Simplified Maximum Power Point Controller For PV Installations
Simplified Maximum Power Point Controller For PV Installations
Simplified Maximum Power Point Controller For PV Installations
Johan HR Enslin (SMIEEE) Danie B Snyman (MIEEE) Dept of Electrical and Electronic Eng., University of Stellenbosch STELLENBOSCH, SOUTH AFRICA, FAX (+272231) 77-4981
ABSTRACT
A new low-cost, simplified control strategy, based on the positive feedback of the output current in an MPPT converter, is introduced. Cost effective Maximum Power Point Tracking (MPFT) reduces the energy cost of renewable energy generators by optimizing the utilization of the renewable energy source. Maximum power point tracking for relative small photovoltaic (PV) systems, with battery back-up, is achieved by employing this simplified positive feedback control strategy to maximize the output current into the battery. This new control strategy is analyzed and practical tested by using a Buck converter topology, but can also be used by the known converter topologies for PV systems and other renewable energy sources with a single power maximum. Experimental werification on the operation of this controller is included.
CHARACTERISTICS OF A PV PANEL
Effect of Irradiance (Insolation)
I-v 1 w / m2
1
80
70
3.5
,a 2 5
1 . :
1
0.5 0
B a
10 12 14
INTRODUCTION
A Photovoltaic (PV) Remote Area Power Supply
Figure 1 shows the I-V curves for a 60 W commercial PV panel at irradiance levels of 500, 750 and 1000 W/m2. As the irradiance level increases, the current also increases. The maximum power point of the panel increases with a steep positive slope, proportional to the irradiance. Higher irradiance levels result in associated higher power maxima and in higher voltages at the power maxima, if the cell temperature is constant. Parameters that influence the isolation levels on a surface at a fixed tilt on earth, is the daily and seasonal solar path; the presence of clouds, mist, smog and dust between the surface and the sun; and the shade of any object positioned such that the insolation level is reduced. Effect of Cell Temperature Variations
TEMPERATURE CHARACTERISTICS
mT*mrc)
Authorized licensed use limited to: Jawaharlal Nehru Technological University. Downloaded on October 21, 2009 at 01:38 from IEEE Xplore. Restrictions apply.
The cell temperature of a PV panel is a function of the insolation levels, the ambient temperature, the rate at which the cell is cooled and the way the PV panel is constructed. Cell temperatures as high as 85O C can easily be reached under 1 sun condition in the warmer areas of southern Africa. The general effect of cell temperature on the different parameters of a silicon solar cell, normalized to 25'C. is illustrated in figure 2. Increased cell temperature results in a lower open circuit voltage and max.power point, and higher short circuit current. The effect of cell temperature on the I-V characteristics of the same 60 W solar panel described in figure 1 is plotted in figure 3. Note the reduction of maximum available PV power and the lower voltage of this peak power point, as a function of higher cell temperatures. The combined effect of lower power maxima and lower voltage at these maxima for higher cell temperatures, can affect available power from PV arrays. By using an MPFT converter some power loss can be avoided.
The matching between the actual and rated PV array power is therefore a function of three parameters namely; insolation levels, cell temperature and array voltage. Figure 4 illustrates the matching of the F panel described in figures 1 V and 3. The influence of all three these mentioned parameters are clearly visible from figure 4. Methods of relieving the negative effects of these three parameters are now considered.
1101
I -~- V ~ * C
\ \
00
p; n)
70
I.V7S*C\
PdVOWPM
32.c5 4
l2
1.s
1
: I
30
Fig. 4: Load matching of a 60 W solar panel The limit imposed by insolation levels on the available PV power can be relieved by solar path trackers. Both daily solar path tracking and seasonal solar path tracking will improve the average insolation levels and therefore the limit of the: average daily matching is improved. Figure 4 shows that the matching limit of the PV panel (at 25'C cell temperature) may be improved from 47% at 0.5 sun to 100% at 1 sun, by using a r,olar path tracker. The limit imposed by cell temperature on the: PV system matching, may be relieved by improved thermal coinductivity between the PV cell and the environment. Although not easy to obtain, the matching limit of the PV panel can be improved from 78% at 75C cell temperature to 91% at 50C cell temperature.
