Abstract: This paper introduces a low-cost thyristor switched capacitor (TSC) device suitable for distributed reactive power compensation in low-voltage (400 V) distribution networks. It achieves hardware-locked protection by using the detection of zero voltage across the thyristor-type contactless switch (SCR ) as a prerequisite for SCR triggering, thus avoiding damage to components caused by inrush current from false triggering. It also prevents reactive power overcompensation, simplifies design, and reduces costs. Keywords: Zero-voltage switch; Reactive power compensation; Thyristor Introduction In recent years, the rapid development of the power system has gradually changed the tight power supply situation in China, with some regions experiencing oversupply. This has made it possible to consider the economic operation of the power system and the activation of the electricity market. Users have also raised higher requirements for power supply quality. Nationwide distribution network transformation is underway, and it is expected that the power supply situation for urban and rural industrial and residential users will be greatly improved. The increase in voltage qualification rate will enhance the enthusiasm of urban and rural residents, especially farmers, for electricity use. To cooperate with the power grid transformation project and meet the needs of voltage control and local distributed reactive power compensation, a new type of low-cost thyristor switched capacitor (TSC) type fast dynamic reactive power compensation device has been developed. The device uses a single-chip microcomputer as the main controller, realizing fuzzy control and various control methods, including symmetrical compensation and phase-by-phase compensation, multi-stage reactive power switching in one step, and voltage reference control. With reasonable configuration of compensation capacity, the overall load power factor cosφ can be guaranteed to be no less than 0.9, and the response time can reach as fast as 0.02 to 0.04. Furthermore, the switching is smooth with no current surges or reactive power backfeed. Compared to capacitor compensation methods with contact switches, it can solve the problems of overvoltage oscillation and arc reignition that often occur during switching, avoiding arc burns on contacts and switch damage caused by frequent operation. With the promotion of power electronics technology and the emergence of a large number of rectifier loads, harmonic pollution in the power grid is becoming increasingly serious. Harmonics flowing into reactive power compensation capacitors can cause them to overload or even be damaged. Therefore, the designed series reactor can limit the harmonic current flowing through the capacitor. By selecting the main wiring method and using a two-phase contactless switch control method, the cost can be effectively reduced and any internal wiring method of the three-phase capacitor bank can be selected. 2. Selection of Main Wiring Thyristor-controlled power electronic switch (TSC) reactive power compensation devices have advantages such as fast response speed, frequent switch operation, long lifespan, and 7-level stepped adjustment of reactive power capacity grouped in a 4:2:1 ratio. As a type of static compensator, it is superior to fixed capacitors or capacitor banks controlled by AC contactors. Common wiring methods are as follows (see Figure 1). (1) Three-phase control delta connection method (Figure 1a) The characteristic of this connection is that the line voltage is a known quantity (terminal voltage is fixed). The thyristor contactless switch is composed of two ordinary thyristors connected in anti-parallel. The thyristor always disconnects when the current crosses zero, so the voltage on the disconnected capacitor is generally at its maximum value (either positive or negative). Due to the self-discharge phenomenon of the discharge resistor installed inside the capacitor, the voltage on the capacitor gradually decreases. When the thyristor is reconnected, the residual voltage of the capacitor needs to be considered. When the system voltage and the residual voltage of the capacitor are equal (a small range of difference is allowed), it is the trigger point for the thyristor contactless switch to be connected. Otherwise, since the voltage across the capacitor cannot change abruptly, when the difference between the system voltage and the residual voltage of the capacitor is large, the triggering of the SCR will generate a large current surge (this is unavoidable for capacitors with contact switching). This surge will directly damage the thyristor. The current surge is mainly reflected in the current surge rate and the maximum value of the surge current when the switch is connected. (L0+L)di/dt=UL(t)-UC(t) (1) In the formula, L0 is the internal inductance of the power supply; L is the series inductance; UL(t) is the grid voltage when the thyristor is triggered; UC(t) is the capacitor voltage when the thyristor is triggered. Usually, the internal inductance L0 of the power supply