Original article
A novel fuzzy controller for photovoltaic pumping systems driven by general-purpose frequency converters

https://doi.org/10.1016/j.seta.2020.100758Get rights and content

Abstract

In this paper, a fuzzy logic based Mandani-type controller is applied to an industrial frequency converter to couple a photovoltaic generator to a conventional centrifugal pump for water pumping applications. The system operates in closed loop and the DC bus voltage regulation of the frequency converter is done indirectly by varying the speed of the pump according to the power provided by the photovoltaic generator. In order to decrease the number of series-connected modules, the working voltage of the photovoltaic generator (maximum power point voltage) must be as close as possible to the minimum voltage limit required by the frequency converter’s input. The proposed algorithm was embedded in a general-purpose microcontroller. Controller performance was evaluated through tests on an experimental bench that allows simulating different water heads. Experimental results shows that the speed of the proposed fuzzy controller allows the operation close to the lower DC voltage limit imposed by the frequency converter, making possible the use of photovoltaic generators with a reduced number of PV modules in series (in this work, 30% fewer modules). The fuzzy controller regulates the photovoltaic generator voltage throughout the day, avoiding frequency converter limitation due to undervoltage errors, even with sudden changes in irradiance.

Introduction

Water and energy are the key drivers of agricultural production while certain regions of the world are facing severe energy and water crisis [1]. In Brazil, despite the abundance of water resources, a considerable part of its population lacks reliable access to water. Riverside populations in the Amazon, for example, devote a significant part of their time to transporting water for consumption and domestic use. Although often surrounded by water, these populations suffer since access to the energy sources necessary for water pumping is very restricted. As these regions are remote, the conventional electricity grid has low penetration and reduced probability of implementation in the future, while fossil fuels are expensive and difficult to access, thus in-site generation alternatives such as the solar photovoltaic (PV) have for some time become a viable reality [2]. Photovoltaic pumping system (PVPS) is an ideal alternative to the electricity and diesel-based water pumping systems [1], [3], [4] and has been a promising field of research for the last fifty years [1].

The structural outline of a PVPS has three main components: PV generator (PVG), power control system (PCS), and eletromechanic system containing motor and pump. Storage components may also be used in this system, such as batteries for electric charge storage or reservoir for water storage [1]. In most applications, the direct coupling between the PVG and the motor pump makes the reservoir the only storage component. The nonlinear relation between the water flow rate and solar power has led to the development of specific PCS for PVPS applications. Such PCS are designed to operate with a specific motor pump or a limited group of motor pumps, which usually restricts the number of options to be used.

A trend that has been worked on by some research groups is the use of conventional electric motor pumps coupled with standard industrial-use frequency converters (FC, also called Variable-speed Driver) [4], [5], [6], [7], [8], without the need of electrical energy storage. These systems are also recognized as directly coupled PV pumping systems, which implies on an appropriately sized water reservoir to meet the water demand when solar radiation is insufficient. In [3], for example, it is presented a simple mathematical model to be used in system sizing, for a given water head and solar resource, to calculate with good precision the water flow rate pumped by a commercial integrated solution. In [4], it is presented the long-term performance results of a direct coupled PV pumping system, operating over 28 years. The results are important to attest the reliability of such configuration. The paper presented in [5] was one of the earliest to present the use of standard FC applied to PV pumping, describing the various advantages of its application in comparison to special purpose PV pumping converters. In [6] it is presented a systematized procedure to tune a PID controlled FC used to supply a standard water pump with PV source. In [7], a directly coupled configuration is used, in a setup like the one presented in [6] but using fuzzy logic to control the converter. A review of variable-speed drivers in PV pumping applications for irrigation in Brazil is presented in [8].

