[align=left] 1. Introduction Independent solar-powered water pumping systems have broad application prospects for solving ecological problems such as drinking water for humans and livestock, agricultural irrigation, and shelterbelt irrigation in remote areas, and have therefore received widespread attention. For water pumps driven by AC motors, an inverter controller is required. The controller must have the following two functions: inverter function—converting the DC power output from the solar array into the three-phase AC power required to drive the motor; maximum power point tracking (MPPT) function—adjusting the output frequency in real time according to the light intensity to ensure the system outputs maximum power. Compared with control systems using a "general-purpose frequency converter + maximum power point tracking controller," dedicated frequency converters integrate inverter and maximum power point tracking control functions, offering advantages such as simple structure, high reliability, good dynamic characteristics, and low cost. In addition to ensuring that the hardware circuitry and structure meet technical specifications, the design of the maximum power point tracking control law is crucial for the design of dedicated frequency converters. Extensive research has been conducted on maximum power point tracking (MPPT) algorithms, which have been widely adopted and applied in solar-powered water pumping systems. Examples include the atmospheric pressure method, perturbation observation method (PAO), and increased admittance method (ICT). However, these classic methods struggle to achieve ideal control performance in battery-free solar-powered water pumping systems. While mature dedicated frequency converters for solar-powered water pumping systems exist internationally, such as the Solartronic series from Denmark's Grundfos, their rated voltage is too low to be compatible with domestically produced water pumps. Domestic research and development of related products has also been undertaken, but due to extremely limited market access in previous years, mass production has not yet materialized. Furthermore, existing products all employ the classic MPPT method, and their control characteristics require further improvement. This development project, based on research into a solar-powered drip irrigation test system for sand-fixing forests along the Xinjiang Hade desert highway, proposes a multi-criteria and hybrid MPPT method. Furthermore, based on the parameters of the solar cell array and the performance of the microprocessor (MPU), the minimum step size of the frequency converter's output frequency is optimized. On this basis, a dedicated frequency converter using a fixed-point MPU is developed. To verify the control effects of various methods, two identical solar-powered water pumping test systems were established for comparative experiments and long-term continuous operation experiments. The results of the two-year experiment show that the new method has better dynamic response characteristics and maximum power point tracking performance; the dedicated frequency converter has excellent technical indicators, stable performance, and high reliability. 2. Composition and Characteristics of an Independent Solar-Powered Water Pumping System The independent solar-powered water pumping system described in this paper consists of three parts as shown in Figure 1: a solar cell array, a dedicated frequency converter, and a water pump using a squirrel-cage three-phase asynchronous motor. Based on the design concept that water storage is better than electricity storage in a solar-powered water pumping system, no battery was configured to simplify the system and reduce costs. A solar cell is a nonlinear power source, and its output characteristics are closely related to factors such as structure, materials, light intensity, and ambient temperature. Figure 2 shows the characteristic curves of output voltage, current, and power. As shown in Figure 2, there is a maximum value in the output power, i.e., the maximum power point. The output current-voltage characteristic is bounded by the maximum power point; the current remains basically constant in the region to the left, called the constant current region, while the voltage changes less in the region to the right, called the constant voltage region. Figure 1. Basic structure of an independent solar-powered water pumping system. Figure 2. PV (IV) characteristics of the solar cell array. A dedicated frequency converter is used to control the pump speed. Its control circuit is based on an MPU to achieve DC-AC inverter control and maximum power point tracking. In an independent solar-powered water pumping system without batteries, the power required by the pump is directly provided by the solar cells. The DC output voltage of the solar cell array decreases as the frequency converter output frequency increases. Therefore, by adjusting the frequency converter output frequency, the operating point of the solar cell array can be controlled. Increasing the output frequency in the constant voltage region and decreasing the output frequency in the constant current region can achieve maximum power point tracking. At the same time, since there is no battery to support the DC voltage, when the pump speed is too high or the light intensity decreases rapidly, if the frequency converter output frequency is not adjusted down in time, the voltage of the solar cell array will drop rapidly, causing the frequency converter undervoltage protection to activate. Therefore, higher requirements are placed on the dynamic characteristics of the controller. 3. Design of the dedicated frequency converter 3.1 Design of the main circuit and structure The circuit structure of the developed dedicated frequency converter is similar to that of the general frequency converter, as shown in Figure 3. The key design points are as follows: (1) The solar cell array outputs DC power, which can directly power the inverter circuit. To prevent current from flowing back to the solar cell array, a reverse blocking diode is installed in the input circuit. (2) The smoothing capacitor provides some support for the DC voltage, which is beneficial to improving the system characteristics. Since the input is DC voltage, a smaller capacitor can be configured, which is half or even smaller than that of a general-purpose inverter. (3) Because it is mainly used in remote areas, the reliability of the inverter is very important. Therefore, the inverter circuit adopts intelligent power modules with high integration and good reliability. (4) The filtering and absorption circuits are reasonably designed to achieve the same anti-static, pulse group and lightning surge interference capabilities as general-purpose inverters. (5) The low-power inverter (below 2.2kw) adopts natural air cooling, and the high-power inverter adopts forced air cooling. The cooling fan is controlled according to the radiator temperature to reduce the running time and extend the maintenance cycle as much as possible. (6) The environmental conditions are relatively harsh. The all-metal shell is adopted with a protection level of not less than IP53. Figure 3 Basic structure of dedicated frequency converter 3.2 Maximum power point tracking control law design The maximum power point tracking control law needs to meet three requirements: tracking speed, tracking accuracy and system stability. In an independent solar pumping system without a battery as voltage support, the stability problem is mainly the voltage stability problem. If the dynamic characteristics of the control law are not ideal, the output voltage of the solar cell array will drop sharply, the frequency converter will be undervoltage protected, and the system will stop running. The constant voltage method can solve the system stability problem better because it directly controls the operating point voltage. Its basic principle is to approximate that the voltage of the maximum power point is constant under different conditions. When the measured voltage is greater than the set maximum power point reference voltage vref during operation, the output frequency is increased; otherwise, the output frequency is decreased. The output frequency f is determined by equation (1), where δf>0 is the frequency adjustment step size. However, the selection of vref depends on the characteristics of the solar cell array. In actual operation, it may deviate from the maximum power point. Therefore, the reference voltage should be adjusted in real time according to the operating status of the system. In the preliminary research of this project, a multi-criteria method as shown in Table 1 was proposed. By detecting DC voltage and input current, the operating point of the system is adjusted directly using the signs of δp, δv, and δi as criteria. Since it can simultaneously discriminate the control effect of output frequency adjustment and changes in light intensity, it overcomes the shortcomings of classical methods, and the dynamic tracking characteristics of the system are improved. The dedicated frequency converter adopts a hybrid maximum power point tracking control combining the multi-criteria method and the constant voltage method. The multi-criteria method and the constant voltage method operate alternately. The execution cycle of the multi-criteria method is less than 1 second, tracking the maximum power point in real time and providing the optimal reference voltage value for the next execution cycle of the constant voltage method (approximately 10 seconds). Therefore, the system has both fast dynamic tracking characteristics and good stability. Appendix: Multi-criteria method logic 3.3 Minimum step size optimization During the execution of the multi-criteria method, the minimum step size δfmin directly affects the tracking accuracy. If δfmin is too large, the operating point will swing around the maximum power point, resulting in low tracking accuracy; if δfmin is too small, the corresponding δv and δi will be very small. After passing through the voltage sensor, current sensor and A/D converter, the MPU cannot identify them, thus causing the maximum power point to be judged incorrectly, and the system may run at the low power point for a long time. Based on the parameters of the solar cell array, voltage/current sensor and A/D converter, combined with the experimental results, the minimum step size can be optimized using formula (2). Where: vo—open circuit voltage of solar cell array vmp, imp—voltage and current of maximum power point frate—rated operating frequency of water pump kv, ki—gain of voltage/current sensor vad—reference voltage of A/D converter n—number of bits of A/D converter 3.4 Design of control circuit The control circuit is based on the Fujitsu 90F462 control chip and mainly has the following control functions: (1) Inverter control function. Two-phase pulse width modulation is adopted to reduce switching losses while ensuring the output sinusoidal current. The carrier frequency is adjustable and is set to 5kHz at the factory. (2) Maximum power point tracking control function. Based on the measured system response characteristics, in order to ensure the normal operation of the maximum power point tracking function and to make the system have the highest possible response speed, the reasonable control cycle is tens of milliseconds. (3) Fully automatic operation function. The system starts or stops automatically according to the voltage of the solar cell array. The optimized start control reduces the number of repeated starts. (4) Operation data storage function. In order to facilitate the monitoring of the unattended system operation status, the control circuit is equipped with a calendar and a storage chip, which can save up to 8 years of operation data and access it through the operation panel or communication port. (5) Water level monitoring function. Multiple water level sensor interfaces are provided to monitor the water level of the reservoir and deep well, which prevents the reservoir from overflowing and the water pump from overheating and being damaged due to dry pumping. (6) Complete protection functions. Including undervoltage protection, overvoltage protection, overcurrent protection and module overheat protection. 4. Experimental system and results In July 2004, two identical systems were built on the campus of Tsinghua University Shenzhen Graduate School (Figure 4). Each system's solar cell array has a peak power of 2200W; the inverter has a rated power of 1500W and a maximum output frequency of 50Hz; the three-phase 220V submersible pump has a rated input frequency of 50Hz, a power of 1500W, and a flow rate/head of 14 (m³/h)/20m; four outlets at heights (6, 11, 16, and 19m) were set up for comparative experiments. Figure 5 shows the output current waveform of the inverter; the 5kHz switching frequency ensures a good sinusoidal waveform for the output current. Figure 6 shows the experimental results confirming the tracking characteristics of the multiple criterion method; the output power curve and the light intensity curve match very well, indicating that the system has excellent tracking characteristics. Figure 7 shows the experimental results using hybrid maximum power point tracking control; it performs well under different weather conditions, especially when the light intensity changes rapidly, it can adjust the output frequency in a timely manner, effectively control the DC voltage, and ensure stable system operation. After a year of operational testing in 2005, the frequency converter operated stably, and the temperature rise of key components remained within acceptable limits. Each system generated a total of 1940 kWh of electricity and pumped 15600 m³ of water (11 m head) throughout the year. Figure 4 shows the independent solar-powered water pumping test system. Figure 5 shows the frequency converter output current waveform. Figure 6 shows the tracking characteristics using the multiple criterion method. Figure 7 shows the output curves under different weather conditions . 5. Conclusion The developed solar-powered water pumping frequency converter exhibits stable performance and high reliability. Employing hybrid maximum power point tracking control, it meets the dynamic control requirements of battery-free independent solar-powered water pumping systems. Its fully automatic operation and data storage functions are suitable for the operation and management of unattended systems in remote areas. Its robust protective structure ensures normal operation in harsh environments. The series of products, with rated power ranging from 400 W to 11 kW, will be mass-produced and marketed.