessary to verify that the EUT system does not negatively influence the grid when inserted/integrated into an existing LV grid. Therefore, it demonstrated that the POWAIR system was able to inject or absorb the required reactive power in order to keep the voltage in the fixed level, also under abnormal conditions.

The aim was to obtain a complete voltage regulation, maintaining voltage variation within the limits in compliance with grid codes. Therefore, the EUT will be able to work in adverse situations and environments such as voltage variations, always taking into account power limitations of the equipment developed.

  • ESS test

These tests were necessary to verify that the EUT system does not negatively influence the grid when inserted/integrated into an existing LV grid. Therefore, it demonstrated that the converter system was able to inject or absorb the required reactive power in order to keep the voltage at the fixed level. Additionally, the equipment was able to store the necessary active power according to control characteristics.

The aim was to facilitate power balancing and frequency control in seconds and to overcome any constraints to keep currents and voltages within acceptable limits. Therefore, the EUT working in ESS mode was able to work in adverse situations and environments, enhancing security and quality of supply.

Similarly to the battery, techno economic analysis of the power converter was carried for different module sizes and total capacities. The modular power converter was more economic above 12 MVA and offered many potential advantages from 2-12 MVA. A comparison of the key features of the modular and classical power converter topologies is shown in Table 1.

Table 1: Topologies comparison

Overall conclusions from the development and testing of the power converter were:

A new multi-modular topology of a power electronics converter (inverter) has been developed which is needed to convert the battery DC output voltage to AC voltage to interface with an AC distribution network.  The converter topology is more complex than conventional technologies, but should theoretically offer some benefits for some battery and relatively high power applications above about 2 to 12 MVA:

  • Provide higher flexibility to connect different batteries to a grid
  • Meet technical requirements according to relevant codes and standards
  • Maximise availability
  • Minimise costs

The inverter has been developed to a prototype level (TRL 6/7), and various functions were tested successfully, although there are still numerous functions to be implemented and tested, as detailed in the report.   It is recommended therefore to further improve the inverter, in particular the control system, to achieve all the required functions and to meet the relevant international standards.  It is also recommended to carry out a scientific calculation of the availability of the new converter topology and compare with conventional topologies.


Consortium Websites

CEST logo

E-on logo Fuma-Tech logo
GreenPower logo DNV KEMA logo
University of Seville logo University of Southampton logo


Funded by

FP7 Logo