As explained above, the insolation levels and the cell temperature determines only the limits of the best obtainable matching. The& matching is determined by the array voltage. Apart from a few PV systems that use some or another form of MPPT, PV arrays for RAPS systems are normally loaded by batteries. The array voltage is therefore determined by the battery state of charge, the voltage drop across the cable from tho PV array to the regulator and the voltage drop across the regulator. From figure 4 it can be seen that serious mismatching may occur during early mornings when the battery voltage is expected to be on its lowest (say 11 Volt), and the cell temperature is expected to be on its lowest (say 25C). For this case, thc: possible matching is loo%, but the PV system would operate at around 70% matching without the use of an MPFT converter. By using MPPT, the matching for this case could be improved to at least 95%. At solar noon the matching limit could for instance be 78% due to a cl temperature of 75C. el If the battery voltages plus the voltage drops are equal to 13.5
-1
7 - - - \ t I\ c - - +\
L
M '
I-10
1'0 12 14 1'6
Panel Voltage M
I,
L2 2 '
24O
Fig. 3: Panel I-V characteristics vls cell temperature According to the international standard for specifying the power rating of a PV panel, output power is specified at an insolation level of 1000 W/m2, a cell temperature of 25C and with an air mass of 1.5.Although the PV panel with the I-Vcharacteristicspresented in figure 1 and figure 3 is specified as a 60Watt panel, it show that this power output is only valid for 25C cell temperature and 1000 W/m2 insolation. From this information apparently a mismatch between the rated PV array power and the actual power generated by the PV array is found. Matchine of rated and actual PV array vowers The ratio between the actual PV array power and the rated PV array power is an indication of how successfully the PV generation facility is using [5]. As indicated in the previous paragraphs, a mismatch between the rated PV array power and the actual PV array power may result due to lower insolation levels and higher cell temperatures. Insolation levels and cell temperatures set the limits of the maximum power available from the PV array, while the loaded voltage, determines the actual amount of power delivered by the PV array.
1241
Authorized licensed use limited to: Jawaharlal Nehru Technological University. Downloaded on October 21, 2009 at 01:38 from IEEE Xplore. Restrictions apply.
V, the system could operate at its matching limit of 78% but if the battery voltage plus the voltage drops are equal to 15.5 V, the system matching would only be 68%. The limits of the mismatch between the actual PV array power and the rated PV array power are therefore controllable by using solar path tracking and/or cooling the PV panels. Real mismatch between the actual PV array power and the rated PV array power is improved by using an MPPT converter. However, when these techniques are used to improve system matching, the cost of solar path tracking, cooling and the MPPT converter must not exceed the cost of PV panels needed to offset the effects of bad system matching.
The advantage of using input and output parameters to achieve optimum or near-optimum control of dc-dc converters, are well known [l]. Feed-forward controlled converters have been implemented with minimum additional hardware for several converter topologies, e.g., for the Buck topology [l]. The feed-forward control method combines current mode control with the feed-forward of output current and input voltage to obtain enhanced dynamic response on transient changes i output current or input voltage. As shown in figure n 5 , the input and output voltage, and the output current is fed forward at ratios determined by the control law generator. An error signal is obtained from the difference between the reference for the output and the output voltage. This error signal is added to the feed-forward control parameter, I,, which is obtained from the control law generator. Note that the sign of the error signal is defined such that for output voltages higher than the reference voltage, the error signal is negative. The mathematical content of the control law generator for the basic topologies of power converters (i.e., Buck, Boost and Buck-Boost) was previously determined by Red1 and Sokal [l], e.g. for the Boost converter the control law generator performs the following mathematical function to obtain I,:
For the Buck converter, the feed-forward control parameter reduces to equation 2. This means that near optimum control of the Buck converter is obtained by feeding only the output current forward. This simplifies the implementation of feed-forward control for Buck converters:
Id = I m
\b
lOENER*TOR
i( ",
Fig. 5 : Diagram of feed-forward controller for PV Systems. Another simplification when using a Buck converter to charge a battery (figure 5 ) is that no output capacitor is necessary when the ripple current in the inductor is small. This also implies that the inductor current in the converter is also the output battery charging current. Using feed-forward control in converters for PV systems, the following advantages can be listed:
1242
Authorized licensed use limited to: Jawaharlal Nehru Technological University. Downloaded on October 21, 2009 at 01:38 from IEEE Xplore. Restrictions apply.
