ON THE SPECIFICATION AND TESTING OF INVERTERS FOR STAND-ALONE PV SYSTEMS. J. Muñoz and E. Lorenzo - PDF

Description
ON THE SPECIFICATION AND TESTING OF INVERTERS FOR STAND-ALONE PV SYSTEMS J. Muñoz and E. Lorenzo Instituto de Energía Solar. Universidad Politécnica de Madrid (IES-UPM). Ciudad Universitaria s/n

Please download to get full document.

View again

of 37
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information
Category:

People & Blogs

Publish on:

Views: 15 | Pages: 37

Extension: PDF | Download: 0

Share
Transcript
ON THE SPECIFICATION AND TESTING OF INVERTERS FOR STAND-ALONE PV SYSTEMS J. Muñoz and E. Lorenzo Instituto de Energía Solar. Universidad Politécnica de Madrid (IES-UPM). Ciudad Universitaria s/n Madrid. Tel: Fax: Keywords: Stand-alone PV system, inverter, testing, efficiency, reliability, technical specification. SUMMARY Inverter features are reviewed from a PV systems perspective, with a view to contributing to possible codes, procurement specifications and testing procedures, in order to assure the technical quality of these systems. A laboratory testing campaign has been carried out on a representative set of sixteen currently available inverters and a set of the most common AC appliances. The results of the tests are discussed with the aim of divulging the particular features of operating AC appliances in PV systems and the provisions to be taken into account in PV system design. The development of testing procedures has followed the motto keep it as simple as possible, in order to make their application easier in conventional laboratories in developing countries. 1 1. INTRODUCTION The presence of inverters is becoming widespread in stand-alone PV systems for rural electrification purposes. In fact, national agencies in different countries, such as Thailand 1 or Brazil 2, have included inverters in their recent initiatives. To a large extent, this is happening because inverter technology is overcoming the barriers that have prevented their use in the past, namely: low reliability, high cost and poor conversion efficiency. Inverters allow PV systems to power conventional AC appliances, which represents an undeniable advantage in terms of delivered service and user satisfaction. However, standard AC appliances are designed for operating in the particular conditions of the conventional grid (well regulated voltage and frequency, low voltage harmonic distortion, large surge power capability, etc.), which are sometimes not maintained by inverters. This implies the risk of improper functioning, even damage, to the appliances, which should be taken into consideration in PV system design. Moreover, while the use of AC appliances connected to the grid is highly standardised and certified using internationally validated procedures, there are no equivalent standards, certifications and procedures available for PV inverters. This paper is intended to contribute to future technical standards for PV inverters. For that, a set of 16 commercially available PV inverters has been tested in combination with most common AC appliances. The results are discussed below in an attempt to assess the different aspects (power capability, voltage regulation, protections, etc.) of interest for general PV standards, and also for the procurement specifications issued by national agencies in the promotion of rural electrification programmes. Testing inverter activities carried out by other laboratories have been also taken into consideration To a great extent, this paper follows along the lines of the Universal Technical Standard for Solar Home Systems 6 and the consecutive testing campaigns of commercial PV components 7-9. It should be made clear that this particular approach for technical quality assurance does not involve the accreditation of testing laboratories by international certification bodies (ISO ) as a prerequisite. Instead, it is based on the idea that technical quality in PV rural electrification is more a matter of will than of technical sophistication. Rather simple local testing (in conventional laboratories of agencies, utilities, universities, etc.) can be a very effective technical quality assurance tool if it is clearly established in contractual agreements between vendors and customers. We should clarify that this ad hoc alternative neither tries to compete with accredited laboratories nor exclude them. Rather it is motivated by the persistence of technical problems in the field that, today, remain out of the scope of international standards. As a matter of fact, our laboratory, at the IES-UPM, despite not being internationally accredited, maintains a regular and increasing activity on PV components testing (PV generators, lamps, regulators, batteries and, now, inverters) upon request. Therefore, the motto keep it as simple as possible has been paramount in the definition of testing procedures, as well as for the selection of the required instrumentation. It is important to note that the users of PV inverters must be protected against electric shock, which is matter of vital importance as a result of the potential risks of AC voltages greater than 50 V. This can be done in different ways, for example, following the recommendations of the well-known IEC standard 11. However, simply for presentation reasons, their discussion takes place in a further paper. Here, we 3 have restricted ourselves to testing the inverters capability for activating the so-called Residual Current Devices, which are at the core of most protective schemes. 