ABERTAX TECHNOLOGIES Ltd.

Abstract

The authors came up with a concept of turning existing grid connected PV systems to run independent of the grid if and when required. This can be done using a smart battery which includes a battery management unit and a converter. However, in order to limit the size of the battery required and ensure electricity supply at all times, the authors propose the use of a micro CHP to assist the battery and supply the heat demand during the winter season. Three grid connected houses with a PV system installed, were monitored over a three year period. This paper will show that the integration of a Smart Battery and a single phase CHP will provide the energy requirements for these houses all year round. The design of the battery capacity and sizing of the CHP for each house will be addressed in this paper. The authors made use of a wireless battery management to control the smart battery energy storage and the CHP.

Introduction

In most households the grid has always been the only source of electrical energy due to its availability, reliability and stability. A typical domestic installation usually consists of a light circuit and a power circuit. The capital and running cost of using the grid has, in most cases, been the cheapest way for providing the household energy demands whenever they are needed. However, over the last decade, renewable energy sources in domestic buildings have increased substantially. In the first 10 years of this century the increase in renewables was mainly due to subsidies and attractive feed in tariffs. However, in the last 4 years, the drop in the price of photovoltaic panels has made PV very attractive, especially in countries with high solar insolation such as Malta [1]. In Northern European countries, grid connected combined heat and power systems (CHPs) are also on the increase due to their high efficiency. Wind is not so popular in household applications due to a number of well known reasons. Most of the renewable technologies installed so far are grid connected and assume that the grid can absorb the excess energy that is instantaneously generated. However, the use of the grid as a dump store will be eventually limited due to problems resulting from high injection of energy into the grid at the same time. Figure 1 shows a particular grid connected PV installation which is experiencing repetitive shutdowns due to the excessive high voltage of the grid resulting from excessive energy injection.

Figure 1: Loss of PV energy due to high voltage on the grid

In southern European countries this problem is not encountered so often, due to the fact that the summer peak energy consumption peaks at the same time with the peak energy generated from the PV mainly due to the airconditioning loads. This was evident from a study[2] that the authors did in 2011.

Figure 2: Maximum instantaneous peak power demand on a hot summer

day and the effect of a 65MWp PV plant (August 2011)

The blue curve in figure 2 shows the hourly power demand on the Malta Power Station during the particular day where the absolute peak power was reached. A peak power of 413MW was reached during the year 2011. The green curve shows the instantaneous power that would be generated from 65MWp of PVs and the red curve shows the net power demand from the power station in the presence of such PV power. One can clearly see that 65MWp of PVs would shave the peak almost completely thus reducing the power station capital investment required to source this peak, that is expected to increase in due course. Apart from this, open cycle gas turbines are usually used to cope with this peak and hence the running cost is also very expensive making photovoltaics and battery energy storage profitable during this time. At the moment the PV installations in Malta have a generating peak of around 35MW which will be increasing by approximately 10MW per year. Therefore, the problems related with excessive renewable energy injection are bound to become more common in a couple of years, especially during spring time, when the photovoltaic energy reaches its absolute peak while the air-conditioning load is not that high. The only solution for this problem is to use or store the excessive energy for self-consumption at some other time of the day.

In fact, in certain countries grid operators are trying to push customers for self-consumption of the renewable energy that they generate. This usually involves the use of a hybrid inverter with a battery storage which tends to bring some of the existing components in the installation redundant. In Germany for example, financial support is given for the installation of self-consumption systems and this has created a market for domestic battery energy storage. Other markets for battery energy storage exist also in countries or regions where the grid is unreliable. In these areas the tendency is to have a standby battery which powers an inverter to supply the critical loads such as the fridge and lighting in the case of a power failure.

Photovoltaic Energy Generation in Domestic Case Studies

Three domestic installations were selected for this study, representing three different sizes of families and houses. In Malta all houses have a flat roof and therefore, most of the installations can easily be mounted at the optimum angle and direction.

