Maximum Power Point Tracking

Transcription

1 Lahore University of Management Sciences Maximum Power Point Tracking [An optimum way to track maximum power point of each panel in a multi solar panel system] Annum Malik Asad Najeeb Joveria Baig Muhammad Sohaib Iqbal

2 Abstract In this project, we propose a design of a Maximum Peak Power Point tracking (MPPT) controller and its implementation using a buck boost convertor topology. Each solar panel in a multiple panel system will be individually monitored by multiplexing in time and ensured that each one is operating at its maximum point by a feedback closed loop algorithm implemented in a microcontroller based design. Duty cycle of the buck-boost convertor will be continuously monitored and adjusted to extract maximum power. As verified by our simulations, this method will be able to give an efficiency of 89.2% under mild lightning conditions (irradiance of 0.5 Suns). The proposed project will involve design and implementation of a new MPPT algorithm and two input Signatures buck convertor of Approval (not used in past designs) for this purpose. Block diagram and the detailed working of each component will be discussed in the paper below

3 Signatures of approval

4

5 Problem Statement Section at a glance Introduction Statistical analysis Solar Irradiance Problem definition Vision

6 Problem Statement This section will cover mostly the motivation behind under taking this particular project as our intended research of study and the implications in future by the implementation by this new design. Past researches in this area covered hardware aspects and designs of algorithms to operate a single solar panel at its maximum efficiency; however they failed to incorporate the importance of tracking of maximum power point for individual panels in a multi solar panel system. In this paper, we propose to design a tracking system which will operate each panel in a multiple panel system at its maximum efficiency and hence give a higher total efficiency. The proposed design will aim to cover the aspects not targeted by previous researches and will answer the question of feasibility of implementing solar panels over their cost. We will start by describing the nature of problem that we are targeting in our design and its importance 1.1 Introduction: Renewable sources of energy acquire growing importance due to massive consumption and exhaustion of fossil fuel. Beginning with the surge in coal use which accompanied the Industrial Revolution, energy consumption has steadily transitioned from wood and biomass to fossil fuels. The early development of solar technologies starting in the 1860s was driven by an expectation that coal would soon become scarce. However development of solar technologies stagnated in the early 20th century in the face of the increasing availability, economy, and utility of coal and petroleum. As time passed by, environmental concerns gradually increased and the use of bio fuels and coal became a serious

7 concern. Moreover, with the advent of the new technology, researchers typically started focusing on the advantages of extraction of solar energy efficiently over other sources. Figure 1 summarizes this effectively, showing clearly that the energy received in a 24 hour span in the form of sunlight is significantly comparable to the energy contained in the fossil fuels all around the world. As years passed by, the need for coal substitutes and more research into solar energy became imperative. With the ever growing demand for energy and the transition of focus from petroleum based fuels to solar fuels, it becomes necessary to consider the implications of using solar energy to the economy of a developing country like Pakistan. 1.2 Statistical Analysis: A little foray into statistics reveals that a total of 20.5 million customers stand connected to the PEPCO system, while another 3.0 million or so are on the KESC rolls. Out of these, 21 million fall under the category of general customers, including both domestic and commercial users. Around eight million of these can be termed as affluent customers with monthly average usage of more than 300 units. The installed load of these customers should actually be more than 5 KW. This translates into the requirement of a 1KW solar power systems for this category, which can cater to at least percent of their demands. In financial terms, a 1 KW solar PV-based system costs Rs.500,000. Keeping in view the prevalent power tariff, the payback period for it has been calculated to be anything between 3-10 years. Now the question arises so to why would existing

8 customers take the trouble and expense to go for a solar power system when they are already using their private generators or UPS. The answer lies in the fact that UPSs in Pakistan have a poor efficiency of 19 to 45 percent at the most, delivering highly impure power. The charging of the standby systems adds hefty amounts to the monthly utility bills. The generators have an efficiency of around 18 percent and remain expensive to maintain. The noise and smoke pollution further tips the ante against generators. PV based solar power systems are almost maintenance free, environment friendly and need no fuel at all. These statistics clearly reveal the need for a highly efficient solar panel system to increase the efficiency of energy supplied to household using this system and so it became sufficiently necessary for us to undertake this area as our research of study. 1.3 Solar irradiance: Irradiance is a characteristic that deals with the amount of sun energy reaching the ground. The irradiance reaching the earth in ideal conditions is 1000W/m2. However, this value is altered significantly depending on where you are located geographically, the angle of the sun, and the amount of haze or cloud cover preventing all of the suns energy from reaching the ground. Since solar panels run strictly off the energy emitted from the sun, their output is affected by the changing irradiance. More details on how the current voltage characteristics of the panel vary with changes in the irradiance will be presented in the sections that follow.