2. AC LOADS AND INVERTER REQUIREMENTS Standard characterization of AC loads assumes that they are operated by the grid, which is (or should be) close to an ideal voltage source, i.e., pure sinusoidal voltage, zero internal impedance, unlimited output current, etc. This approach has the advantage that it is independent of the voltage source and closely represents the load behaviour when operated by different utility grids. However, inverters are far from being ideal voltage sources because their output current is limited, and also, because their voltage waveform cannot necessarily be sinusoidal. As a matter of fact, square and quasi-square waveforms are also present in the current market (Figure 1) and can be acceptable in many practical situations. Therefore, AC loads can behave differently when they are operated by inverters. Furthermore, standard parameters, such as the power factor, PF, or the total harmonic distortion, THD 12 X (where x=v,i for voltage and current waveforms, respectively), can differ significantly under non-sinusoidal conditions. Figure 1. Voltage waveforms of different inverters found in the current PV market. (a) Sinusoidal. (b) Quasi-square. (c) Square. As an illustrative example, figure 2-a shows the starting of a domestic electric drill (induction motor) operated on the grid. It can be seen that the surge power (1800 W) is eight times greater than the rated one (225 W). Figure 2-b shows what happens when the same electric drill is powered by an inverter. The surge power is now reduced by a factor of 5.6 (1800/320), but the motor is correctly operated, although it takes more 4 time (1.3 versus 0.5 seconds). Hence, this inverter should be accepted in practice, despite its inability to provide the surge power associated to its grid operation. Figure 2. Starting of a domestic electric drill (induction motor) operated on (a) the grid, (b) an inverter. These operational differences between the grid and inverters limits the real effectiveness of possible inverter specification approaches based on standard parameters. For example, faced with a specific PV application (SHS, schools, etc.) one can imagine requiring something like: Inverter power capability must be X times larger than the rated power. However, this represents rather little help in anticipating whether the inverter is capable of starting a particular load, because the X value is only slightly related to its surge power under grid conditions. Therefore, this approach to inverter specification should be restricted to the case in which the elements making the load cannot be anticipated. Think, for example, of the provision of standard PV systems to a large number of schools whose AC equipment is from diverse origins. Then, it could be of interest to specify the surge inverter capabilities by simply referring to the inverter itself, and disregarding the load. It is worth mentioning that, for such cases, the GEF/World Bank specifies the following requirements 13, 14 : The inverter must operate safely at an ambient temperature of 25 ºC for a minimum of: a) four hours at full rated output power, b) one minute at 125% of the rated power, and c) two seconds at 150% of the rated output (to simulate high surge currents due to starting of motors). However, when all the particular elements making up the load can be precisely defined, the inverter specifications can focus simply on assuring the proper operation of all these elements and disregarding other electrical parameters. This is just the case in 5 many PV projects. Particularly when rural electrification is concerned, the elements making up the load used to be provided within the same project frame as the PV systems themselves. In such cases the load can be defined, for example, in terms of: a XX inch, XX W television (or even, the television model XX from XX, or similar); a standard XX W drill, etc. Moreover, additional information describing the expected pattern of load use is also required in order to determine, both, the simultaneously allowed load elements, and the energy requirements. The first is needed for inverter specification, while the second should be known for PV-system sizing purposes. Keeping in mind the research into this way of inverter specification, we have carried out a laboratory testing campaign combining commercially available inverters and real AC loads. The following aspects have been analysed separately: On the AC side: - Rated power and surge capabilities. - Voltage and frequency regulation. - Harmonic distortion. On the DC side: - Low voltage disconnection. - Ripple. General features: - Power efficiency. - Reliability. - Other aspects. 