Household 1 is a semi-detached villa with a floor space of 290 m²and has a grid connected PV system of 2.82kWp. A family of six people live in this house which is equipped with all the necessary electrical appliances including a dishwasher, 2 washing machines, electric water heater and air conditioning. Figure 3 shows the monthly and running average of PV energy when compared to the total electrical energy consumed. The graph also shows the amount of grid import and export for the different seasons during two and a half years. The heating in winter was assisted by gas fired heaters during winter 2012. The total gas consumed over this winter was approximately 6 cylinders, each with an energy content of 150kWhr (i.e 900kWhr of heat from gas). As can be seen from Figure 3, during this winter the net electrical import from the grid was approximately 1700kWhr while that of winter 2012 was approximately 950kWhr. From a master research at the University of Malta it was concluded that due to the high relative humidity during the winter season, the heating required is much higher. Therefore, an experiment was conducted in this house during last winter. Five dehumidifiers were spread around the house and set to control the humidity to 60%. This has obviously resulted in an increase in the electricity consumption. The grid import last winter has eventually increased from 950kWhr to 1700kWhr, however the use of the 6 gas cylinders normally consumed to assist the heating were not required. This means that by controlling the humidity, a net energy saving of 150kWkhr was possible. It is clear that for this house the PV generates around 80% of its electrical energy needs and the intention is to increase the system by four PV panels of 240Wp each, to reach the yearly needs generated from PV. One can also conclude that storing the seasonal excess energy into a battery storage is not feasible as this peaks to 1100kWhr and will be even higher when the 4 additional panels are added to bring the yearly average to 100%.

Figure 3: Percentage of PV generated relative to the energy consumption for household 1

Household 2 is an elevated maisonette with a floor space of 151 m²and a PV system of 1.38kWp. A family of four people live in this house with most of the appliances being electric as well, but without air-conditioning. However, the heating in this house is mainly done through a wood burning fire place rated at 12kW. During the winter season a gas fired heater is also used to assist with the heating and during the winters considered here, four cylinders were used every year (i.e. 600kWhr of heat from gas). Water heating was electric till March 2012, after which a solar water heater was installed. The electrical heating element for this heater was changed from 3kW to 1.5kW around February 2013. During the winter of 2013, this water heater, which supplies two bathrooms that are used at different times during the day, was switched on for a total of 36 hours. These actions have already brought this household above 100% renewable for its yearly electrical energy needs as can be seen in Figure 4. Though the yearly electrical energy needs are supplied from the PV system installed, the grid plays a very important role in absorbing the excess energy during spring and summer while delivering the necessary energy during autumn and winter. Again here the seasonal fluctuations cannot be stored in a battery storage as peaks reach 450kWhr.

Figure 4: Percentage of PV generated relative to the energy consumption for household 2

Household 3 is also an elevated maisonette with a floor space of 133 m², initially having a 1.38kWp PV system. A family of 3 people lives in this house with most of the appliances being electrical including air-conditioning. Three gas cylinders (450kWhr) were used to assist with heating during the winter of 2012. The family had a newborn in 2012 which resulted in increased activity due to the mother staying at home with the child. One can notice that the energy fluctuations here increased and the percentage went down as expected. The large dip in the winter of 2012 was partly attributed due to the regular use of an electrical halogen heater. At the end of August, two PV panels of 240Wp each, were increased to bring the total kWp to 1.92kWp (8x240Wp).

Figure 5: Percentage of PV generated relative to the energy consumption for household 3

A 1.4kW air conditioner with inverter was used sparingly during the winters of 2012 and 2013. During the winter of 2013 this house also tried the same experiment as household 1 and two dehumidifiers were added. The result was identical to household 1 that is, the 3 gas cylinders which were normally used were not required during winter 2013 and the use of the halogen heater was reduced. As can be seen in Figure 5, apart from eliminating the use of the gas heaters the electrical energy consumption was actually less than the winter of 2012.

From the above case studies it was concluded that while the yearly electrical energy needs can easily be supplied from a few PV panels, it is very evident that a battery energy storage to store all the seasonal energy fluctuations for self-consumption use is not feasible. One can also notice that even in a warm southern European countries such as Malta, the electrical energy demand in winter, due to the heating load which is mostly electric is quite substantial. Based on these studies, the proposed system that the authors present in this paper recommends the careful design of a scalable, non-invasive setup, that would eventually make use of a combination of, a grid PV system, a smart battery energy storage, a smart heating/cooling load and a CHP.

