Efficiency Analysis of PV-Inverter System Connected to a Network of Adjustable Speed Drives

Makbul Anwari1, Muhammad Ikram Mohd Rashid2 and Taufik3

Faculty of Electrical Engineering, University of Technology Malaysia, Johor, Malaysia (Tel : +607-5535235; E-mail: Makbul@fke.utm.my) 2 Faculty of Electrical and Electronic Engineering, Universiti Malaysia Pahang, Pahang, Malaysia (Tel : +609-4242153; E-mail: Mikram@ump.edu.my) 3 Department of Electrical Engineering, California Polytechnic State University, US (Tel : + 805-756-2318; E-mail: taufik@calpoly.edu)
1

Abstract This paper presents efficiency analysis of a distributed generation system consisting of Photovoltaic (PV) Inverter system as the renewable energy source connected to a network of Adjustable Speed Drives (ASDs) as the load. A smallscale laboratory of the Photovoltaic, Inverter, ASDSs and Loads system were setup. Two power supplies from Photovoltaic (PV) and grid were connected to two ASDSs with two separate induction motors. Measurements were recorded from which the interaction between the source and the load in this system can be analyzed. Analysis of the result was carried out in which results of both experiments were compared. Furthermore, the worst and best case operating scenarios were indentified from which base-line information for optimum operating conditions is achieved. Index TermsPV-Inverter, Efficiency. Adjustable Speed Drives,

I. INTRODUCTION enewable energy is no longer just an option nowadays. The renewable energy plays an important role in our live due to rise demand for electrical power. Photovoltaic is one of the renewable energy that has been used for about decade ago. The word "photovoltaic," first used in about 1890, is a combination of the Greek word for light and the name of the physicist and electricity pioneer Allesandro Volta. So, "photovoltaic" can be translated literally as "light-electricity." Photovoltaic or known as solar energy is a renewable resource that is vast and is locally available. It is a clean energy source that allows for local energy independence. The sun's energy flow reaching the earth is typically about 1,000 Watts per square meter (W/m2), although ease of use varies with location and time of year. Simple PV systems provide power for many small consumer items, such as calculators and wristwatches. More complicated systems provide power for communications satellites, water pumps, and the lights, appliances, and machines in some people's homes and workplaces. Many road and traffic signs along highways are now powered by PV. In many cases, PV power is the least expensive form of electricity for performing these tasks [1]. Photovoltaic produce DC voltage to the load. The voltage can be converting to AC by using DC-AC converter which is commonly known as inverter. The inverter conversion process with power electronics is generally known as the switched-mode inversion. One major application of switchedmode is Adjustable Speed Drive (ASD). Adjustable Speed Drives (ASDs) are power electronic circuits used to control the speed of motors. In practice, many ASDs are connected together to form a network of adjustable speed drives to control the speed of motors in manufacturing lines, buildings (for

HVACs), agricultural sectors (for irrigation pumps), and household applications (such as those found in energy-saving washers and dryers). Today, 55% of the total electrical power in the US is consumed by the motor drives industry. Due to the advent in power electronics, adjustable speed drives employing solidstate switches have become popular in motor applications due to the significant energy saving that they offer. It has been estimated that the widespread use of adjustable speed drives causes as much as 20% reduction in energy consumption. This is equivalent to the total electric energy produced by about 162 power plants. So, in this project, PV system and ASD will be used and connected to each other to see how it can save the energy usage and although to see their efficiency. Further, the results of this project, henceforth, will be useful in developing a smart or innovative load management PV-inverter-ASD system that ensures maximum system efficiency at any given time. This, in turn, will result in significant energy savings, suppression of operating cost, and provision of efficient planning tool for future expansion of PV system as the alternate source of energy. Power quality is an aspect of power engineering that has been with us since the inception of power systems. However, topics in power quality have risen to the forefront since the advent of high power semiconductor switches and networking of transmission and sub-transmission systems. Also, the trends in modern power engineering have been extracting the most from the existing installed system, and this has placed stress on issues of sinusoidal waveform fidelity, absence of high and low voltage conditions, and other ac waveform distortion. It is estimated that industrial and digital economy companies collectively lose $45.7 billion a year to outages and another $6.7 billion each year to power quality phenomena [1]. Harmonic distortion is the most common power quality problem and it is found in both the voltage and the current waveform [2]. Current harmonic cause increased loses to customer and utility power system, they produce poor power factor, distorted voltage waveform (causing voltage harmonics) and they could produce dangerous resonant oscillations in the utility power supply. Previous studies have shown that the drive load factor and the system impedance at the drive are the two factors impacting the input harmonic distortions in ASD. The current harmonic distortion and lower order harmonic currents in percent of the fundamental vary significantly with the drive load factors. However, the higher order harmonic currents above the 17th do not change much with the drive loading and system impedance