9 These changes in temperature and irradiance make the V-I characteristic near impossible to predict or control. The only thing we have control over is the operating point of the solar panel. Without control over this operating point, the output of the solar panels will be very unpredictable resulting in an even more inefficient solar power system. Location: Worcester, MA N W Average Clear sky Solar tracking January February March April May June July August September October November December Average Average Clear Sky Solar Tracking Table 2-1: Insolation Levels in Massachusetts in kwh/m2/day

10 1.4 Problem definition: Photovoltaic arrays are used in many applications such as water pumping, battery charging, hybrid vehicles, and grid connected PV systems. A power Voltage graph of a solar panel shows that there is an optimum operating point where the panel is able to deliver maximum possible power to the load at a particular voltage or current. This point varies significantly with increasing temperature for every individual panel depending mainly on the cell temperature, irradiance and insignificantly on some other factors. Therefore, on line tracking of the maximum power point of a PV array is an essential part of any successful PV system. A variety of maximum power point tracking (MPPT) methods are developed. The methods vary in implementation complexity, sensed parameters, required number of sensors, convergence speed, and cost. The goal behind our implementation of the solar panel is to come up with a design to operate each panel to operate at its maximum power and to extract maximum power from the PV panels even in times of partial shading or minimum irradiance. 1.5 Vision/Goals: To cope up with the so formed bottleneck between the excessive demand and limited supply of energy in Pakistan To identify substitutes to petroleum resources We believe macroscopic implications of a supply shock induced energy crisis are large because energy is the resource used to exploit all other resources Sustained energy crisis becomes a humanitarian crisis Principles of green energy and sustainable living movements have gained popularity

11 Literature Review Section at a glance MPPT algorithms Dc to dc convertors MPPT controllers Charge Controllers

12 2.0 Literature review 2.0 Literature Review: Before coming up with a design to implement an effective algorithm for maximum power point tracking, it was necessary to perform a literary survey on existing designs of the controllers already implemented in this area. In this section, we present existing designs and programs for controllers implemented for single panels. Upon studying theses various techniques, it was evident that most of them were either used to control the current or voltage indirectly rather than controlling the voltage directly. These methods showed some serious limitations and disadvantages that we plan to overcome in terms of simplicity, high convergence speed and independence on PV array characteristics. Most of these algorithms failed to incorporate the effect of sudden changes in irradiation level. Detailed explanations of current researches in this area have been explained in this section. 2.1 MPPT algorithms MPPTs play a very important role in a PV system since they are responsible for maximizing the power output from a PV system for a given set of conditions of irradiance and temperature and hence maximize array efficiency. Thus, an MPPT can help minimize the overall system cost. Many such algorithms have been proposed up to date, ranging from direct methods to some indirect ones, these will be discussed in detail below and feasibility of each of these to be used will be considered to draw a conclusion as to which of these to use in our design.

13 The MPPT methods can be classified as direct and indirect methods. The direct methods include those methods that use PV voltage and/or current measurements. These direct methods have the advantage of being independent from the priori knowledge of the PV generator characteristics. Thus, the operating point is independent of isolation, temperature or degradation levels. The direct methods include the techniques of differentiation, feedback voltage, perturbation and observation (P&O) [2], incremental conductance, as well as fuzzy logic and neural network. The indirect methods are based on the use of a database of parameters that include data of typical P-V curves of PV systems for different irradiances and temperatures. Details of each of these algorithms follow in this section. There are a number of algorithms implemented in past researches, however after considering the feasibility of each of these, we managed to shortlist them, according to their feasibility in practical design and reliability of data, to two different methods from which to choose our proposed design. These two are discussed in detail below: Incremental Conductance Method: It may be observed that when a power- voltage equation is differentiated, the two sides of the equation gives incremental conductance and conductance of the other side. This forms the basis for this model Equations above shows that using the incremental conductance model on an IV

14 characteristic, MPP can be tracked. From the flow chart below, (figure ). The present value and the previous value of the solar array voltage and current are used to calculate the values of di and dv. If dv=0 and di=0, then the atmospheric conditions have not changed and the MPPT is still operating at the MPP. If dv=0 and di>0, then the amount of sunlight has increased, raising the MPP voltage. This requires the MPPT to increase the PV array operating voltage to track the MPP. Conversely, if di<0, the amount of sunlight has decreased, lowering the MPP voltage and requiring the MPPT to decrease the PVarray operating voltage. If the changes in voltage and current are not zero, the relationships in Equations can be used to determine the direction in which the voltage must be changed in order to reach the MPP. If di=dv > _I=V, then dp=dv > 0, and the PV array operating point is to the left of the MPP on the P V curve. Thus, the PV array voltage must be increased to reach the MPP. Similarly, if di=dv < _I=V, then dp=dv < 0 and the PV array operating point lies to the right of the MPP on the P V curve, meaning that the voltage must be reduced to reach the MPP. This describes the primary advantage of incremental conductance over the perturb-and-observe algorithm: incremental conductance can actually calculate the direction in which to perturb the array s operating point to reach the MPP, and can determine when it has actually reached the