6 3. TESTING CAMPAIGN Sixteen commercial inverters from 13 different suppliers and 6 different countries have been tested (table 1). Rated power ranges from 0,14 to 10 kva. Fifteen are onephase 230V/50Hz and only one is three phase 230V/400V. According to their voltage elevation topology, two inverter types can be distinguished: high-frequency, which are switched at high frequency and use a small and light transformer made of ferrite, and low-frequency, which refers to inverters that incorporate bulky iron transformers. According to the output voltage waveform, inverters can be classified as sinusoidal, quasi-square and square, as already described in figure 1. Inverters I1 to I10 have been directly acquired by the IES-UPM, while inverters I11 to I16 have been tested at the request of private companies and organizations involved in PV rural electrification projects. Certain testing results may not be available for all inverters because some of them broke down during the tests. Table 1. Main characteristics of the tested inverters. Table 2 lists the AC loads that we have used in the testing campaign. Roughly, they represent the vast majority of AC appliances currently powered by PV systems. Most tests (voltage and frequency regulation, power efficiency, overload etc.) require the load power to be varied within a suitable range, adapted to the particular rated power of each inverter. For that, we have relied on resistive loads. The simple combination of several commercial incandescent lamps allows the range below 1 kw to be covered by steps of 25 W (figure 3-a). For high power levels, up to 15 kw, a resistive load bank has been implemented using water heaters (figure 3-b). These resistive loads are widely available, cheap and easy to use, which are important advantages when looking into 7 local testing application. Other loads have been used for testing particular inverter features: surge capabilities, harmonics, electromagnetic interference, etc. Table 2. Electrical characteristics of the AC loads used for inverter testing. Figure 3. Resistive loads used for inverter testing. (a) Incandescent lamps (up to 1 kw) (b) Water heaters (up to 15 kw). Inverters must be tested at the maximum ambient temperatures of the final site, which is usually greater than the ambient temperatures of our laboratory. This has led us to develop a climatic chamber. Again, looking for local applicability, this chamber has been built using conventional materials: a garden shed, thermally-isolated with expanded polystyrene and aluminium sheets to protect it against fire (figure 4-a). The control of temperature is carried out using a standard thermostat that measures the chamber temperature and controls an electric heater placed inside (figure 4-b). Figure 4. (a) Climatic chamber. (b) Ambient temperature control. Finally, it must be mentioned that the precise characterization of general AC performances require specific instrumentation (watt meters, power quality analyzers, digital oscilloscopes, etc.) capable of measuring parameters such as PF or THD X. However, when the elements making up the load can be precisely anticipated, an alternative means of characterization requiring only common instrumentation (true RMS voltmeters and calibrated shunts) can be applied. Because of its simplicity, special attention has been paid to this last alternative. 8 4. RESULTS AND DISCUSSION 4.1 Rated and surge power Obviously, the PV system design must ensure that rated inverter power is equal to or greater than the sum of the rated power of all the simultaneously permitted individual AC loads. Furthermore, the inverter must ensure the safe starting of whatever individual load in any normal operating condition, i.e., with the rest of the simultaneously permitted individual loads maintained in steady state operation (it is reasonable to assume two or more individual loads will not start at the same time). On the lack of a precise load definition, we have tested the inverters at their rated power at 25 ºC for one hour using pure resistive loads. All the inverters passed this test. We have also tested the inverter starting capabilities with all the compatible loads, i.e., loads whose rated power is equal to or lower than the rated power of the inverter. As a result, most inverters are capable of starting any individual load when no other loads are operated simultaneously. However, in some cases, even starting a single load can cause inverter shutdown. For example, the inverter I1 requires several ON/OFF switchings to start the TV. In general, as the total of rated power of simultaneously connected loads increases, the greater the difficulty for the inverter to start a particular load, especially motors. However, it is difficult to generalize and anticipate the inverter behaviour for a particular combination of elements making up the load, which advises matching the inverter to the specific load as far as is possible. Under starting conditions, the surge power can exceed the rated one and some inverters behave as sources of current limiting the output current and, therefore, reducing the output voltage, which can drop more than 40% below the nominal value (figure 5). Such transient voltages can negatively affect the proper operation of other 9 AC loads connected in parallel, which may require a certain load management in order to avoid the simultaneous operation of incompatible loads. Figure 5. Voltage and current evolution at inverter I12 output during the starting of a motor. For about 3 seconds, the voltage drop reaches 40%. 