[3]. Harmonic currents in Amps will increases with the increase of drive load factors. One common application for ASD is for variable torque loads under different loading conditions. With multiple drives and multiple motors operating in the system, the harmonics created from various loads could crosstalk and could further positively or negatively affect the magnitude of distortion. However, little has been known about the harmonic interaction between these drives on a single point of common coupling (PCC). There has been a few of attempts done in the past to investigate harmonic distortions. As an example, previous research has been done for a single drive with two induction motors connected to the drive [4]. Besides that, there were also experiments for network of two drive system and two motors connected to each drive [5], network that contain different types of ASD with one motor connected to each drive [6], and system with mixture of ASD and loads connected together at PCC [7]. In this paper, an efficiency analysis of a network of three ASDs with one motor connected to each ASD will be presented. System impedance at the drive input is considered to be constant and the input harmonic distortions are determined by drive load factors only. Experimental results from this research will be important to further understand the interaction and hence reveal how the drives interact with each other. II. BACKGROUND Adjustable speed drive (ASD) is a general term applied to any device whose motor's speed can be manually or automatically varied and adjusted according to the circumstances. ASD is set up using power electronics technology and is also classified as the extension of switchedmode inverter. It converts AC input with a fixed rms and frequency to an AC output with variable rms and frequency. Usually, the drive is said to be "adjustable" speed if speeds may be selected from several different pre-set range. While the term "variable speed" means the output speed of the drive can be changed without steps over a range. The increase of electrical applications inevitably brings more non-linear loads installed on the power system consequently responsible for drawing the non-sinusoidal currents from the power supply. A typical example of a non-linear load is an ACDC converter, where the energy is usually converted by a standard diode rectifier due to the simplicity, reliability, lower cost and no need of bi-directional power flow, which make this topology commonly met in many of the actual ASDs. Voltage distortions and current distortions can cause severe problem to the utility provider, other equipment in the same bus and to other user too. Due to the high non-linear loads, the installations may receive a faulty alarm from one of the electrical equipments. This phenomenon occurs even if all the electrical equipments are operating well and is operated within their ratings. One assumption made to explain this phenomenon is that the high level of voltage distortion (due to the ASDs harmonic currents) causes the fault trip. In order to understand the effects of the distortion, the Total Harmonic Distortion (THD) at end user and utility provider will be analyzed. Voltage and current will be measured at several points to observe how the harmonics interact with each

other. Multiple ac drive system will be also set up in the laboratory. In this paper, the THD of the system is obtained and the methods of characterizing each type with measurements are presented. Then, the respective loading characteristic to the ASD network is studied and analyzed. III. EXPERIMENTAL SETUP AND PROCEDURES

