Analysis Of Grid-Connected Photovoltaic Systems
Conventionally, grid connected photovoltaic energy conversion systems are composed of a DC-DC converter and an inverter. The DC-DC converter is controlled to track the maximum power point of the photovoltaic array and the inverter is controlled to produce current in such a way that the system current has low Total Harmonic Distortion (THD) and it is in phase with the utility voltage. The efficiency of the conventional system is low because the DC-DC converter and the inverter are connected in series. The purpose of this article is to design a photovoltaic generation system for connection in a three-phase system using only a DC/AC inverter...
- Dr. L Ashok Kumar
Photovoltaic (PV) energy has great potential to supply energy with minimum impact on the environment, since it is clean and pollution free. A large number of solar cells connected in series and parallel set up the photovoltaic or solar arrays. One way of using photovoltaic energy is in a distributed energy system as a peaking power source.
On the other hand, strict regulations have been applied to the equipment connected to the utility lines. Some of these regulations are related to harmonics distortion and power factor. However, with the development of power electronics, many equipment tend to increase the levels of harmonic distortion. The line current at the input to the diode bridge rectifier deviates significantly from a sinusoidal waveform and this distorted current can also lead to distortion in the line voltage. Moreover, many modern equipment use digital controllers, based on microprocessors sensitive to variations in the voltage and current waveforms. Therefore, to increase the PV system utilisation the power conversion can be designed to also provide functions of a unified power quality conditioner.
The utilisation of two DC/AC fully controlled converters make the system have the most versatile structure of converters applied as energy conditioner. In this case, depending on the controller, the converters can have different functions of compensation. For instance, they can realise active series and shunt filters combined to compensate simultaneously load current and harmonics of the supplied voltage. In this way, the equipment is called Unified Power Quality Conditioner (UPQC). An active shunt filter is a suitable device for current-based compensation. This configuration includes current harmonics and reactive power compensations. The active shunt filter can also balance unbalancing currents.
The active series filter is normally used for voltage-based compensation. In this case, voltage harmonics and voltage sags and dips are compensated. Other applications can be found in literature for purposes of compensation of the fundamental frequency, such as reactive power compensation, flux control of active power and voltage regulation. In this case, it is called Unified Power Flow Controller (UPFC).
Conventionally, grid connected photovoltaic energy conversion systems are composed of a DC-DC converter and an inverter. The DC-DC converter is controlled to track the maximum power point of the photovoltaic array and the inverter is controlled to produce current in such a way that the system current has low Total Harmonic Distortion (THD) and it is in phase with the utility voltage. The efficiency of the conventional system is low because the DC-DC converter and the inverter are connected in series. The purpose of this article is to design a photovoltaic generation system for connection in a three-phase system using only a DC/AC inverter.
The proposed system increases the conversion efficiency and also provides useful function any time, operating as power supply as well as harmonic and reactive power compensator when the sun is available. At low irradiation, the system operates only as harmonic and reactive power compensator. Other DC/DC converter is used to provide voltage harmonic compensation. Cost estimation shows that the use of additional components increases the cost in less than 12% to have another function to improve power quality. Also, this converter does not change the efficiency of the PV energy conversion since the converters are connected in parallel.
The control was implemented with the Synchronous Reference Frame (SRF) method. The system and controller were design and simulated. Different Pulse-Width-Modulation (PWM) techniques have been compared to suggest a configuration with optimal efficiency. The system provides approximately 2.8 kW of photovoltaic generation.
The use of photovoltaic (PV) systems as a safe and clean source of energy from the sun has been rapidly increasing. The application of PV systems in power systems can be divided into two main fields: off-grid or stand-alone applications and on-grid or grid-connected applications. Stand-alone PV systems can be used to provide power for remote loads that do not have any access to power grids while grid-connected applications are used to provide energy for local loads and for the exchange power with utility grids.
The first large grid-connected PV power plant with 1 MW capacity was installed in Lugo, California, USA. The second plant with 6.5 MW capacity was installed in Carissa Plains, California, USA. Currently, many large grid-connected PV systems with different ranges of power are operating in various countries.