4.2 Voltage and frequency regulation AC loads should generally be operated at a fixed voltage and frequency. Several problems can rise from both over voltage and under voltage: load damage, not-ignition, poor luminosity, etc. Besides, frequency variations can affect equipment using frequency as a reference for their operation, such as internal clocks or timers. Regulation refers to the inverter s ability to maintain the steady state AC output voltage, V OUT, and the frequency, f OUT, close to the nominal value face fluctuations of DC input voltage and load power demand (the previously described starting transient phenomena are excluded from this concept). We have tested this inverter capability, again with resistive loads, by measuring V OUT and f OUT for DC input voltages ranging from 11 to 15 V (12 V reference), and for load power ranging from zero to the inverter rated power. The resulting V OUT range has thus been described in terms of the average RMS value, V o, the maximum, V o + ΔV, and the minimum, V o - ΔV. The frequency range is also described in this way. Table 3 summarizes the results. This test must cover the expected ranges of operating output power and DC input voltages. In general, it is sufficient to carry out the test at three output power levels (0%, 50% and 100% of inverter rated power) and three input voltages (90%, 100% and 120% of nominal input voltage). 10 Table 3. Voltage and frequency regulation of the tested inverters. DC input voltage is varied from 11 to 15 V; meanwhile load power is varied from zero to the inverter rated power. It can be seen that frequency regulation is usually good. Conventional grid standards 15, 16 usually establish a limit of 2% (± 1 Hz), or even 1%, for frequency fluctuations. These rules can be extended to PV inverters. A good possibility is to choose 2% as a compulsory value, and 1% as a recommended one. All the tested inverters, except I1, comply with these comfortably. In general, voltage regulation is not as good as frequency regulation, and several inverters behave poorly. It is worth commenting that most inverters have a better regulation when they operate at a fixed input voltage and a varying output power. However, their regulation is worse when DC input voltage changes. As representative examples, figure 6-a shows the variations in V OUT versus the DC input voltage for three tested inverters operating at a fixed output power. Quasi-square inverters deserve a further comment. In order to regulate the RMS value of V OUT, the pulse width is modulated in each semi period of the voltage waveform, by decreasing the width pulse when the DC voltage increases and vice versa. The important point is that, despite this procedure stabilizing the RMS value, the peak voltage remains unregulated and can reach high values. Figure 6-b shows an example of this. Figure 6. (a) RMS output voltage versus DC input voltage for three different tested inverters operating at a fixed resistive load. (b) AC output voltage (RMS and peak values) of inverter I9 versus DC input voltage operating at a fixed resistive load. 11 Voltage regulation standards for the conventional grid vary from country to country. For example, in Europe, the voltage regulation must be within ± 10% 15, while in the USA, national standards require a regulation within ± 5% of the nominal voltage 17. These rules can also be extended to the RMS value of PV inverters. A good possibility is to choose 10% as a compulsory value, and 5% as a recommended one. As far as the grid is concerned, as the voltage is sinusoidal, the peak voltage is proportional to the RMS one ( 2xVrms) and their relative fluctuations are the same. However, as stated above, the same is not true for quasi-square inverters, whose peak voltages can reach high values. For these inverters, it seems advisable to limit the peak voltage to the maximum allowed for the sinusoidal ones, for example, to 1,55 ( 2x1,1) times the nominal RMS voltage as compulsory limit. 4.3 Harmonic distortion Grid related standards involve a divided responsibility between consumers and utilities. The latter being responsible for the quality of supply, they must ensure that the total harmonic distortion of voltage waveform, THD V, is always kept below certain limits. And, for that, they resort to limiting the harmonic content of the current consumed by users, applying harmonic limits to individual appliances 18, 19. For example, in Europe, the THD V of public and industrial supply networks must always be (regardless of the load condition) less than 8% 15. The IEEE recommends 19 a lower limit: THD V 5%. However, the alternative of extending these strict THD V limits to the case of common
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks