1 00:00:05,460 --> 00:00:10,950 In this block, I will discuss the concept of maximum power point tracking, or as is 2 00:00:10,950 --> 00:00:14,300 commonly called - MPPT. 3 00:00:14,300 --> 00:00:20,340 This concept is very unique to the field of PV systems, and this also brings about a very 4 00:00:20,340 --> 00:00:25,100 special application of power electronics in the field of photovoltaics. 5 00:00:25,460 --> 00:00:33,000 The concepts discussed in this block are equally valid for cells, modules, and arrays, although 6 00:00:33,190 --> 00:00:41,200 Maximum power point tracking (MPPT) concepts are usually employed at PV module/array level in practice. 7 00:00:41,300 --> 00:00:47,079 For reasons discussed before, the solar cell has an I-V curve as shown in the figure (graph). 8 00:00:47,079 --> 00:00:53,749 As a quick recap, you might remember how the I-V curve of a PV module is similar to that 9 00:00:53,749 --> 00:01:01,300 of a single solar cell, as the PV module is merely an interconnection of several solar cells. 10 00:01:01,539 --> 00:01:08,539 A series or parallel interconnection of cells increases the voltage or current respectively, 11 00:01:08,540 --> 00:01:12,540 but the overall nature of the I-V curve remains the same. 12 00:01:12,540 --> 00:01:19,140 In a similar manner, it can be said that the interconnection of the PV modules to form a PV 13 00:01:19,140 --> 00:01:25,689 array would yield comparable I-V curves, albeit at different I-V levels. 14 00:01:25,689 --> 00:01:33,400 Thus, it is reasonable to consider the I-V curve of a solar module or array similar in 15 00:01:33,400 --> 00:01:36,200 nature to that of a solar cell. 16 00:01:36,200 --> 00:01:40,810 So let's head to the simple I-V curve of a solar module. 17 00:01:40,810 --> 00:01:45,430 Now, the I-V curve is nothing but the current-voltage curve. 18 00:01:45,580 --> 00:01:52,950 This I-V curve of a PV module is defined for a unique set of temperature and irradiance conditions. 19 00:01:53,260 --> 00:02:00,359 For instance, if the irradiance (or illumination) were to increase, the I-V curve would also 20 00:02:00,359 --> 00:02:02,789 change to a higher level. 21 00:02:02,789 --> 00:02:09,380 In general, a higher irradiance gives a better I-V curve, but a higher temperature gives 22 00:02:09,380 --> 00:02:13,700 a worse I-V curve and vice versa. 23 00:02:13,960 --> 00:02:21,040 Now let's try to understand the concept of the operating point. 24 00:02:21,050 --> 00:02:28,050 At any given point in time, the solar module operates at a particular voltage and current. 25 00:02:28,220 --> 00:02:34,340 The point on the I-V curve where the solar module operates is called the operating point. 26 00:02:34,520 --> 00:02:40,400 For a given irradiance and temperature, an operating point on the I-V curve corresponds 27 00:02:40,400 --> 00:02:44,800 to a unique (I,V) value. 28 00:02:45,400 --> 00:02:50,240 We know that power = V*I. 29 00:02:50,400 --> 00:02:57,380 If we draw the power-voltage curve, or the P-V curve, it takes a form as shown. 30 00:02:57,380 --> 00:03:03,640 Of course, an operating point on the I-V curve will also correspond to a unique operating 31 00:03:03,640 --> 00:03:06,310 point on the P-V curve. 32 00:03:06,310 --> 00:03:13,300 Now what is the significance of the P on the P-V curve of a solar module? 33 00:03:13,300 --> 00:03:21,800 Well, this is the power that is produced and delivered to the rest of the PV system, and eventually the load. 34 00:03:21,850 --> 00:03:27,980 Therefore, it is clearly advantageous that the solar module operates at maximum power 35 00:03:27,980 --> 00:03:33,020 as seen in the figure as the peak of the P-V curve. 36 00:03:33,020 --> 00:03:40,020 Now, without any external electrical manipulation, the PV module's operating point is largely 37 00:03:40,220 --> 00:03:47,220 dictated by the electrical load seen by the PV module at its output. 38 00:03:47,470 --> 00:03:54,350 To get maximum power delivered by the PV module, it is therefore imperative to force the module 39 00:03:54,350 --> 00:04:01,300 to operate at the operating point corresponding to the maximum power, or as it's generally called, 40 00:04:01,340 --> 00:04:04,260 the maximum power point (MPP). 41 00:04:04,420 --> 00:04:11,660 This point corresponds to the peak of the P-V curve or the "knee" of the I-V curve. 42 00:04:11,660 --> 00:04:17,640 The simplest way to do this, is to force the voltage of the PV module to be that at the 43 00:04:17,680 --> 00:04:25,220 MPP (called Vmpp) or regulate the current to the right amount as that at MPP (called Impp), 44 00:04:25,319 --> 00:04:28,599 using converters. 45 00:04:28,900 --> 00:04:36,500 But what if, after forcing the PV module to operate at MPP, the ambient conditions, 46 00:04:36,500 --> 00:04:44,060 like irradiance or temperature change and in turn cause the I-V/P-V curve to change as well? 47 00:04:44,700 --> 00:04:51,500 This would mean that the old MPP is no longer valid under these conditions. 48 00:04:52,000 --> 00:05:00,100 Thus, to be continuously at the MPP at all times, we would need to track any such changes 49 00:05:00,120 --> 00:05:04,560 in the I-V curve, and find out the new MPP. 50 00:05:04,580 --> 00:05:11,000 This process is called maximum power point tracking or MPPT, and the devices that perform 51 00:05:11,009 --> 00:05:14,329 this process are called MPP trackers. 52 00:05:14,580 --> 00:05:18,479 So how do these MPPT devices work? 53 00:05:18,479 --> 00:05:27,800 An MPPT device is nothing but a hardware implementation of an MPPT algorithm or algorithms. 54 00:05:27,800 --> 00:05:32,280 There are several algorithms to track the MPP effectively. 55 00:05:32,340 --> 00:05:36,020 These are also called MPPT techniques. 56 00:05:36,030 --> 00:05:43,030 Some of the modern literature talks about two broad categories of MPPT techniques: 57 00:05:43,380 --> 00:05:48,240 • Indirect MPP tracking, like fractional open-circuit voltage method, or 58 00:05:48,340 --> 00:05:57,700 • Direct MPP tracking, like the Perturb and Observe method or the Incremental Conductance Method 59 00:05:57,840 --> 00:06:02,400 Each of these techniques have their own advantages and disadvantages. 60 00:06:02,520 --> 00:06:07,720 Let's talk about the indirect MPP tracking first. 61 00:06:07,729 --> 00:06:14,729 In this kind of tracking, simple assumptions and periodic estimations of the MPP are made 62 00:06:14,819 --> 00:06:16,490 with easy measurements. 63 00:06:16,490 --> 00:06:20,969 An example from practice is the fixed voltage method. 64 00:06:20,969 --> 00:06:29,700 This technique is based on adjusting the operating voltage of the solar module only on a seasonal basis. 65 00:06:29,900 --> 00:06:36,800 This is under the assumption that the higher MPP voltages are expected during winter, and lower 66 00:06:36,800 --> 00:06:42,400 MPP voltages are expected in the summer, for the same level of irradiance. 67 00:06:42,400 --> 00:06:48,229 Of course, this method is not very accurate and accuracy is increased if such an idea 68 00:06:48,229 --> 00:06:55,229 is implemented in a place with minimal irradiance fluctuations on a particular day. 69 00:06:55,680 --> 00:07:02,259 One of the most common MPPT techniques in indirect tracking category is the fractional 70 00:07:02,259 --> 00:07:04,599 open-circuit voltage method. 71 00:07:04,599 --> 00:07:12,100 This method exploits the fact that a good approximation of Vmpp is given by constant 72 00:07:12,100 --> 00:07:16,330 k multiplied by Voc. 73 00:07:16,330 --> 00:07:24,300 This constant k, for crystalline silicon is usually found to be around 0.7-0.8 in practice. 74 00:07:24,800 --> 00:07:30,449 In general, constant k is based on the PV technology and the kind of solar cells in use. 75 00:07:31,800 --> 00:07:37,639 How is this technique rationalized? Let's discuss that for a moment. 76 00:07:37,639 --> 00:07:42,770 We know that every illumination level and therefore irradiance level corresponds to 77 00:07:42,770 --> 00:07:48,759 a particular I-V curve and therefore a specific level of Voc. 78 00:07:48,759 --> 00:07:59,700 It is found from practice that for a range of I-V curves, the ratio of Vmpp to Voc shows little variation. 79 00:07:59,900 --> 00:08:07,800 Thus, for changes in irradiance, if the Voc can be measured quickly, the Vmpp can be easily 80 00:08:07,800 --> 00:08:12,140 estimated as a fraction of the Voc. 81 00:08:12,280 --> 00:08:18,820 This technique is easier to implement compared to some of the more complicated techniques. 82 00:08:18,830 --> 00:08:22,659 What then, are the drawbacks of such technique? 