System requirements and design criteria

Today there are a number of households that have a grid connected PV system and/or a CHP installed. In both cases they tend to rely on the grid to ensure that;

i) the PV can generate the maximum power at all times to harvest all the maximum possible energy

ii) The CHP runs at its maximum efficiency operating point

In both cases the grid is assumed and used as an “infinite storage system” with the grid operators having to adjust their energy supply to cope with the variations caused. This situation has a limit especially when such decentralised sources of energy start to approach a high percentage of the total grid energy demand. The authors have designed a system to control this by a scalable and non-invasive system that will utilise any existing grid connected generators. This can be done by the addition of a smart energy storage setup and a smart load that can be controlled from a master management and control unit.

Figure 6: A system using the smart battery storage for operation with and without the grid

The proposed system will utilise existing grid connected components and will be able to operate with and without the grid. The operation of the system will obviously become more critical to manage and control without the grid, as the “infinite storage” is not present in the system any more. When designing such a system it is very important to consider the following three main parameters;

1. The yearly energy consumed by the consumer

2. The seasonal energy demands required

3. The maximum instantaneous power demand

As can be seen in the case studies above all three houses have large seasonal variations which cannot be stored in a domestic battery energy storage. In case study 2 the yearly energy generated from the PV system is more than what is required, however, the seasonal energy demand fluctuations are very high. So, it is very clear that while parameter 1 is easily achieved with a few PV panels (in Malta), coping with parameter 2 and even more parameter 3 are design issues which make it much more difficult and expensive to achieve without the presence of the grid. It is very clear that in order to run the system without the grid, battery energy storage is a must. The question is “what is the capacity of the storage required?”

To satisfy design criteria 2, it very clear that even in southern European countries where the weather is warmer, a CHP or some other form of heating is required. The authors propose the use of a CHP and heat pumps as this will drive the efficiency of the combined setup over 100%. The presence of a self-excited CHP makes it possible to satisfy the heat demand and partly assist in meeting design criteria 3; i.e. to supply part of the load during peak power demands which tend to occur during evenings in winter. The proposed setup has three sources of energy apart from the grid and in order to control the peak power (in the absence of the grid), it is important that the loads during the day are managed properly and efficiently. While the PV source is limited by the instantaneous maximum power being generated during daylight, both the CHP and the energy storage can usually cope with a relatively high peak power at any time of the day. So peak evening loads should be limited to the peak power possible from the battery storage and the CHP. This means that rather than a high energy capacity, the battery storage and its inverter should be able to handle a high discharge current. The energy into the battery can be replenished at other times when the load demand is not high. In fact the maximum energy demand in a day ever registered in the above cases was of 24kWhr during a cold winter day. This means that, if we assume that on such a day no PV energy could be harvested, a 2kW CHP running for 12 hours a day would supply the energy demand and therefore provide a very warm cosy environment. Limiting the absolute peak power to 6KW and assuming a 48V battery bank would result in a peak power from the batteries of 4kW (assuming the CHP is on and delivering 2kW). Therefore, assuming that the battery is supplying the load via a 4kW inverter with an efficiency of 90%, the peak battery current demand would be 93Amps. Assuming this peak power to occur for a maximum of 2hrs and a minimum state of charge of 20%, the battery energy storage capacity would be 10kWhr. Therefore, for the above 3 cases the luxury of supplying the complete house with electricity at any time with and without the grid would be met by the use of a 2kW/6kW (electrical/heat) CHP and a 48V, 250Ahr battery with a 4kW inverter. Apart from the advantage of ensuring electricity at all times the management system can be programmed to make use of the most economical source of energy. For example during peak hours the grid might be more expensive than running the CHP or discharging the energy from the battery storage.

System Efficiency Considerations

Having calculated the power and energy requirements, it is important to consider the overall system efficiency at the design stage. A high quality charger and inverter usually operate with an efficiency of 90% to 92%. The battery efficiency is usually considered to be around 80%. The efficiency of the CHP can vary from 0% (running idle without any demand) to 94% at its maximum efficiency operating point. With this in mind it is important to note that:

a. It is very important to keep the source to load efficiency as high as possible,

b. When used, the CHP should run at the optimum power and any excess electricity and heat should be stored or consumed.

c. In each voltage conversion 8 to 10% of energy is lost

d. Storing energy in the battery would mean a loss of 20%

e. The maximum PV energy should be harvested at all times.