Figure 1. System block diagram: Grid connection

Figure 2. System block diagram: PV-Grid connection The experimental setup of this project consists of two power supplies. The first source of energy supply to the drive was from the grid transformer system meanwhile the second source of energy supply to the drive was from the photovoltaic inverter system. Those two power supplies were connected to three adjustable speed drives of 0.5 hp, 0.7 hp and 1.5 hp. Each of adjustable speed drives (ASD) powered three separate induction motors. 0.5 hp ASD powered 0.25 hp induction motor that was referred as motor 3. 0.7 hp ASD powered 0.7 hp induction motor that was referred as motor 2 whereas 1.5 hp ASD powered 2 hp induction motor that was referred as motor 1. Table 1 below shows list of induction motors used in the experiments. Table 1: List of Induction Motors used in Experiments
Induction Motor Motor 1 ( Carpanelli) Motor 2 (Marelli) Motor 3 (Feedback) Power (hp) 2 0.7 0.25

In this project, 2 types of experiment have been carried out. Figure 1 and Figure 2 show the block diagram of laboratory setup for data collection from the experiments. Figure 1 shows that only energy supply from the grid has been used while

Figure 2 shows that energy has been used from both supplies which are grid and PV. They are simplified diagram of the system indicating the labelling convention of the data and graphs. All experiments were conducted under room temperature. For experiment 1, energy supply from the grid was connected to the loads and energy supply from photovoltaic system was disconnected as shown in Figure 1. Measurements were recorded at point which is also shown in Figure 1. Certain motor loads were varied from 0% to 100% of load whereas certain motor loads were fixed at 40% of load. The variations of motor loads are shown in Table 2 below: Table 2: Variations of Motor Loads in Experiments
Experiment Experiment 1 (Without PV) Fixed Motor Load Motor 1 Motor 2 Motor 3 Motor 1 and Motor 2 Motor 1 and Motor 3 Motor 2 and Motor 3 Motor 1 Motor 2 Motor 3 Motor 1 and Motor 2 Motor 1 and Motor 3 Motor 2 and Motor 3 Varied Motor Load Motor 2 and Motor 3 Motor 1 and Motor 3 Motor 1 and Motor 2 Motor 3 Motor 2 Motor 1 Motor 2 and Motor 3 Motor 1 and Motor 3 Motor 1 and Motor 2 Motor 3 Motor 2 Motor 1

Figure 3: Graph for motor 1 fixes while motor 2 & 3 variables

Figure 4: Graph for motor 2 fixes while motor 1 & 3 variables

Experiment 2 (With PV)

After every data collections were completed, the data were compiled into graphs to monitor the efficiency of the system. Meanwhile for experiment 2 energy supply from the grid and PV were connected to the loads as shown in Figure 2. Measurements were recorded at point which is also shown in Figure 2. In this experiment, steps in the previous experiment were repeated. The measurement was conducted in the afternoon started at 12.28 pm till 1.13 pm (20 th June 2009). Load for motor 1 can only reach up to 40% load because the power of torque load measurement was small compared to motor1. All experiments were conducted in 50 Hz. IV. RESULT AND DISCUSSION Discussion for Experiment 1 In this experiment, only supply from the grid has been used. Figure 3 to figure 8 above show the graphs of efficiency verses torque load in percentage. Figure 3 to figure 5 show the graphs that only one motor was fixed at 40% of torque load meanwhile figure 6 to figure 8 show the graphs that two motors were fixed at 40% of torque load. Figure 4 shows the graph efficiency verses torque load for motor 1 and motor 3 when motor 2 was fixed at 40% of torque load.