PV systems can enhance the operation of power systems by improving the voltage profile and by reducing the energy losses of distribution feeders, the maintenance costs, and the loading of transformer tap changers during peak hours. Nonetheless, in comparison with other renewable technologies, PV systems still face major difficulties and may pose some adverse effects to the system, such as overloading of the feeders, harmonic pollution, high investment cost, low efficiency, and low reliability, which hinder their widespread use. Moreover, variations in solar irradiation can cause power fluctuation and voltage flicker, resulting in undesirable effects on high penetrated PV systems in the power system. Some control methods, such as Maximum Power Point Tracking (MPPT) can be used to improve efficiency of PV systems.
In such controllers, both the produced voltage and the current of the PV array should be controlled. This may complicate the PV system structure with increased possibility of failure while tracking maximum power in unexpected weather conditions. With respect to system protection scheme, the PV system-based Distributed Generations (DGs) should energise the local loads after the system has been disconnected from the utility grid during faulty conditions.
In these situations, any unintentional islanding may increase the risk of safety problems or damage to other parts of the system components, which can decrease system reliability.
These problems mean that accurately analysing the effects of installing large grid-connected PV systems on the performance of the electric network is necessary.
This evaluation is important because it can provide feasible solutions for potential operational problems that grid-connected PV systems can cause to other components in distribution systems.
In the literature, many works focus on steady-state modeling and analysis of PV systems. However, no attempt has yet been made to study the effects of grid-connected PV systems on the dynamic operation and control of the system before real-time implementation.
PV System Modeling
High-penetrated grid-connected PV systems, which are known as a type of DG in the megawatt range, are rapidly developed. These cover the majority of the PV market in different countries worldwide.
The main components of a grid-connected PV system includes a series/parallel mixture of PV arrays to directly convert sunlight to DC power and a power-conditioning unit that converts DC power to AC power; this unit also keeps the PVs operating at maximum efficiency. Figure 1 shows the general diagram of grid-connected PV systems.
Notably, in many cases, energy storage devices such as batteries and super-capacitors are also considered the third component of grid-connected PV systems.
These devices enhance the performance of PV systems, such as power generation at night, reactive power control over the PV systems, peak load shifting, and voltage stabilising of grids.
To provide proper interface between grid-connected PV systems and the utility grid, some conditions must be satisfied, such as phase sequence, frequency and voltage level matching. Providing these conditions strongly depends on the applied power electronics technology of PV inverters.
Figure 1: Simplified diagram of the grid-connected PV system...
Figure 2: Equivalent circuit of the PV module...
The electric characteristics of a PV unit can generally be expressed in terms of the current-voltage or the power-voltage relationships of the cell.
The variations in these characteristics directly depend on the irradiance received by the cell and the cell temperature.
Therefore, to analyse the dynamic performance of PV systems under different weather conditions, a proper model is required to convert the effect of irradiance and temperature on produced current and voltage of the PV arrays.
Figure 2 shows the equivalent electrical circuit of a crystalline silicon PV module. In this model, I is the output terminal current, IL is the light-generated current, Id is the diode current, Ish is the shunt leakage current, Rs is the internal resistance, and Rsh is the shunt resistance.
In practice, the value of Rs strongly depends on the quality of the used semi-conductor. Therefore, any small variation in Rs value can dramatically change the PV output.
Possible Effect Of Grid-Connected PV Systems On Distribution Systems
Renewable energy sources, especially PV systems, have become more significant sources of energy, attracting considerable commercial interest. Nonetheless, the connection of large PV systems to utility grids may cause several operational problems for distribution networks.
The severity of these problems directly depends on the percentage of PV penetration and the geography of the installation. Hence, knowing the possible impact of large grid-connected PV systems on distribution networks can provide feasible solutions before real-time and practical implementations.
The aim of this section is to introduce possible effects that PV systems may impose on distribution systems. Inrush Current. The small inevitable difference between PV systems and grid voltages may introduce an inrush current that flows between the PV system and the utility grid at connection time, and decays to zero at an exponential rate. The produced inrush current may cause nuisance trips, thermal stress, and other problems.
Grid Connected Photovoltaic System
The proposed PhotoVoltaic (PV) energy conversion system has high efficiency, low cost and high functionality. Figure 3 shows the block diagram of the proposed system. The converter 1 (PV converter) in Figure 3 is responsible to convert the PV energy to the grid as well as to compensate current harmonics and reactive power. The converter 2 (Dynamic Voltage Restorer — DVR converter) in Figure 3 is responsible to compensate voltage harmonics or voltage sags.
Figure 3: PV generation with UPQC function...
Figure 4: Conventional load with voltage minimum at end of line...
The utilisation of two controlled converters makes the system to have the most versatile structure applied as energy conditioner. In this case, depending on the controller, the converters can have different functions of compensation.
For instance, they can realise active series and shunt filters combined to compensate simultaneously load current and harmonics of the supplied voltage.
Safety is one of the major concerns in PV systems due to unintended islanding at the time of fault occurrence at the grid side. Here, PV systems continue to feed the load even after the network is disconnected from the utility grid, which may lead to electric shock of workers.
PV systems usually are designed to operate near unity power factor to fully utilise solar energy. In this case, the PV system only injects active power into the utility grid, which may change the reactive power flow of the system.
Therefore, voltages of nearby buses can be increased because of the lack of reactive power. The produced over-voltage can have negative effects on the operation of both the utility and customer sides. Output power fluctuation, The fluctuation of the output power of PV systems is one of the main factors that may cause severe operational problems for the utility network. Power fluctuation occurs due to variations in solar irradiance caused by the movement of clouds and may continue for minutes or hours, depending on wind speed, the type and size of passing clouds, the area covered by the PV system, and the PV system topology. Power fluctuation may cause power swings in lines, over- and under loadings, unacceptable voltage fluctuations, and voltage flickers.
Output Power Fluctuation
The fluctuation of the output power of PV systems is one of the main factors that may cause severe operational problems for the utility network. Power fluctuation occurs due to variations in solar irradiance caused by the movement of clouds and may continue for minutes or hours, depending on wind speed, the type and size of passing clouds, the area covered by the PV system, and the PV system topology. Power fluctuation may cause power swings in lines, over- and under loadings, unacceptable voltage fluctuations, and voltage flickers.
Harmonic distortion is a serious power quality problem that may occur due to the use of power inverters that convert DC current to AC current in PV systems. The produced harmonics can cause parallel and series resonances, overheating in capacitor banks and transformers, and false operation of protection devices that may reduce the reliability of power systems.
Frequency is one of the more important factors in power quality. Any imbalance between the produced and the consumed power may lead to frequency fluctuation. The small size of PV systems causes the frequency fluctuation to be negligible compared with other renewable energy based resources. However, this issue may become more severe by increasing the penetration levels of PV systems. Frequency fluctuation may change the winding speed in electro motors and may damage generators.
Limits Of Grid Transmission Capacity
Conventional design of a power grid considers a load flow directed from the transformer to the load. Passive loads with sinusoidal currents have been assumed for the rating of transformers and distribution lines. Figure 4 shows the voltage decreasing with the distance from the transformer.
Therefore, the design is usually made to keep the voltage at the transformer above the nominal voltage in order to achieve a voltage drop which is below the minimum specified value. In the last few years the usage of distribution grids has changed heavily as many devices are using uncontrolled bridge rectifiers at the mains input side. In many rural areas large decentralised power generation (e.g. photovoltaic, wind, micro turbines and combined generation) has been installed. In some areas the installed generation power is significantly higher than the consumption and often reaches the rated grid power. Due to high levels of generated power from decentralised generation stations the load flow may change its direction. Particularly in high solar gain periods, when solar plants feed their highest power levels into the grid, while the power consumption can be fairly low, reverse power flow may occur. Therefore, solar generated power is fed into the medium voltage grid over the transformer of that branch. If the power is in the range of the nominal power of the branch, the voltage at the connection point of the generation plant may significantly increase. If the voltage exceeds the tolerance of usually 10% above nominal voltage, other devices and equipment might be damaged.
Figure 5: Voltage maximum or minimum at the end of line...
Figure 6: Data acquisition and control structure...
Figure 5 shows the possible voltage variation with the distance from the transformer for different load and generation conditions. Therefore, the design is usually made to keep the voltage at the transformer above the nominal voltage in order to reduce voltage drops below the minimum specified value.
With decentralised generation the voltage may increase at the connection point as shown in Figure 5. With the voltage at the transformer being set above the nominal value it is very likely to exceed the specified maximum voltage. In Germany, a maximal voltage increase of 2 or 3% in the future caused by distributed power plants in low voltage grids is recommended. In case of reverse power flow the maximum permitted voltage will be reached even below nominal power of the grid branch.
Therefore, the grid needs to be improved to offer new services and new functionality to deal with the new requirements. Avoiding high installation or operating costs promotes further growth in decentralised power generation. In the past grid extension was required to increase the transmission capacity, resulting in additional cabling and higher investment cost, even if the additional capacity is only being used for a few operating hours per year, usually on solar gain days, when additional grid capacity is actually needed. In the short term, additional connection of solar generation systems can often not be permitted until grid extension had been carried out.
Increasing Voltage Quality And Grid Capacity
While the grid capacity and grid quality have primarily been provided by network expansion so far, this project aims to use the installations which are distributed in the grids effectively. This is done by the use of distributed measurement technology, intelligent control of power electronics, new information and communication technology and the possibilities of the grid control. The concept is developed and tested on the example of distributed PV systems. However, the use is not restricted to this application. In all networks with controllable feed-in installations and loads network efficiency can be increased by distributed network services.
The operational status of the grid has to be measured continuously at connection points of large loads and decentralised generation. Solar inverters are equipped with data acquisition capabilities because they need to synchronise their voltage and frequency to the grid voltage. For load connection points measuring technology is to be installed. As shown in Figure 6 a main computer is networked to a number of data acquisition devices and solar inverters. Data acquisition devices and solar inverters monitor voltage, current and power flow at their locations on the grid. Data acquisition devices are located at large loads (e.g., industrial plants) and grid nodes. The main computer receives the grid status data and then calculates the values for the required reactive power for the individual solar inverters that will be sent over the data network to the inverters.
The control structure consists of three different controls. The first part is the limitation of the grid voltage by reactive power absorption of the inverters. To avoid unnecessary losses only as many inverters as needed have to absorb only as much reactive power as needed to limit the grid voltage. Thus, the main computer only activates the inverters with the highest voltage levels in the grid. Additionally, voltage fluctuations due to fast load and generation changes e.g., moving clouds can be compensated and smoothed by injecting and absorbing reactive power through the solar inverters. The inverters can also be used for local compensation of reactive power required by other loads in order to minimise power losses in the grid.
Figure 7: Voltage drop at a line when feeding in active (left) as well as active and reactive power (right)...
Solar inverters above 8 to 10 kW are usually connected by three phases to the grid. They can operate in all four quadrants thus being able to inject or absorb reactive power while active power is fed into the grid. Figure 7 shows in a qualitative way the voltage drop at a transmission line. While the voltage at the end of the line U2 is lower than the voltage U1 at the beginning (transformer side) in case of normal load conditions, this changes when active power is fed in at the end of the line (left part of Figure 7).
The voltage may be significantly higher at the end of the line than at the transformer. By additionally absorbing reactive power (or current) the overvoltage can be decreased (right hand side of Figure 7). This is also the case in low voltage distribution grids with a relative high R/X ratio especially when taking the transformer impedance into consideration.
The reactive power flow results in an additional current that has to be driven from the inverter. Studies on the reactive power have shown that a minimum power factor of cos y = 0.9 in typical low voltage grids is sufficient to keep the voltage within the permissible limits. A power factor cos y = 0.9 provides reactive power of 43% of the active power. This causes a 10 % higher current of the inverter. If the reactive power is only absorbed at increased voltage levels, the higher rating of the solar inverter may be lower or it may even not be necessary. If reactive power is used for limiting the grid voltage additional power losses are generated in the inverter and in the grid lines due to the higher grid current. But the benefit is that higher active power can be transmitted and the surplus solar generated electrical power can be fed in to the grid. Therefore, it is appropriate to provide the reactive power not by a static characteristic of the inverters, but to minimise the reactive power absorption by individually activating those inverters which have the most significant effect to the grid voltage. The communication of each inverter with a central computer ensures the optimisation of the reactive power absorption.
Figure 8: Voltage increase due to PV power plants...