is very small, and the series inductance L used to limit harmonic current is also not very large. Therefore, when the voltage difference is large, the current change rate di/dt when the SCR is turned on is very large. At the moment when the thyristor is just triggered, since the switching area near the gate of the thyristor does not have time to spread the conductive area to the entire crystal chip, the excessive di/dt at this time can easily damage the thyristor switch. The most serious situation is when the thyristor is triggered at the moment when UL(t) and UC(t) are out of phase and reach their maximum values. The resulting surge current is very dangerous. Taking 400 V as an example, the maximum voltage difference UL(t)-UC(t) can reach 1130 V. If a set of 50 kvar capacitors in delta connection is installed in a 400 V grid, the maximum voltage difference UL(t)-UC(t) can reach 1130 V. For the low-voltage side of a kVA transformer, assuming Uk = 4%, then C = Qc/3ωU2×106 = 332μF (2) L0 = U2×Uk(%)/ω×S×10 = 51μH (3) Where C is the compensation capacitor; U is the line voltage; QC is the reactive capacity of the compensation capacitor; S is the transformer capacity; Uk(%) is the percentage of the transformer short-circuit voltage. Considering the series inductance L = 50μH and ignoring the resistance, the impact current period T is where Im is the peak value of the impact current. Equation (4) shows that, ignoring the resistance, when the thyristor switch is triggered at the maximum voltage difference, the current reaches the maximum value of 2048A after about 0.287 ms. The thyristor current selected according to the normal load current should be 55 A. Thyristors can withstand 8-10 times the overcurrent for a short time, but they are difficult to withstand such a large overcurrent. Increasing the series reactor can reduce the current surge, or a larger thyristor can be selected, but both require increased investment. Therefore, the problem must be solved through control. To determine the appropriate triggering time, the residual voltage of the capacitor needs to be measured in advance, which is usually not easy to do. To solve this problem, a pulse sequence is selected as the trigger signal for the thyristor. In addition, no matter how high the residual voltage of the capacitor is, it is always less than or equal to the amplitude of the power supply. In one cycle, the thyristor will always be at zero voltage or reverse voltage at certain times. Each time the thyristor is triggered, the moment when it withstands the reverse voltage is selected as the start of the trigger pulse sequence. In this way, when the thyristor changes from reverse bias to forward bias, it automatically enters a stable conduction state, thus solving the problem of measuring the residual voltage of the capacitor. (2) A semi-controlled switch wiring method using diodes instead of some thyristors (Figure 1b). Its characteristic is that the capacitor always maintains a certain voltage (the positive maximum value) each time it is disconnected (thyristor blocking). In this way, when the thyristor switch is turned on, as long as the pulse train is triggered from the maximum value of the system voltage, a smooth transition can be guaranteed. Moreover, fewer thyristors are used to reduce costs, and the control is also simpler. Its disadvantage is that a current surge will still occur when the power is first supplied. In addition, since delta connection is not recommended for capacitor banks, Figures 1(a) and (b) require custom capacitors, which increases the complexity of the wiring. (3) A Y0 type wiring method is used (Figure 1c). With this wiring, the thyristor voltage rating can be reduced, but the current rating increases. The reduction in capacitor voltage will increase its unit price, and a short-term unbalanced neutral current will be generated when it is turned on. (4) Wiring method without B-phase thyristor switch, controlling only phases A and C (Figure 1d). Due to the uncertainty of the residual voltage of the capacitor, the maximum voltage that the thyristor can withstand is the same as that in Figure 1(a). This method can use the most common three-phase capacitor bank and uses fewer expensive thyristors, so it has advantages. This paper selects the wiring method of Figure 1(d) and develops a zero-voltage trigger circuit, making the control of this wiring method an economical and feasible solution. 