15 MPP. Thus, under rapidly changing conditions, it should not track in the wrong direction, as P&O can, and it should not oscillate about the MPP once it reaches it. The simulation results for incremental conductance were simulated on Matlab R2009a and the results showed fair consistency with expected outcome as shown by figure The values taken were from a solar panel data sheet. For the exact code and its implementation procedure, refer to the appendix Perturb and Observe algorithms In P&O method, the MPPT algorithm is based on the calculation of the PV output power and the power change by sampling both the PV current and voltage. The tracker operates by periodically incrementing or decrementing the solar array voltage. If a given perturbation leads to an increase (decrease) in the output power of the PV, then the subsequent perturbation is generated in the same (opposite) direction. So, the duty cycle of the dc to dc convertor is changed and the process is repeated until the maximum power

16 point has been reached. Flowchart summarizes this behavior. Maximum power is initialized to a certain value and actual value is computed using the measured values of current and voltage. If the change in power comes out to be more than the limit set, the value of maximum power is perturbed and this keeps happening till the actual power becomes equal to the maximum power point on the graph. The simulation results, shown in figure show fair consistency and reliable tracking, however, it tends to oscillate at the maximum point and takes a fair amount of time to reach the maximum point. There are some serious drawbacks to this algorithm but hardware feasibility makes it easier to implement Conclusion After considering both the methods for the algorithm, we decided on implementing the perturb and observe (P&O) method for our design. We considered the shortcomings of each of the algorithms and came up with ways to overcome the shortcomings involved in the P&O method and thereby increasing its efficiency considerably. This method will be described in the next section.

17 2.2 dc-dc convertors: Since we intend to implement a multiple input DC-DC converter for our solar panels, we first have to go through the one input simple dc-dc converter theory and then based on the current research on multiple input dc-dc converters, we will have to design one that will work for multiple solar panels. Figure 1

18 The diagram for the buck dc-dc converter is shown in figure 1(a). The input from our solar panel is Vd and its is passed to the rest of the circuit through a switch which can be a MOSFET. The input to the MOSFET can be provided by a controller. The shape of the waveform at across the diode is given in figure 1(b); The duty cycle, D, can be changed by using PWM generated by a controller. During the on time (switch closed), the current passes through the inductor, charging the capacitor and going through the load, which is modeled by a resistor. And during the off time (switch open), the charged inductor and capacitor use their stored energy to provide current to the load (through the diode) while keeping Vo relatively constant. The 2 nd order low pass filter is basically there to get the average value of the waveform in figure 1(b) which is Vo. The spectrum for Vo can be seen in figure 1(b) and we make sure to get the dc component only by keeping the corner frequency, fc much farther away from zero frequency as shown in figure 1(c). Now, by solving for the output voltage, Vo in terms of other parameters we get: Vo = D * Vd And so given that Vd, voltage from the panel is varying, we need to continuously change the duty cycle, D to get a constant output voltage. The capacitor value is kept large so that the ripple in output voltage is kept to a minimum. The inductor makes sure that the current flowing through the load is almost constant too, by discharging during the time the switch is open (off time). It can be shown that, in steady state, the average inductor current is equal to the average load current, i.e. the average capacitor current, in steady state is zero. This also brings up a point about the two possible modes of operation: 1. Continuous-conduction mode 2. Discontinuous-conduction mode

19 In continuous-conduction mode, the current through the inductor never reaches zero, i.e. the switch closes before the current dies out. In discontinuous-conduction mode, the current throught the inductor does die out and reaches zero before the switch is closed. We are going to be focused on the continuous-conduction mode for our project and it can be shown that the relationship between Vo, Vd and D remains the same. The output current waveform is shown in figure 2 (top part). And the circuit diagram for on time and off time is shown in figure 2(a) and figure 2(b) respectively. (Figure 2 is on the next page). Figure 2

20 Project outline Section at a glance Overview Dc to dc convertors MPPT algorithm Charge Controllers and overall picture