Smoothing Of Voltage Fluctuations
Fluctuating power input to PV systems due to passing clouds or highly fluctuating loads cause voltage fluctuations in the low voltage grid. Reactive power consumption (capacitive) at negative voltage peaks and reactive power absorption (inductive) at positive voltage peaks by the distributed solar inverters can smooth voltage fluctuations in the grid. The risk of flickers can be reduced by such an additional control that is implemented locally in the inverters. The smoothing does not need any communication of the inverters with a central computer.
Reactive Power Compensation
Reactive power compensation to this date requires additional equipment and associated installation and commissioning costs which should be recovered by greater efficiencies. So far, compensation is mainly used in large industrial plants. Therefore, generating decentralised reactive power for compensation significantly lowers the power losses due to short transmission distances of the reactive power. For generating reactive power short term energy storage is required. This can be done with capacitors or inductors. Voltage link based solar inverters usually have capacitors, so the already installed capacity can be used for reactive power. The existing reactive power reserves which are inherently present by the distributed inverters can be used to provide reactive power to the overlaid middle voltage grid or to reduce the reactive power consumption of the low voltage grid to minimise the losses.
The field testing is done in a real low-voltage grid with a high penetration of PV power plants.
Overview Of The Test Grid
Figure 8 shows the structure of the test grid. The grid is fed by two transformers (rated power 630 kVA) and operated meshed. The installed PV system capacity is 400 kWp and is already higher than the average network load. On sunny days, active power is fed back regularly in the medium voltage grid. There are numerous relatively large PV power plants in the grid due to the high number of agricultural buildings with large roof areas.
Figure 9: Number of 10-minute averages depending on the active power flow of the test grid...
The voltage distribution and the loadings of cables and transformers were calculated by a commercial power system analysis software. Figure 8 also shows the voltage distribution in the grid area as a result of PV power plants. According to the VDEW recommendations, the voltages are calculated without loads and with the inverters feeding in their rated power. It is evident that in this grid a voltage increase < 2 % is observed only near the transformers. The increase is above 2% between the transformers and over 3 or 4 % at the critical network extensions. Despite the voltage increase the transformers and cables in the grid are loaded at 40%.
Figure 10: PV feed-in and voltage...
Data from both transformers have been available in 10 minute averages over a period of a year. Figure 9 shows the number of measured 10 minute averages depending on the reactive power flow of the grid. On sunny days the power generated by the PV power plants in the grid exceeds the load. Thus, there is an active power flow from the test grid to the overlaid middle voltage grid.
Two measurement points at the inverters of PV power plants were available to evaluate the state of the test grid in advance. One is at a PV power plant which is at the end of a critical long line and the other is located between the transformers.
The upper chart of Figure 10 shows the developing of the PV-feed-in in p.u. based on the rated power of the inverter, which was a sunny summer day. The rated power is not achieved because of the strong heating of the PV modules. The lower chart of Figure 10 also shows the corresponding voltages at both measurements points (green: measuring point at the critical grid extension, red: measurement point between the two transformers). The zero values of voltage and power are the result of short-term transmission errors in the measurement. The voltage profile follows the PV feed-in very well. The left transformer was out of service due to maintenance on this day. That is the reason why there are high voltage increases. These values correspond well with the results of the grid calculation. Figure 11 shows a close-up of the PV feed-in and the corresponding voltages, an unsettled day (green: measuring point at the critical network extensions, red: measurement point between the two transformers).
Figure 11: Detail of PV feed-in and voltage...
On this day, the left transformer was also out of service due to maintenance. The gradients of the voltage peaks or drops are usually smaller than the gradients of power peaks or drops. This is due to the distribution of the PV systems in the test grid. Thus, the power drops caused by passing clouds are staggered. These staggered power drops cause staggered voltage drops.
The largest power gradient measured so far is 0.07 p.u./s relative to the rated power. The largest voltage gradient measured is so far 0.002 p.u./s relative to the rated voltage. The concept described in this paper provides an improved voltage quality and higher transmission capacities in low voltage grids with a high penetration of PV power plants. The technology described above is currently under development and being tested with solar inverters on the low voltage grid. Generally speaking, the technology can be applied to any power electronic inverter which is either permanently or temporarily connected to the grid. Due to the inbuilt data communication and data acquisition facilities the system can be automatically configured after connecting a new inverter to the grid.
Dr. L Ashok Kumar
Dept. of EEE
PSG College of Technology,
If you want to share thoughts or feedback then please leave a comment below.