The above facts are very important when designing a system and its management. For example charging the batteries from a PV system should, if possible, be avoided as this limits the maximum power transfer, since the charging profile and the state of charge of batteries will dominate the flow of power. The battery should always be designed and used as a buffer and not as a component in the energy flow process. Figure 7a and b show the difference in the PV/Solar to load efficiency with the same components but a different connection. If one considers the sun as the source (the real energy source), the values will obviously become much lower due to the PV panel conversion efficiency (15%-20%).

Figure 7: Difference in efficiencies from PV/Solar to load

The important observation here is the difference in the energy efficiency values. Sometimes customers try to get a higher solar energy harvest through the use of PV panels with slightly higher efficiency for which they have to pay dearly. (A standard panel with 15% efficiency will cost half the price of a panel with an efficiency of 20%, i.e. paying double the price for 33% extra). It would be really a costly waste if the efficiency gained from the panel will then be lost completely due to the connection and management of the system. One can also observe that in Figure 7b, the charger has been changed to an inverter giving the added advantage that apart from charging, the inverter can source the power to the load thereby utilising better the equipment cost while meeting the maximum instantaneous power demand discussed earlier.

Experiment Setup

In their experimental setup represented by the block diagram of Figure 8 the authors made use of a standard 1.2kVA grid connected inverter, a purposely built CHP capable of delivering 2kVA electrical / 6kVA heat [3], a 48V, 50Ah lead acid gel battery pack (i.e. 1.92kWh with 80% DoD), a variable load (smart load), an LED lighting circuit and a power circuit.

Figure 8: Block diagram of experimental setup

In view of the above discussions, the authors concluded that the battery should and can actually be the heart of the system, as it has to manage the flow of energy and act as an ideal reservoir of energy that is always there when needed but at the same time ensuring the highest efficiency of the complete system. The role of the battery management will become extremely important especially when the grid is off. To achieve this, the authors made use of the Abertax wireless battery management system for the efficient management of the power flow to and from the battery by also controlling the various energy sources using the wireless version. This is a novel product that the authors have developed at Abertax Research and Development to make the battery play the ideal role.

The Abertax e²BMS is much more than just a battery monitor. It can transmit and receive data through an internet connection via GPRS. However, it has also an embedded RF transceiver that can communicate and control any other device as shown in Figure 9. This feature can be used to control, the ON/OFF operation of the CHP, the smart loads and the ON/OFF of each individual appliance via the RF transceiver.

Figure 9: Abertax e²BMS system used for smart management

Due to the development of light emitting diode (LED) technology it was decided that the lighting should be powered directly from the battery since they are in any case powered from DC and this has the advantage of eliminating the AC/DC conversion loss. Apart from this loss, research at the University of Malta has shown that most of the LED lighting fixtures actually fail due to the embedded converter being too close and subjected to the heat emitted by the LEDs. Based on this research, the authors have designed and developed the ideal LED luminaires and together with Abertax Kemtronics, they are already producing novel luminaires that are directly powered from either 12V or 24V DC.

The CHP has also been developed by the authors as they could not find an ideal, off the shelf product that would fit their needs. One of the drawbacks in most CHPs available on the market is the fact that they cannot be switched on if the grid is lost. Another drawback is the size and related to it, is also the price. The PV inverter is a standard 1200 VA grid inverter from SMA and here one has to point out that in order to use this in the absence of the grid, the battery and system management plays a very important role as it has to ensure that the maximum energy that is automatically delivered from the PV system has to be absorbed at all times either by the battery or by the loads. In the case that the battery becomes fully charged the use of smart loads become necessary. The smart loads have to be controlled by the management system to absorb any excess energy. The smart loads are controlled variable loads and are ideally heating and/or cooling loads. The easiest smart load is a thyristor controlled water heater which is used to store any excess energy as heat in water. Heat Pumps are also ideal type of smart loads especially in warmer countries where the PV peak coincides perfectly with the load demand for the cooling required. The high coefficient of performance (COP) of modern inverter controlled heat pumps render them ideal for use in the above setup. Even in winter, the fact that ideally the CHP should run at its peak constant power (to ensure the highest efficiency), a heat pump can be used to convert any excess electrical energy also to heat. In such a case one should note that the combination of a CHP and heat pump/s would be running at over 100% efficiency. Assuming the CHP used in this paper and a heat pump with a COP of 3, the net heat output will be 12kW (i.e. 2kW electric x 3 + 6kW heat). If we assume that the CHP runs at 90% efficiency this combination will therefore result in 135% efficiency.