Figure 5: Graph for motor 3 fixes while motor 1 & 2 variable

Figure 6: Graph for motor 1&2 fixes while motor 3 variable

Figure 7: Graph for motor 1&3 fixed while motor 2 varied

Figure 10: Graph for motor 2 fixes while motor 1 & 3variables Figure 8: Graph for motor 2&3 fixes while motor 1 variable For this graph, the highest efficiency was recorded at 68% when motor 1 and motor 3 were at 40% of torque load while motor 2 was fixed at 40% of torque load. Figure 5 shows the graph efficiency verses torque load for motor 1 and motor 2 when motor 3 was fixed at 40% of torque load. For this graph, the highest efficiency was recorded at 70% when motor 1 and motor 2 were at 40% and 100% of torque load while motor 3 was fixed at 40% of load torque. Figure 6 shows the graph efficiency verses torque load for motor 3 when motor 1 and motor 2 were fixed at 40% of load torque. For this graph, the highest efficiency was recorded at 73% when motor 3 was at 20% of torque load while motor 1 and motor 2 were fixed at 40% of torque load. Figure 7 shows the graph efficiency verses torque load for motor 2 when motor 1 and motor 3 were fixed at 40% of torque load. For this graph, the highest efficiency was recorded at 72% when motor 2 was at 100% of torque load while motor 1 and motor 3 were fix at 40% of torque load. Figure 8 shows the graph efficiency verses torque load for motor 1 when motor 2 and motor 3 were fixed at 40% of torque load. For this graph, the highest efficiency was recorded at 67% when motor 1 was at 40% of torque load while motor 1 and motor 3 were fixed at 40% of torque load. From these graphs, the highest efficiency was identified at 20% to 30% of torque load. The highest efficiency was recorded at 20% of torque load in figure 3 which is 74%. This highest efficiency was recorded when motor 1 was in fixed torque load while motor 2 and motor 3 were in variable torque load.

Figure 11: Graph for motor 3 fix while motor 1 & 2 variables

Figure 12: Graph for motor 1&2 fix while motor 3 variables

Experiment 2

Figure 13: Graph for motor 1&3 fix while motor 2 variable

Figure 9: Graph for motor 1 fixes while motor 2 & 3 variables

Motor 2 (40%) Motor 3 (40%) Motor 1 and Motor 2 (40%) Motor 1 and Motor 3 (40%) Motor 2 and Motor 3 (40%) Motor 1 (40%)

(20%) Motor 1 (40%) and Motor 3 (40%) Motor 1 (40%) and Motor 2 (100%) Motor 3 (20%) Motor 2 (100%) Motor 1 (40%) Motor 2 (20%) and Motor 3 (20%) Motor 1 (20%) and Motor 3 (20%) Motor 1 (40%) and Motor 2 (40%) Motor 3 (20%) Motor 2 (40%) Motor 1 (20%)

68% 70% 73% 72% 67% 70% 75% 71% 71% 72% 67%

Figure 14: Graph for motor 2&3 fix while motor 1 variable Discussion for Experiment 2 In this experiment, both supply from the grid and PV have been used. Figure 9 to Figure 14 above show the graphs of efficiency versus torque load in percentage which is exactly the same as previous experiment. Figure 9 to Figure 11 show the graphs that only one motor was fixed at 40% of torque load meanwhile Figure 12 to Figure 14 show the graphs that two motors were fixed at 40% of torque load. Figure 9 shows the graph efficiency versus torque load for motor 2 and motor 3 when motor 1 was fixed at 40% of torque load. For this graph, the highest efficiency was obtained at 70% when motor 2 and motor 3 were at 20% of torque load while motor 1 was fixed at 40% of torque load. Conversely, Figure 10 shows the graph efficiency versus torque load for motor 1 and motor 3 when motor 2 was fixed at 40% of torque load. For this graph, the highest efficiency was obtained at 75% when motor 1 and motor 3 were at 40% of torque load while motor 2 was fixed at 40% of torque load. While Figure 11 shows the graph efficiency versus torque load for motor 1 and motor 2 when motor 3 was fixed at 40% of torque load. For this graph, the highest efficiency was obtained at 71% when motor 1 and motor 2 were at 40% of torque load while motor 3 was fixed at 40% of torque load. On the other hand, Figure 12 shows the graph efficiency versus torque load for motor 3 when motor 1 and motor 2 were fixed at 40% of torque load. For this graph, the highest efficiency was obtained at 71% when motor 3 was at 20% of torque load while motor 1 and motor 2 were fixed at 40% of torque load. Figure 13 shows the graph efficiency versus torque load for motor 2 when motor 1 and motor 3 were fixed at 40% of torque load. For this graph, the highest efficiency was obtained at 72% when motor 2 was at 100% of torque load while motor 1 and motor 3 were fixed at 40% of torque load. Figure 14 also shows the graph efficiency versus torque load for motor 1 when motor 2 and motor 3 were fixed at 40% of torque load. For this graph, the highest efficiency was obtained at 67% when motor 1 was at 40% of torque load while motor 1 and motor 3 were fixed at 40% of torque load. From these graphs, the highest efficiency was identified at 20% to 40% of torque load. The highest efficiency was obtained at 20% of torque load in Figure 10 which is 75%. This highest efficiency was recorded when motor 2 was in fixed torque load while motor 1 and motor 3 were in variable torque load. Table 3: Summary of experimental results
Experiment Experiment 1 (Without PV) Fixed Motor Load Motor 1 (40%) Varied Motor Load Motor 2 (20%) and Motor 3 Highest Efficiency (%) 74% Motor 2 (40%) Experiment 2 (With PV) Motor 3 (40%) Motor 1 and Motor 2 (40%) Motor 1 and Motor 3 (40%) Motor 2 and Motor 3 (40%)