The magnitude, duration and energy content of output voltage transients, after a transient change of output current or input voltage is reduced. In general, smaller size and lower cost output filter capacitors can be used. Since the converter is controlled as a current source, several MPPT converters can be used in parallel to charge the same battery.
Simplified control for MPPT Several advantages i.e., high efficiency, low cable losses [SI and easy feed-forward control [l] makes the Buck converter, or Buck derived topologies, an excellent choice for MPPT converters. Normally the MPPT control algorithm that is used for maximum power point tracking converters requires a microprocessor [2], or the maximum power point is tracked by means of an indirect method [lo]. Some of these methods are therefore only feasible at higher power levels (1 kW power range). MPPT converters should however be cost-effective and should have a high reliability. For low power levels these requirements are only obtainable by converters of low complexities.
When the time constant of changes in the output voltage of the converter is small compared to the switching period of the switching element, it is only necessary to maximize the output current of the converter in order to track the maximum power point of the PV array. Maximization of the output current by means of feeding the output current back in a positive way, is obtained since the IV characteristics of the PV array are responsible for negative feedback when the output current tends to exceed the optimum current for maximum power delivery by the PV array. Voltage feedback from the load is only necessary to protect the load from overvoltage i.e. protecting a battery against overcharge and protecting a DC motor for water pumping against overspeed. The current regulator is implemented in such a way that the power electronic switch has a minimum duty cycle. The current regulator may comprise a pulse width modulator (PWM), a delta modulator (DM) or a hysteresis controller in the current-mode control arrangement [9]. A higher output current (and therefore higher output power) of the converter results therefore in a higher duty cycle for the current regulator CR, which is also the onloff duty cycle of the switching device@) in the converter.
-'
Near optimum feed-forward control for enhanced converter dynamics is already a simplified implementation of optimum control as proposed by Bums and Huffman [ll]. Feedforward control for maximum power point tracking in PV systems is further simplified [I21 to need only the output converter current as a feed-forward parameter. Figure 6 is a representation of the simplified feed-forward control approach for maximum power point tracking with any of the switch-mode power converters. The simplification comprises that the control law generator is removed and that only the output current is used as a feed-forward control parameter. A further simplification is that the voltage regulation loop is converted into an overvoltage protection loop. This overvoltage protection loop becomes only active when the output voltage rises above a predetermined limit. The error signal from the overvoltage protection loop will then reduce the control signal to the current regulator (CR) to limit the output voltage. Positive feedback of a signal proportional to the output current of the converter is therefore used as the input control signal to the current regulator, to perform maximum power point tracking. When the overvoltage protection loop is active, no maximum power point tracking can be performed. The fundamentals of the MPPT are:
A V , = - [1- A + B ]
c,
(4)
From equation 4 evidently when the integral part A is larger than the part B, the input capacitor voltage will decrease. The opposite applies when A is smaller than B. For the Buck converter, the relation between the variation in output current over one switching period of the converter as function of PV array voltage is given by equation 5 . In equation 5 , A i is the change in output current as function of the duty cycle (d), the PV array voltage () V , and the output voltage of the converter (Vb). From equation 6 it is clear that when the values of the integral function D is larger than the value of the integral function C, the inductor current will decrease. The opposite applies when D is smaller than C.
