Implementation of Type-2 Fuzzy Logic Control in an Internet of Things-Based Solar-Wind Hybrid Power Generation System to Improve Stability and Energy Efficiency

2.1 System design

This study proposes a HRES that integrates solar PV generation, wind power generation, battery energy storage, an intelligent control system, and IoT-based monitoring. The proposed system is designed to manage the distribution of electrical energy generated by the solar PV panel and the Savonius wind turbine to the battery bank and electrical loads using T2-FLC. The primary objectives of the proposed system are to ensure continuous power supply, improve battery charging performance, reduce relay chattering, and provide real-time system monitoring. The main hardware components consist of a 200 W PV panel, a Savonius Vertical Axis Wind Turbine (VAWT), a 24 V battery bank, a Maximum Power Point Tracking (MPPT) hybrid charge controller, a Direct Current-Alternating Current (DC-AC) inverter, a PZEM-017 direct current power sensor, a Direct Current Solid-State Relay (DC-SSR), an Arduino Mega microcontroller, and an ESP32 communication module. The PV panel and the wind turbine function as renewable energy sources, whereas the battery bank stores surplus electrical energy for subsequent use when renewable energy generation decreases. The Arduino Mega executes the T2-FLC algorithm to determine the optimal energy distribution strategy, while the ESP32 transmits monitoring data to the IoT platform for remote observation. The operating principle of the proposed system begins with electricity generation from the PV panel and the Savonius wind turbine. The generated electrical parameters, including voltage, current, and power, are measured by the PZEM-017 sensor and transmitted to the Arduino Mega. These measurements are processed as input variables for the T2-FLC algorithm. Based on the fuzzy inference results, the controller determines whether the available energy should be directed to charge the battery bank, supply the electrical load, or activate the backup power source. Finally, the ESP32 continuously transmits the operating data to the IoT monitoring platform, enabling real-time monitoring, data logging, and remote system supervision. The overall architecture of the proposed hybrid solar-wind energy management system is illustrated in Figure 1.

Figure 1. Design of a hybrid solar-wind system with Fuzzy Logic Control (FLC) and Internet of Things (IoT) monitoring

2.2 Hardware implementation

The hardware implementation integrates renewable energy generation, energy storage, sensing, control, actuation, and communication components into a single laboratory-scale HRES. A 200 W PV panel serves as the primary energy source during daytime operation. Under standard operating conditions, the PV panel provides an operating voltage of approximately 18-20 V and a maximum output current of 10-11 A, as shown in Figure 2(ii). The electrical output of the PV panel is influenced by several environmental factors, including solar irradiance, ambient temperature, panel orientation, tilt angle, and PV conversion efficiency. The performance of PV systems is strongly influenced by several environmental and operational factors, including solar irradiance, ambient temperature, PV module technology, and surface soiling, all of which directly affect conversion efficiency and electrical power output [33-36]. These factors should therefore be considered when evaluating the performance of PV-based HRES. A Savonius VAWT, shown in Figure 2(i), is employed as the secondary renewable energy source. The Savonius turbine was selected because of its simple mechanical structure, high starting torque, ability to operate at low wind speeds, and suitability for environments with multidirectional wind conditions. The turbine converts the kinetic energy of the wind into mechanical energy, which is subsequently converted into electrical energy by a permanent magnet generator. The generated electrical power varies according to wind speed, turbulence intensity, and wind direction. Savonius vertical-axis wind turbines have been widely investigated for small-scale renewable energy applications because of their high starting torque, simple mechanical construction, and reliable operation under low and multidirectional wind conditions. These characteristics make Savonius turbines particularly suitable for HRES intended for distributed generation and smart grid applications [37, 38]. A 24 V battery bank functions as the energy storage unit by storing surplus electrical energy generated from the PV panel and the wind turbine. During periods of low renewable energy generation, the battery bank supplies electrical power to maintain continuous operation of the load. Battery voltage is used as one of the principal input variables in the control system because it serves as a practical indicator of the battery operating condition and supports energy management decision-making.