V. CONCLUSION In this paper, an efficiency analysis of PV-inverter system connected to a network of ASDs has been presented. As a results, the optimum operating condition based on the worst and best case operating scenarios has been obtained. The experiment results show that the efficiency increases from 0% up to 20% torque load and then decrease when the torque load increases up to 100%. From the experiments that were conducted show that graphs of efficiency in experiment 1 and experiment 2 have approximately same pattern but the highest efficiency was obtained in experiment 2 which is 75% when motor 1, motor 2 and motor 3 at 40% torque load.

VI. ACKNOWLEDGMENT The authors thank the Malaysian Government, Ministry of Science, Technology and Innovation (MOSTI) and Universiti Teknologi Malaysia for the ScienceFund Grant, Project No. 0101-06-SF0205. REFERENCES
[1] K. Lee, G. Venkataramanan, and T.M. Jahns, Source Current Harmonic Anlysis of Adjustable Speed Drives Under Input Voltage Unbalance and Sag Conditions, IEEE Transactions On Power Delivery, Vol.21, No.2, April 2006. J.M. Hernandez Cid and J.F. Velazquez Moran, Adjustable Speed Drives Test Bench for Harmonic Distortion Evaluation, Department of Electronics, ITESO, Mexico. X. Liang, Y. Luy, and D.O. Koval, Investigation of Input Harmonic Distortions of Variable Frequency Drives, IEEE 1422-1291-9, 2009. C.Mitsui, Taufik, and M. Anwari, Efficiency Study of Adjustable Speed Drive with Dual Motor Connection, 8th International Power Engineering Conference- IPEC, 2007, Singapore. Taufik, E. Matsumoto, and M. Anwari, Impact of Multiple Adjustable Speed Drive System to Power System Harmonics, 2nd IEEE International Conference on Power and Energy (PECon 08), December 1-3, 2008, JB, Malaysia. M. Anwari, M. Imran Hamid, and Taufik, Modeling and Simulation of A Network of Adjustable Speed Drives, 3rd

[2] [3] [4] [5]

[6]

International Conference on Modeling, Simulation and Applied Optimization, 2009. [7] L. Asiminoaei, S. Hansen, and F. Blaabjerg, Predicting Harmonics by Simulations. A Case Study for High Power Adjustable Speed Drives, Electrical Power Quality and Utilisation, Magazine Vol.2, No.1, 2006. [8] W. Xu, Status and Future Directions of Power System Harmonic Analysis, Department of Electrical and Computer Engineering, University of Alberta, Canada, 2003. [9] Makbul Anwari, M. I. M. Rashid, M. Imran Hamid, Taufik, Experimental Harmonic Analysis of a Network of Adjustable Speed Drives, TENCON November 2009. [10] F. Jurado, N. Acero, A. Lpez, J. I. Rojas, Adjustable Speed Drives and Voltage Stability, IEEE, 2001.