21 3. Project Outline Project Outline: 3.1 Overview

22 4.2 Proposed MPPT algorithm The MPPT method proposed in our design will consist of a curve fitting technique to approximate the P-V curve as a fourth order polynomial and the coefficients that model this fourth order polynomial can be evaluated in terms of the cell temperature using the values given. The polyfit function in matlab was used to fit the power-voltage graph on the fourth order polynomial. Figure shows that the polynomial models the p-v graph to considerable precision. The flowchart in figure summarizes the method used in its implementation. In the method, an approximation fourth order polynomial function of the P-V curve is first obtained and then used to derive the coefficient functions of the P-V curve in terms of the cell temperature. The tracking process starts with sensing the ambient temperature and the sun s radiation. The reference temperature, irradiation and rated load voltage, are obtained from the PV module s data sheet. Then, the cell temperature is calculated by using the equation relating cell temperature to other parameters. Using the cell temperature value, the coefficients are calculated using equations. To determine an optimum PV voltage at which the PV power is equal to its maximum value from the fourth order polynomial function, the condition is considered. The PV module s voltage that is sensed by the controller is used as to determine the dppv/dvpv value. If it is equal to zero, it means that the PV generator is operated at its optimum voltage whereas if it is greater than zero, the process of searching the optimum voltage

23 is repeated by incrementing or decrementing the PV voltage with a constant, C set at a value of 0.1. After finding the optimum PV voltage, the optimum duty cycle is calculated. The simulation results in figure show a considerable efficiency in tracking the maximum power point of the p-v curve. The results obtained below by simulating the algorithm shows that its able to keep it between considerable limits.

24 The detailed algorithm for this section is given in the appendix attached.

25 3.3 Proposed two input dc-dc convertors Traditionally, two dc voltage sources are connected to two independent dc/dc power converters to obtain two stable and equivalent output voltages, which are then connected to the dc bus, to provide the electric energy demanded by the load (Figure 1).However in order to simplify the power system and reduce the cost, a double-input dc/dc converter (Figure 2) can be used. It is possible to connect two dc voltage sources either in parallel or in series to form an input voltage source for the dc/dc converter to transfer the desired power to the load. According to one design two dc sources can be put in parallel to implement the doubleinput pulse width-modulation (PWM) dc/dc converter while using coupled transformer. But this has one big drawback i.e. because of the voltage amplitude differences between two dc sources, only one of them can be connected to the input terminal of the dc/dc converter at a time. Hence the control scheme is based on the time-sharing concept. Consequently, power from difference dc sources cannot be transferred to the load simultaneously. A much better approach is to connect the two dc inputs in series to form a single voltage source and using the traditional dc-dc convertor to direct power to the load.

26 Look at Figure 3. Here VHI>VO>VLO. SHI and SLO are two power switches which are controlled by PWM. DLO and DHI are the two diodes that provide bypass path for the inductor to flow continuously when the power switches are switched off. By applying the PWM control scheme to the power switches SHI and SLO, the proposed double-input dc/dc converter can draw power from two voltage sources individually or simultaneously. The working of the circuit can be explained by dividing the working into 4 modes of operation (depending on the status of the power switches).

27 THE FOURMODES OF OPERATION MODEI (SHI: ON/SLO: OFF): The equivalent circuit of this mode is shown in Figure 4(a).When SHI is switched on with SLO switched off, the power diode DHI is reverse biased and hence can be treated as an open circuit. Meanwhile DLO is forward biased and as a result provides the bypass path for the inductor current IL to flow. In this mode VHI will charge the inductor and the capacitor as well as provide the current to the load.

28 MODE2 (SHI: OFF/SLO: ON): The equivalent circuit of this mode is shown in Figure 4(b).When SLO is switched on with SHI switched off, the power diode DLO is reverse biased and hence can be treated as an open circuit. Meanwhile DHI is forward biased and as a result provides the bypass path for the inductor current IL to flow. In this mode VLO will charge the inductor and the capacitor as well as provide the current to the load. MODE3 (SHI: OFF/SLO: OFF): The equivalent circuit of this mode is shown in Figure 4(C).Here both the power switches are switched off. Both DLO and DHI will provide the bypass current path for the inductor current. Both the voltage sources are disconnected from the circuit and thus the energy stored in the inductor and capacitor will be released to provide the required current to the load. MODE4 (SHI: ON/SLO: ON): The equivalent circuit of this mode is shown in Figure 4(d).Here both the power switches are switched on and both the power diodes are reverse biased. Two input voltages VOL and VHI are connected in series to charge the inductor. The capacitor provides the current to the load. In this mode of operation both the input voltages will transfer their energies simultaneously.

29 The switching frequencies of both the power switches are fixed to be the same. SHI and SLO are synchronized by the same turn-off transistor with the same turn-on moment. Figure 5 shows the waveforms of the key components of the circuit. Here power switches SHI and SLO have different conducting time i.e. the duty ratio for SHI is larger than the one for SLO. The voltage waveform across the inductor has three different levels which are determined by the ON OFF status of the main power switches SHI and SLO. The inductor-current waveform also has three different current slopes based on the three different inductor voltage values. The input currents reveal that two input voltage sources can provide electric energy for the proposed double-input dc/dc converter individually or simultaneously. The capacitor current will compensate the unfiltered output current which will result in a stable dc load current.

30 3.4 Charge controller and overall picture

31 Tentative timeline

32 References

33 Appendix