Results

Simulations of various instances were carried out on the test rig that was built in the laboratory. The time graph in Figure 10 shows the results that were achieved. The test started with the system connected to the grid and power was being consumed to charge the batteries. At T=1min, the PV inverter was started and this is evident from the reduction in power consumption from the grid. A resistive load is connected to the system at T=2min, since PV power is at its maximum the extra power is taken from the grid source. Upon disconnection from the grid at T=3min, the battery current reverses so that power for the load is taken through the DC-AC inverter which was previously serving as a charger. At T=4min, the CHP is connected and the PV power is disconnected, this is done since the battery inverter is at full load and will trip if PV power is lost. Finally at T=5min, the battery current is reduced drastically by increasing the throttle of the CHP to provide most of the electrical power that is required.

The below time graph gives a clear overview of how the domestic power can be sourced from different sources so as to maintain electricity supply even during grid power loss. The graph, however, does not show the transients during switching from one source to another, this is important especially when connecting the CHP.

The developed CHP employed the use of an asynchronous generator for various reasons [3]. During this research, however, it was noted that high inrush current may develop while connecting the generator. This results in a problem if performed in the absence of the grid since it may trigger an overcurrent fault in the DC-AC battery inverter as seen in Figure 11. A work-around is to purchase a larger inverter which can withstand this transient, however, this increases the system costs.

Figure 10: Smart Battery System Test Result

The authors have developed a synchronising algorithm that considers the phase difference of the generator and inverter output waveform before switching on. Figure 12 shows the current demand when the synchronising algorithm is used. As can be seen the starting current demand is now much less and the CHP could be phased in smoothly without disturbing the system operation.

Figure 11: Connection of Asynchronous Generator without synchronization

Figure 12: Connection of Asynchronous Generator with synchronization

Conclusion and Future Work

The authors have shown that a smart non-invasive and scalable domestic system is the way forward to ensure a reliable and economical way of becoming self-sustainable. The battery storage plays a very important role and should be the heart of the system to ensure optimum performance and efficiency. The authors will continue their work of integrating their wireless battery management to control the whole system. The latter involves the research of intelligent algorithms that should be able to deal with any possible energy flow scenario and the ideal use of the energy mix from the three available sources.

Acknowledgements

The authors would like to thank Abertax Technologies for their financial support in this ongoing interesting research work. Particular thanks go to the founder and chairman of the Abertax Group, Mr. Werner Schmidt and vice–chairman Dr. Martin Florin who have been the motivators of this research work with their aim to ensure a better environment for future generations.

References

[1] Joseph Cilia, Klaus Dieter Merz, MalcolmTabone, Neville Azzopardi, Björn Mentzer, “The Future of a Smart Battery Energy Storage”,13th ELBC, Paris, 25th-28th September 2012

[2] Joseph Cilia, Mark Scicluna, Neville Azzopardi, Klaus Dieter Merz, Björn Mentzer, ”The potential of battery energy storage for grid connected domestic renewable sources of energy” 6^th International Renewable Energy Storage Conference 22-24^th November 2011

[3] Matthew Schembri, David Zammit, Joseph Cilia, ” The Design of a Smart Micro-CHP Combined with Energy Storage” 8^th International Renewable Energy Storage Conference November 2013

Joseph Cilia (joseph.cilia@um.edu.mt, joseph.cilia@abertax.com), Matthew Schembri (msch0037@um.edu.mt) ,Malcolm Tabone (malcolm.tabone@abertax .com), Mark Scicluna (mark.scicluna@abertax.com) , Neville Azzopardi (neville.azzopardi@abertax.com), KD Merz (KD.merz@abertax.com)