83 00:08:22,659 --> 00:08:29,300 Well, first, as the concept of k is just an approximation, this method does not always 84 00:08:29,300 --> 00:08:38,500 point us to the true MPP, but only in the region around MPP or as it's commonly called, 85 00:08:38,570 --> 00:08:41,590 the MPP region. 86 00:08:41,590 --> 00:08:47,520 Secondly, every time the system needs to respond to a change in illumination conditions, the 87 00:08:47,520 --> 00:08:53,540 MPPT algorithm needs to measure the Voc. 88 00:08:53,540 --> 00:09:00,540 How does the MPPT device measure the Voc of a solar module under operation? 89 00:09:01,800 --> 00:09:08,280 This is done by disconnecting the PV module from the load momentarily, or in other words, 90 00:09:08,290 --> 00:09:12,770 the PV current is zero so as to measure the Voc. 91 00:09:12,920 --> 00:09:19,640 Although temporarily, this results in a loss in production of the PV power. 92 00:09:19,650 --> 00:09:26,650 This handicap only worsens if this measurement of Voc has to be made more frequently. 93 00:09:28,510 --> 00:09:34,470 A slight modification to this method can help in overcoming this apparent drawback. 94 00:09:34,470 --> 00:09:39,750 This involves the inclusion of a pilot PV cell that is highly matched with the rest 95 00:09:39,750 --> 00:09:44,000 of the PV cells that constitute the PV module. 96 00:09:44,000 --> 00:09:50,960 The idea here is that the lone pilot PV cell receives the same illumination as the rest 97 00:09:50,960 --> 00:09:56,780 of the PV module, and a measurement of the pilot PV cell also gives an accurate representation 98 00:09:56,780 --> 00:10:00,030 of the PV module under consideration. 99 00:10:00,030 --> 00:10:09,600 Thus, while the PV cell provides for a good estimate of the module's Voc, the Vmpp can 100 00:10:09,600 --> 00:10:20,800 be calculated as Vmpp = k*Voc and the PV voltage can be adjusted accordingly, without having 101 00:10:20,800 --> 00:10:23,370 to disconnect the PV module. 102 00:10:23,370 --> 00:10:30,300 Note that the inclusion of PV cell merely avoids the losses due to module disconnection. 103 00:10:31,030 --> 00:10:38,500 This method would still suffer from the problem of depending on the k value for estimating the Vmpp. 104 00:10:38,700 --> 00:10:47,400 Now let's look at the more involved kind of MPPT - the direct MPP tracking. 105 00:10:47,490 --> 00:10:56,200 This kind of MPPT involves direct measurements of current, voltage or power and a more accurate/faster 106 00:10:56,200 --> 00:10:59,990 response than the indirect method. 107 00:10:59,990 --> 00:11:06,160 Due to time constraints, we shall look at a couple of the most popular kind of algorithms, 108 00:11:06,160 --> 00:11:11,790 which are sometimes also classified as "hill climbing" algorithms. 109 00:11:11,790 --> 00:11:18,330 We start first with the Perturb and Observe, or P&O algorithm. 110 00:11:18,330 --> 00:11:25,330 In this algorithm, a perturbation is provided to the PV module or array voltage. 111 00:11:26,060 --> 00:11:29,610 This would translate to an increase or decrease in power. 112 00:11:29,610 --> 00:11:36,480 If an increase in voltage leads to an increase in power, this means that the operating point 113 00:11:36,480 --> 00:11:43,230 is to the left of the MPP, and hence further voltage perturbation is required towards the 114 00:11:43,230 --> 00:11:46,670 right to reach the MPP. 115 00:11:46,670 --> 00:11:53,670 Conversely, if the increase in voltage leads to a decrease in power, this means that the 116 00:11:53,810 --> 00:11:59,920 current operating point is to the right of the MPP, and hence further voltage perturbation 117 00:11:59,920 --> 00:12:05,110 is required towards the left to reach the MPP. 118 00:12:05,110 --> 00:12:12,110 In this way the algorithm converges towards the MPP over several perturbations. 119 00:12:12,180 --> 00:12:17,910 You would have noticed that this algorithm takes advantage of the fact that the P-V curve 120 00:12:17,910 --> 00:12:25,300 has an increasing nature to the left of the MPP and a decreasing nature to the right of the MPP. 121 00:12:25,300 --> 00:12:32,330 The problem with this algorithm is that the operating point is never steady at the MPP. 122 00:12:32,560 --> 00:12:39,320 It is always hovering around, in the MPP region, although this could be minimized using very 123 00:12:39,320 --> 00:12:44,360 small perturbation steps around the MPP. 124 00:12:44,360 --> 00:12:48,790 This algorithm also struggles under rapidly changing illuminations. 