Battery Energy Storage Systems (BESS) play a crucial role in HRES by balancing the intermittent output of renewable energy sources, improving supply reliability, and supporting intelligent energy management. Effective battery monitoring and charging management are therefore essential to maintain battery performance, operational safety, and long-term system reliability. Various battery management and state estimation techniques have been developed to improve the efficiency and reliability of renewable energy storage systems [39-41].

Electrical parameters, including voltage, current, power, and accumulated energy, are measured using a PZEM-017 direct current power sensor. These measurements are transmitted to the Arduino Mega as input variables for the proposed T2-FLC algorithm. Load switching is performed using a DC Solid-State Relay (SSR-40DD), which is controlled by the Arduino Mega according to the control decisions generated by the T2-FLC. The SSR-40DD operates with a control input voltage of 3-32 VDC, an output operating voltage of 5-60 VDC, and a rated switching current of 40 A, making it suitable for low-voltage DC switching applications in HRES. The control algorithm is executed every 5 s, and the relay switching state is updated at the same interval. Compared with conventional electromechanical relays, the SSR provides high-speed electronic switching without mechanical contacts, thereby eliminating contact bouncing, reducing mechanical wear, and improving switching reliability under frequent operating conditions. In this study, relay chattering is defined as repeated and unnecessary switching events occurring within a short operating interval due to rapid fluctuations in PV power generation, wind turbine output, and measurement uncertainty. The proposed T2-FLC suppresses these unnecessary switching operations, thereby improving battery voltage stability and reducing stress on the switching device. Battery charging is managed using a hybrid MPPT charge controller, which maximizes energy harvesting from both renewable energy sources while providing built-in battery protection against overcharging and excessive discharge through its charging management algorithm. The MPPT controller also incorporates reverse-polarity and overcurrent protection to ensure safe operation of the laboratory-scale prototype throughout the experimental period. A DC-AC inverter converts the stored direct current electrical energy into alternating current electricity suitable for supplying conventional electrical loads. The complete hardware implementation, including the PV panel, Savonius wind turbine, battery bank, sensing devices, control unit, switching actuator, and power conversion components, is presented in Figure 2(iii).

Figure 2. Hardware implementation of the proposed hybrid solar–wind power generation system: (i) Savonius wind turbine, (ii) photovoltaic (PV) panel, and (iii) laboratory-scale prototype

2.3 Internet of Things monitoring system

The proposed monitoring system employs an Arduino Mega as the primary controller for data acquisition and execution of the T2-FLC algorithm. The Arduino Mega was selected because it provides a sufficient number of input-output pins, supports multiple communication interfaces, and is compatible with various sensors and actuators used in the proposed HRES. The controller continuously receives electrical measurements from the PZEM-017 direct current power sensor, including voltage, current, and power generated by the PV panel and the Savonius wind turbine, as well as the battery voltage. The acquired sensor data are processed by the Arduino Mega to determine the operating conditions of the HRES. These measurements are used as input variables for the T2-FLC algorithm, which determines the appropriate energy distribution strategy according to the available renewable energy and the battery charging condition. Based on the fuzzy inference results, the controller regulates the DC SSR to control energy flow between the renewable energy sources, the battery bank, and the electrical load. This adaptive control strategy minimizes unnecessary relay switching caused by small variations in sensor measurements, thereby improving system stability and reducing relay chattering.

To support remote monitoring, an ESP32 microcontroller is employed as the IoT communication module. The ESP32 transmits operational data to the IoT monitoring platform through a wireless network in real time. The transmitted parameters include PV panel voltage, PV panel current, PV panel power, wind turbine voltage, wind turbine current, wind turbine power, battery voltage, battery power, relay status, and system operating mode. The IoT monitoring platform provides continuous visualization of the operating conditions of the HRES and enables real-time data logging, remote supervision, and historical performance analysis. These capabilities allow system operators to evaluate energy generation, battery charging performance, relay operation, and overall system efficiency without requiring direct physical access to the experimental prototype. The developed monitoring interface is presented in Figure 2(iii).