The direct coupling configuration using standard general-purpose FC has proven to be a viable alternative in relation to equipment designed specifically for PVPS [5], [6], [9]. The advantage of a well-consolidated market, a high degree of reliability, easy acquisition, even in underdeveloped countries, and a wide range of manufacturers and power ratings are some of the important advantages of FC use, as highlighted by [5], [6]. However, in practical applications there have been some problems related to the adaptation of the FCs to the PVPS operating requirements [9]. According to [9], one of the problems is the difficulty of implementing maximum power point tracking (MPPT), which was solved through the addition of a basic industrial programmable logic controller (PLC) to the system. In such application, the photovoltaic generator (PVG) directly feeds the FC's DC bus and the FC’s internal Proportional-Integral-Derivative (PID) controller operates in closed loop, being the PV voltage the variable to be controlled. By varying the speed of the pump according to the available PV power, the PID must keep the PV pumping system (PVPS) operating as close as possible to the PVG’s maximum power voltage, with good regulation in order to optimize the system performance and minimize loss of efficiency and interruptions in pumping caused by sudden variations in irradiance.

The interruptions in system operation are caused by sudden variations in the solar resource and occur due to the controller's inability to maintain the voltage of the FC bus at an adequate level, which causes the device to halt due to an undervoltage error on the DC bus [7]. The sudden stop produces overvoltage on the FC bus in addition to pressure spikes in the pipeline (water hammer) caused by the abrupt change in the pumping flow [10]. To overcome this, in [10] was developed an algorithm that enables or disables the PID controller of the FC according to the absence or presence of clouds, and was implemented in FC’s controller in order to avoid blockage the FC (sudden stops). The voltage compatibility between the PVG and the FC limits the number of PV modules in series to keep the PVG voltage between the maximum and minimum input voltage of the FC [5], [11]. This limitation can produce losses in terms of PV production and pumped water [11].

The FC-coupled PVPS has a simple structure, however, the implantation cost for small systems is high [8] due to the need of a significant number of photovoltaic panels connected in series, as the adequate operating voltage must comply with the voltage range of the FC and the control dynamics employed. Thus, reducing the number of modules in series can help to reduce the costs for PVPS with FC in domestic applications or for irrigation of small crops, like those applications presented in [12], [13].

In PV power systems, both PV modules and switching-mode converters present nonlinear and time-variant characteristics, which result in a difficult control problem [3], [14]. The PID is an automatic controller that have their performance compromised when working with systems that have these characteristics, not presenting satisfactory responses for operating conditions other than the tuned conditions. In this scenario, the use of intelligent controllers based on fuzzy logic becomes attractive due to their ability to deal with non-linear and time-variant systems, in addition to dispensing the knowledge of the mathematical model of the system. Although it adds one more element in the PVPS configuration, an external controller allows the use of FCs that do not have an embedded PID controller [7], furthermore, other features can be added to the control firmware to improve system performance.

In [9] a standard programmable logic controller (PLC) was added to a PVPS that uses an industrial FC. The PLC allows the implementation of several control algorithms, such as the monitoring of MPPT and other improvements. Sudden stops due to drops in irradiance have been overcome and the overall performance of the system has been improved. However, the PLC is not a tool that can be easily embedded inside the FC and requires more specialized personnel to handle it.

In [7], [15] fuzzy controllers are proposed, external to the FC, operating in open loop and whose control action is based on obtaining the dynamic behavior of the irradiance. To support MPPT functionality, two different sensing methodologies have been proposed: using an LDR (Light Dependent Resistor) [15] and a low power reference photovoltaic module [7]. The reference value (set point) in these strategies is variable and dependent on irradiance. One of the disadvantages observed is the use of one more external component in the system (irradiance sensor), increasing the probability of failure and hindering its implementation in the existing structure of the standard FCs. The control proposed in this work is easy to implement in the existing structure of the FC (basically programming) and consists of using of fixed set points (may evolve to MPPT in the future), that when are well-adjusted according on the GFV configuration and the local characteristics, have very satisfactory results when compared to systems with MPPT. In [4] the losses occasioned by this practice are less than 4% with respect to a system with MPPT.

Unlike what was proposed in [7], [15], in this work the fuzzy controller operation occurs without sensors in closed loop being the FC’s DC bus voltage the controlled variable. The operating point was set to 207 V DC, which is close to both the minimum voltage allowed for the FC bus and the maximum power point (MPP) of the PVG used in the tests.