125 00:12:48,790 --> 00:12:54,800 For example, if the illumination (and therefore irradiance) changes in between two sampling 126 00:12:54,800 --> 00:13:03,000 instants in the process of convergence, then the algorithm essentially fails in its convergence efforts. 127 00:13:03,200 --> 00:13:06,000 This situation is illustrated in the figure as well. 128 00:13:06,000 --> 00:13:12,650 In the latest perturbation, the algorithm has determined that the MPP lies to the right 129 00:13:12,650 --> 00:13:20,290 of point A, and hence the next step is a perturbation to converge towards the MPP accordingly. 130 00:13:20,840 --> 00:13:28,320 However, as the illumination changes rapidly before the next perturbation, the next perturbation 131 00:13:28,410 --> 00:13:35,410 shifts the operating point to B, while the new MPP actually moves to the left of B. 132 00:13:36,500 --> 00:13:47,500 This can be severely erroneous because the algorithm has now reached a point B such that PB > PA. 133 00:13:48,000 --> 00:13:54,380 But the MPP still lies to the left of point B, which makes the algorithm think that the 134 00:13:54,380 --> 00:13:57,750 MPP is to the right of point B. 135 00:13:57,750 --> 00:14:04,500 This is detrimental to the speed of convergence of the P&O algorithm, which is one of the 136 00:14:04,500 --> 00:14:09,150 critical figures of merit in the field of the MPPT techniques. 137 00:14:09,150 --> 00:14:16,150 Thus, drastic changes in weather conditions severely affect the algorithm's efficacy. 138 00:14:16,290 --> 00:14:26,800 Next, we look at another kind of direct MPPT algorithm, called the Incremental Conductance Method. 139 00:14:26,800 --> 00:14:32,960 To better understand the algorithm, we shall first arrive at a relation between conductance 140 00:14:32,960 --> 00:14:37,120 and incremental conductance. 141 00:14:37,120 --> 00:14:42,430 Conductance of an electrical component is nothing but a ratio of current to voltage, 142 00:14:42,430 --> 00:14:46,980 or in other words, reciprocal of resistance. 143 00:14:46,980 --> 00:14:56,300 We know, at the MPP, the slope of the P-V curve is zero, i.e. dP/dV = 0. 144 00:14:57,350 --> 00:15:04,350 Now dP/dV could be written as d(I*V)/dV. 145 00:15:05,350 --> 00:15:15,300 Using basic differentiation, we get dP/dV = I + V.dI/dV. 146 00:15:15,600 --> 00:15:23,300 If the sampling steps are small enough , then dI/dV could be approximated to ΔI/ΔV. 147 00:15:23,300 --> 00:15:32,500 Thus, at MPP, ΔI/ΔV = - I/V. 148 00:15:32,500 --> 00:15:45,400 To the left of the MPP on the P-V curve, ΔI/ΔV > - I/V and to the right of the MPP 149 00:15:45,400 --> 00:15:52,940 on the P-V curve, ΔI/ΔV < - I/V. 150 00:15:54,160 --> 00:16:01,160 The algorithm exploits these basic facts about the I-V and P-V curve of a solar module. 151 00:16:02,910 --> 00:16:09,910 In general, the algorithm imposes a voltage on the PV module at every iteration, measures 152 00:16:09,910 --> 00:16:17,120 the incremental change in conductance, compares it with the instantaneous conductance, and 153 00:16:17,130 --> 00:16:21,690 decides if the operating point is to the left or to the right of MPP. 154 00:16:21,690 --> 00:16:27,150 I will explain this in a brief conceptual flowchart. 155 00:16:27,150 --> 00:16:31,440 Note that this flowchart is not exhaustive. 156 00:16:31,440 --> 00:16:38,280 In this example, we see an MPPT algorithm based on Incremental Conductance Method. 157 00:16:38,280 --> 00:16:44,470 The instantaneous voltage and current are the observable parameters, while the instantaneous 158 00:16:44,470 --> 00:16:48,930 voltage is also the controllable parameter. 159 00:16:49,080 --> 00:16:56,800 V_ref is the voltage value forced on the PV module by the MPPT device. 160 00:16:56,800 --> 00:17:02,660 It is the latest approximation of the Vmpp. 161 00:17:02,660 --> 00:17:08,620 For any change in the operating point, the algorithm compares the instantaneous and incremental 162 00:17:08,620 --> 00:17:09,789 conductance values. 