2.4 Type-2 Fuzzy Logic Control design

The proposed T2-FLC algorithm is designed to determine the optimal energy distribution strategy based on the operating conditions of the renewable energy sources and the battery bank. The controller receives three input variables, namely PV power, wind turbine power, and battery voltage. PV power represents the electrical power generated by the PV panel, wind turbine power represents the electrical power generated by the Savonius wind turbine, whereas battery voltage is used as a practical control variable to represent the battery operating condition during charging and discharging. Each input variable is represented by three linguistic terms: Low, Medium, and High. The Low membership function represents insufficient power generation or a low battery voltage, the Medium membership function represents normal operating conditions, and the High membership function represents abundant renewable energy generation or a high battery voltage.

In this study, battery voltage was used as an indicator of the battery operating condition for control purposes rather than as a direct State of Charge (SOC) measurement. The proposed controller utilizes battery terminal voltage to determine the charging and discharging operating mode because voltage can be measured continuously using the installed DC sensor. However, battery terminal voltage is affected by charging current, load variations, and temperature; therefore, it does not represent the actual SOC under all operating conditions. More accurate SOC estimation methods, such as coulomb counting or model-based estimation, are beyond the scope of this study and will be considered in future work.

Unlike conventional Type-1 FLC, the proposed T2-FLC represents each membership function as an interval bounded by the Upper Membership Function (UMF) and the Lower Membership Function (LMF). The region enclosed between the UMF and LMF is referred to as the FOU, which enables the controller to model uncertainties arising from sensor noise, measurement errors, and environmental disturbances. Consequently, the controller is capable of producing smoother and more robust control actions under dynamic operating conditions. The T2-FLC algorithm consists of four sequential stages: fuzzification, rule inference, type reduction, and defuzzification.

The theoretical foundation of fuzzy set theory was first introduced by Zadeh [42]. Subsequent developments in Interval Type-2 Fuzzy Logic Systems (IT2FLS) have demonstrated improved capability for representing uncertainties arising from measurement noise, parameter variations, and nonlinear system behavior. Recent studies have shown that IT2FLS-based controllers provide enhanced robustness and more reliable decision-making than conventional Type-1 fuzzy systems, particularly in renewable energy and intelligent control applications [43-47]. During the fuzzification stage, the measured input variables are transformed into interval-valued fuzzy sets according to their corresponding membership functions. Subsequently, the inference engine evaluates a predefined rule base to determine the appropriate operating mode of the HRES. Examples of the fuzzy rules employed in this study are presented in Table 1.

Table 1. Representative fuzzy rules

The proposed T2-FLC employs a complete fuzzy rule base consisting of 27 rules generated from three input variables (solar power, wind power, and battery voltage), each having three linguistic states (Low, Medium, and High). Owing to space limitations, only representative rules are presented in Table 1. All 27 rules were fully implemented in the Arduino Mega controller during experimental testing. The fuzzy inference process generates one of several operating modes, including Grid, Battery, Balanced, Solar Priority, and Wind Priority. The Grid mode is activated when renewable energy generation is insufficient and the battery voltage is low.

The battery mode is selected when renewable energy generation is adequate and stored energy is sufficient to supply the electrical load. The Balanced mode distributes power proportionally between the renewable energy sources and the battery bank. The Solar Priority and Wind Priority modes prioritize the dominant renewable energy source according to the available generation. Following the inference stage, the type reduction process converts the interval Type-2 fuzzy output into an equivalent Type-1 fuzzy set. In this study, the Karnik-Mendel algorithm is employed to calculate the left and right endpoints of the type-reduced set, denoted by $y_l$ and $y_r$, respectively. The final crisp control output is then obtained through the defuzzification process using:

$y=\frac{y_l+y_r}{2}$          (1)

where, $y$ is the crisp control output obtained after defuzzification; $y_l$ is the left endpoint produced by the type-reduction process; $y_r$ is the right endpoint produced by the type-reduction process.