One of the objectives of this work is to reduce as much as possible the number of modules in series that compose the PVG, without the need to introduce additional power converters, which implies making the system operate at a voltage level closer to the minimum voltage required by the FC on its DC bus, demanding greater regulation efficiency from the controller. Taking into account the inability of the PID controller to maintain the appropriate voltage level for this condition when the system is subjected to variations in irradiance, a fuzzy controller based on Mamdani-type rule external to the FC is proposed in this work. Mamdani type control is widely used in industry and has proven usefulness in control engineering [16].

The novelty of the work lies in the development of a control strategy oriented to reduce the number of PV modules in series that are needed to supply a FC-coupled PVPS, while keeping minimal reduction in performance when compared to traditional PID control. To the best of authors’ knowledge, this is the first paper that presents a control framework with this objective. In [11] it was evaluated the optimal number of modules in series in a PVPS from a performance standpoint in large pumping systems, while our work aims to reduce the number of modules required to guarantee proper and efficient system operation, and is mainly focused on small to medium scale pumping systems.

In practice, in the specific case of the control methodology proposed in this work, it can be easily embedded in the commercial FCs programming, allowing it to be easily parameterized, like any other functionality of the FC, in order to support practical use by lay men. It is important to note, however, that basic knowledge is required on industrial variable speed driver parametrization, as would be the case for any methodology using such industry standard devices.

The paper is structured as follows: in “Photovoltaic pumping system and test facility description” is presented the technical description of the photovoltaic pumping system and the test bench used for the experimental evaluations; “Proposed fuzzy controller” shows the design of the fuzzy controller; “The embedded controller” discusses the system embedded in the microcontroller; in “Results and discussions” the experimental results are presented as well as the discussions on the results, followed by the conclusions of the work.

Section snippets

Photovoltaic pumping system and test facility description

The PVPS has a PV generator composed by the series association of 13 modules model S-55P, from the manufacturer Solares, where each module is rated at 55 Wp and is composed by 36 cells connected in series. The rated parameters of this module are listed in Table 1.

The FC used is the CFW-10, model 7.3 A/200–240 V – WEG manufacturer, being all FC operation configured to be commanded via external terminals. In their parameterization (programming) the acceleration and deceleration times must be

Concept

The fuzzy controller is designed to automate the way a human expert, who is successful at controlling the PVPS, performs such a procedure. Initially, the expert (the fuzzy controller designer) chooses what information will be used as inputs to the decision-making process. Usually, an error function is used and the error function variation as the variables on which decisions are based. The variable to be controlled is the PVG voltage and the control variable (controller output) is the increment

The embedded controller

Fig. 7 shows the block diagram of the PVPS control system embedded in a PIC 18F2550 microcontroller. The set point is defined on the code and the voltage divider adjusts the PV voltage to be properly applied on the microcontroller (0–5 V). The resulting voltage from the voltage divider is converted to digital signal by the Analog to Digital Converter (A/D), which is embedded in the microcontroller.

The error value, e(k), and its variation, de(k), are calculated and then processed through the

Results and discussions

Fig. 10 shows, for a complete pumping day, the incident irradiance in the PVG plane, the pressure in the pipe (head height of a pumping system), the PVG voltage and power, and water flow rate. The incident solar resource for that day was 6.49 kWh/m2 or 6.49 peak sun hours (PSH). The fuzzy controller regulates the PVG voltage at the specified reference value of 207 V throughout the day, avoiding FC limitation due to undervoltage errors, even with sudden changes in irradiance. Some previous

Conclusions

In this paper a fuzzy logic based Mandani-type controller was applied to an industrial frequency converter to couple a photovoltaic generator to a conventional centrifugal pump for water pumping applications. The tests presented a robust performance of the fuzzy controller under irradiance and temperature variations. The controlled variable (PV generator voltage) followed the specified reference, over the entire pumping cycle, even with abrupt irradiance variations, which avoided the FC’s halt

Funding

This work was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) under the Instituto Nacional de Ciência e Tecnologia de Energias Renováveis e Eficiência Energética da Amazônia (INCT-EREEA) and by the Grupo de Estudos e Desenvolvimento de Alternativas Energéticas (GEDAE-UFPA).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (31)

Cited by (0)

View full text