1243
Authorized licensed use limited to: Jawaharlal Nehru Technological University. Downloaded on October 21, 2009 at 01:38 from IEEE Xplore. Restrictions apply.
A iL
1 -
[C -
D]
As the cell temperature of the PV array increases such that C4 applies, the same reasoning as in the above paragraph can be used to show that the input voltage for the converter reduces. This reduction in input voltage will result in a reduction of the output current and due to positive feedback, the duty cycle is reduced. This will reduce the output current even further until the optimum output current and maximum power point is reached. Note that the internal characteristic of the PV array is again used to stop the positive feedback "run away". Finally, for the case of a combined reduction in insolation and cell temperature, (i.e. typical operating condition during midday) the input voltage and output current dropped to the level where the PV array current is higher than the output current. The input voltage will then increaseaccording to curve C5 until the maximum power point is reached. The maximum power point is thus tracked from point 0 to 5 in figure 7 .
10 15 Panel Voltage M
20
The MPPT operation is now further described by means of the I-V characteristics of the PV input source. A variety of I-V curves, representing different operating conditions for ii PV array using two solar panels in series is presented by figure 7 . The MPFT controller has to follow the maximum power points 1 to 5 as the PV array temperature and insolation parameters vary during a typical day. The minimum duty cycle for the converter is set to allow for maximum power point tracking of all insolation levels higher than 0.1 kW/m2, e.g., point 0. When the insolation level increases to curve C1, more current from the PV array is available and from equations 3 and 4, it is evident that the averaged input voltage of the converter increases. From equation 5 and 6, the averaged output current increases because of the increased input voltage (Vp). Because the output current is fed back in a positive way, the higher output current increases the duty cycle (d). From equations 5 and 6, an increase in duty cycle (d), increases the output current even further in the same direction. This output current "run away" continues until the maximum power point is reached. Note that the time constant for this loop is much longer than the switching period of the power switch. Note also that during the output current "run away", the input voltage of the converter reduces accordingly to the I-V curve C1 and that this slows the "run away" down as the point of maximum power is approached. If the output current would increase to a value higher than the optimum output current, the input voltage ol the converter would drop suddenly and the "current run away" is stopped. The internal IV characteristic of the PV array is therefore used to stop the effect of the positive fecdb:ick. Similarly the maximum power point is tracked from point ? to 3 when the insolation level increases to C3.
G - S O : K - l / l m O [v/A]
Fig. 8: Implementation of simplified Buck MPFT controller This feed-forward controller is implemented using the Buck converter topology. The accompanying diagram, figure 8, shows a circuit diagram of the practical implementation of the MPFT system, employing the feed-forward controller. The system comprised a PV array, consisting of 2x 55 W PV panels in series as energy source, an input capacitor C,, with associated series resistance R,, an MOSFET switch IRF 130, associated gate drive circuitry, a UES1402 freewheeling diode D, a large output inductor L, the 12 V battery as load and controller. The controller comprises a current shunt CS with associated attenuation of K = 1/1000 VIA, the feed-forward gain of the output current I,, G = 250, which forms the input Iff to the 3520 PWM integrated circuit. The duty cycle output d is then controlling the transistor T directly. The input of the PWM I,, is thus connected to the amplified output current of the converter that is proportional to the converter output power. The output of the PWM is connected to the converter to control the switching element in a way that maximizes the output power of the PV array. The higher t h e output power, the higher the duty cycle, thus positive feedback of the output current. This may be performed since the I-V characteristics of the PV array has only one maximum power point.
1244
Authorized licensed use limited to: Jawaharlal Nehru Technological University. Downloaded on October 21, 2009 at 01:38 from IEEE Xplore. Restrictions apply.