The resulting crisp output is subsequently used to control the DC SSR and determine the charging priority, energy distribution strategy, and operating mode of the HRES. By incorporating the FOU into the decision-making process, the proposed T2-FLC provides greater robustness against measurement uncertainty, sensor noise, and rapid fluctuations in renewable energy generation. The overall control algorithm is illustrated in Figure 3.

Figure 3. Flowchart of the proposed energy distribution algorithm based on Type-2 Fuzzy Logic Control (T2-FLC)

2.5 Testing scenario

The experimental evaluation was conducted to assess the performance of the proposed hybrid solar-wind power generation system in terms of renewable energy generation, battery charging performance, intelligent energy management, relay switching behavior, and IoT-based monitoring. The experiments were performed under actual operating conditions using the laboratory-scale prototype described in the previous section. The first testing scenario evaluated the performance of the PV panel. Electrical parameters, including voltage, current, and output power, were measured under different levels of solar irradiance to investigate the relationship between environmental conditions and PV power generation. The second testing scenario focused on the performance of the Savonius VAWT. The output voltage, current, and electrical power generated by the wind turbine were measured under various wind speed conditions. This experiment was conducted to evaluate the capability of the turbine to generate electrical energy under low and fluctuating wind conditions. The third testing scenario evaluated the charging and discharging performance of the 24 V battery bank. Battery voltage was continuously monitored during charging and energy supply processes to verify charging stability and to ensure that the proposed energy management strategy effectively prevented overcharging and deep-discharge conditions. The fourth testing scenario assessed the performance of the T2-FLC algorithm. The controller was evaluated by applying different combinations of PV power, wind turbine power, and battery voltage as input variables. The resulting operating modes and energy distribution decisions were analyzed to verify their consistency with the predefined fuzzy rule base. The fifth testing scenario investigated relay switching performance and relay chattering reduction. The number of relay switching operations before and after implementing the T2-FLC algorithm was recorded and compared. The observed parameters included switching frequency, relay operating stability, and controller response under fluctuating renewable energy conditions. The final testing scenario evaluated the performance of the IoT monitoring system. Operational data transmitted by the ESP32 communication module, including voltage, current, power, battery voltage, relay status, and system operating mode, were monitored to verify the reliability of real-time data transmission, remote monitoring capability, and data logging performance. The IoT monitoring system was considered successful when the transmitted data accurately represented the actual operating conditions of the hardware prototype throughout the experimental period.

This study successfully implemented a T2-FLC algorithm in a laboratory-scale hybrid solar-wind power generation system integrated with an IoT-based monitoring platform. The proposed system combines a PV panel, a Savonius wind turbine, a battery bank, an Arduino Mega controller, an ESP32 communication module, and a PZEM-017 direct current power sensor to perform intelligent energy management under actual operating conditions. The experimental results demonstrated that the proposed controller effectively maintained the battery voltage within the operating range of 23.9-24.5 V throughout the seven-day testing period, indicating stable battery charging and energy distribution despite fluctuations in renewable energy generation. Furthermore, the implementation of the T2-FLC algorithm reduced relay switching frequency by approximately 40%, thereby mitigating relay chattering and improving the operational reliability of the HRES. The developed IoT monitoring platform successfully provided real-time data acquisition, remote supervision, and historical data logging, enabling continuous observation of system performance and supporting more efficient operation and maintenance. These capabilities improve the practical applicability of the proposed system for small-scale smart grid and distributed renewable energy applications.

Overall, the experimental findings confirm that integrating T2-FLC with IoT technology provides an effective solution for improving voltage stability, reducing unnecessary relay switching, and enhancing intelligent energy management in hybrid solar-wind renewable energy systems. Future research will focus on implementing adaptive optimization techniques for automatic fuzzy rule tuning, integrating additional renewable energy sources and battery management strategies, and evaluating the proposed system under larger-scale microgrid operating conditions.