163 00:17:09,960 --> 00:17:16,539 If incremental conductance is more than the negative of the instantaneous conductance, 164 00:17:16,539 --> 00:17:25,500 this means the current operating point is to the left of the MPP; consequently, V_ref is incremented. 165 00:17:26,129 --> 00:17:31,909 Conversely, if the incremental conductance is lower than the negative of the instantaneous 166 00:17:31,909 --> 00:17:38,909 conduction, the current operating point is to the left of the MPP and is decremented. 167 00:17:39,110 --> 00:17:45,549 This process iterates until the incremental conductance is the same as the negative of the 168 00:17:45,549 --> 00:17:53,500 instantaneous conductance, in which case the chosen reference voltage is equal to the Vmpp. 169 00:17:53,800 --> 00:18:03,200 This MPPT algorithm can be more efficient at the MPP as it doesn't hover in the MPP 170 00:18:03,240 --> 00:18:08,350 region under steady state like the P&O algorithm. 171 00:18:08,350 --> 00:18:15,350 Also, low sampling intervals make it less susceptible to the changing illumination conditions. 172 00:18:15,460 --> 00:18:22,460 However, under very highly varying conditions and partial shading, the incremental conductance 173 00:18:22,700 --> 00:18:27,220 method might also be rendered less efficient. 174 00:18:27,220 --> 00:18:33,669 The main drawback of this algorithm is the complexity of its hardware implementation. 175 00:18:33,669 --> 00:18:40,049 It needs to not only measure the currents and voltages but also calculate instantaneous 176 00:18:40,049 --> 00:18:42,139 and incremental conductance values. 177 00:18:42,139 --> 00:18:49,139 I will not go into the details of the hardware implementation of the MPPT techniques, it is beyond 178 00:18:49,190 --> 00:18:51,100 the scope of this course. 179 00:18:51,100 --> 00:18:58,500 Usually, a DC-DC converter is used for implementing the current or voltage regulation at the PV output. 180 00:18:59,360 --> 00:19:06,200 A typical example from power electronics is the buck-boost DC-DC converter, like the one shown here. 181 00:19:06,600 --> 00:19:12,809 The algorithm needed to decide what voltage and current to force the PV output to, could 182 00:19:12,809 --> 00:19:18,669 be any of the ones discussed here or several others from literature. 183 00:19:18,669 --> 00:19:25,669 The choice of the algorithm would dictate the complexity of the implementation. 184 00:19:26,419 --> 00:19:33,419 In modern PV systems, the function of the MPPT is often implemented within other system 185 00:19:33,590 --> 00:19:39,480 components like solar inverters and charge controllers, like shown here. 186 00:19:39,480 --> 00:19:44,900 Please note that the list of techniques and typical implementation are definitely not exhaustive. 187 00:19:45,190 --> 00:19:52,670 In fact, several algorithms and implementation techniques exist in practice as well as in literature. 188 00:19:53,559 --> 00:19:59,580 There are scores of scientific papers, patents, and proprietary technologies existing in this 189 00:19:59,580 --> 00:20:01,940 rapidly growing space. 190 00:20:01,940 --> 00:20:08,940 In view of time we have covered only the most common MPPT techniques. 191 00:20:10,059 --> 00:20:18,300 In summary, maximum power is delivered from the PV cell, module or array, if the operating 192 00:20:18,300 --> 00:20:23,169 point is the same as that of the "knee" of the I-V curve. 193 00:20:23,169 --> 00:20:30,000 An ideal MPPT device is not only able to track that knee under varying conditions, 194 00:20:30,000 --> 00:20:34,320 but also maintain the voltage and current accordingly. 195 00:20:34,320 --> 00:20:41,210 In doing so, the device is expected to use as low an energy amount as possible, so as 196 00:20:41,210 --> 00:20:47,450 to not undo the additional energy gains the MPPT process provides. 197 00:20:47,450 --> 00:20:54,940 The MPPT device is also expected to respond swiftly and accurately to the changes in the 198 00:20:54,940 --> 00:21:00,139 ambient conditions that impact the PV output. 199 00:21:00,139 --> 00:21:05,960 In the next blocks this week, we shall look at the other components of the PV system that 200 00:21:05,960 --> 00:21:12,960 make the harnessed, intermittent solar energy usable for various applications. 201 00:21:12,960 --> 00:21:14,900 See you in the next block.