[2] AF Boehringer: "Self-adaptive dc converter for solar spacecraft power supply", IEEE Trans on AerosDace and Electronic Systems, AES-4, no 1, pp 102-111, Jan 1968. [3] JP Femeux; E Toutain; H LeHuy: "Multifunction DC-DC converter as versatile power conditioner", E = Trans on Indus. A d . , IA-22, no 6, 1037-1043,NovlDec 1986. [4] JHR Enslin; DB Snyman: "Combined low-cost, high efficient inverter, peak power tracker and regulator for PV applications", IEEE Trans on Power Electronics. PEL-6, no 1, 73 - 82, Jan 1991.
:
10 18
18
[S] J Applebaum: "The quality of load matching in a direct coupling photovoltaic system", IEEE Trans on Enerev Conversion, vol EC-2, no4, pp 534-541, Dec 1987.
--
20
22
94
Fig. 9: Experimental results with simplistic MPPT. The above-mentioned MPPT controller was tested on a 2x 55 W PV panels. Figure 9 shows the measured results from this maximum power point tracker employing the feed-forward controller. The output power of the converter is plotted as a function of the array voltage for the 2 PV panels, connected in series. The converter was charging a 12 V battery. The battery current is the only feedback parameter and is used in a feed-forward configuration to maximize the PV array output power. The maximum power points in figure 9 (A, B and C) were determined very quickly and tracked accurately. The power curves Power1 to Power3 were measured using the converter in an open-loop mode, before closing the control loop with the feed-forward current controller. The controller operates excellently under dynamically varying insolation, temperature and battery voltage.
[6] JHR Enslin; DB Snyman: "Field evaluation of a novel ma.ximum power point tracking regulator in PV installations", Proc. of EPE-91, vol I, pp 265-270, Florence, Italy, 3-6 Sept 1991.
[7] JHR Enslin: "Positive feedback controller for maximum power point tracking system", Provisional Patent Suecification, no 92/1387SA, Pretoria, South Africa, Feb 1992.
[8] DB Snyman; JHR Enslin: "Analysis and experimental evaluation of a new MPPT converter topologies for PV installations", Proc. IECON-92, San Diego, CA, USA, 9-13 Nov 1992.
[9] C Deisch: "Simple switching control method changes power converter into a current source", Proc. 9 th IEEE PESC78, Syracuse, NY, pp 300-306, 13-15 June 1978.
[lo] JJ Schoeman; JD van Wyk: "A simplified maximal power controller for terrestrial photovoltaic panel arrays", proC. 13 th IEEE PESC-82, pp 361-367, June 1982. [l:L] SD Huffman; WW Bums; TG Wilson; HA Owen: "Fast-response free-running dc-todc converter employing a state trajectory control law", Proc. 8 th IEEE PESC-77, Pa10 Alto, CA, pp 180-189, June 1977. [l:!] JHR Enslin; DB Snyman: "Simplified feed-forward control of maximum power point in PV installations", proC. IECON-92, San Diego, CA, pp 548-553, 9-13 Nov 1992. [13] JHR Enslin: "Riickkopplungsregler", German Patent Specification, no G 93 02 756.7, Germany, Feb 1993.
CONCLUSIONS
A new simplified, low cost MPPT feed-forward controller is described. Experimental results evaluating the performance of the controller under dynamic varying conditions are given. This controller can be utilized in small (one or two panel) and larger installations so that the total MPPT converter cost may be a fraction of the solar panel(s) cost. This controller may also be integrated into a power integrated circuit (PIC) for mass-production purposes. The characteristicsof PV panels are discussed, describing the principle and advantages of MPPT.
REFERENCES
[l] R Redl; NO Sokal: "Near-optimum dynamic regulation of dc-dc converters using feed-forward of output current and input voltage with current-modecontrol", IEEE Trans on Power Electronics, PE-1, no 3, pp 181-192, 1986.
1245
Authorized licensed use limited to: Jawaharlal Nehru Technological University. Downloaded on October 21, 2009 at 01:38 from IEEE Xplore. Restrictions apply.