3 Thyristor Voltage Zero-Crossing Trigger Circuit Due to the change in the residual voltage of the capacitor, the voltage on the thyristor is a value that cannot be calculated based on the power supply voltage. Therefore, this paper designs a thyristor voltage zero-crossing trigger circuit. The circuit diagram is shown in Figure 2. The voltage across the thyristor contactless switch is stepped down by a resistor and sent to the optocoupler. When the instantaneous AC voltage equals the residual voltage of the capacitor, the voltage across the thyristor is zero. At this time, the optocoupler outputs a negative pulse with a width of approximately 150μs. After the pulse is inverted, it is ANDed with the TSC activation command to start a multivibrator outputting a pulse train. This pulse train then passes through a power amplifier and isolation circuit to trigger the corresponding thyristor. Once triggered, the thyristor remains on, and the corresponding capacitor is activated. Because the voltage at the thyristor's on-state is close to zero, the trigger pulse train is continuously output as long as the TSC activation command exists, ensuring reliable thyristor conduction. When the TSC activation command is canceled, the trigger pulse stops, and the thyristor disconnects when the current crosses zero, only re-activating at the zero-voltage point when the microcontroller issues the next TSC activation command. This hardware interlocking circuit is relatively reliable and can avoid the accidental trigger pulses caused by software interlocking methods under conditions of large power supply interference (such as a remote short circuit fault in the AC system). For three-phase symmetrical compensation devices, only two phases need to be equipped with zero-voltage triggered contactless switches; however, for phase-by-phase reactive power compensation devices, all three phases need to be equipped with such contactless switches. For flexible adjustment, generally three groups of capacitors of different capacities must be installed to achieve seven levels of reactive power capacity adjustment. For example, a 100 kvar reactive power compensation device might have capacitor groups of 60, 30, and 15 kvar (for a 105 kvar reactive power compensation device). 4. Fuzzy Control of Reactive Power Compensation Low-voltage reactive power compensation equipment is generally divided into network-type and user-type. The former mainly reduces network losses and improves voltage quality; the latter mainly improves the power factor, reduces reactive power losses, and also improves voltage quality. Since the grid voltage level depends on the reactive power balance of the entire system, not just on a single reactive power device, the relationship between the voltage at a certain operating point and the reactive power compensation capacity cannot be represented by a single function curve. When the system operating conditions are used as parameters, the relationship between the voltage at the terminals of parallel capacitor banks and the reactive power is a set of curves, which brings difficulties to the automatic control of reactive power compensation capacitor banks. Under some operating conditions, the requirements for voltage quality and power factor may conflict. When the voltage value is near the rated value, the power factor may be relatively low; when the power factor is increased, the voltage may be too high, which is equivalent to a situation of excess reactive power in the system. For user-type reactive power compensation equipment, the reactive power compensation power and the load power factor also have a non-linear relationship. Under the same load conditions, the closer the power factor is to 1, the larger the compensation capacity required to adjust the same power factor, and the worse the compensation effect of using capacitors of the same capacity. Therefore, for the selection and operation control of user-type reactive power compensation equipment, even if the power factor is compensated to 1, it may not be economically worthwhile. At low loads, even with a very low power factor, using a small set of capacitors may result in overcompensation, potentially causing oscillations in the capacitor compensation equipment due to repeated switching. Frequent switching reduces the lifespan of both switches and capacitors. Even though using contactless switches can significantly extend switch life, frequent switching still causes interference and impacts on the system, a prominent problem with current capacitor compensation devices. Furthermore, the power output of reactive power compensation capacitors is proportional to the square of the voltage; the rated power output of each capacitor set varies with different voltage levels. All these factors complicate the control algorithm of reactive power compensation equipment. Therefore, this paper proposes a fuzzy control method. Since the TSC reactive power compensation device uses a tiered capacitor switching method to regulate reactive power, it cannot be continuously adjusted, and the adjustment amount does not need to be too finely divided, which perfectly suits the characteristics of fuzzy control. Considering the influence of several nonlinear factors and multi-valued functions, the fuzzy logic reasoning method can be used to calculate the appropriate capacity increment for compensation simply and quickly. Fuzzy control is a control method that uses a computer to mimic expert control strategies. The design of a fuzzy controller consists of three processes: input fuzzification, derivation of fuzzy control laws, and fuzzy output determination. The inputs of the fuzzy controller for the parallel capacitor compensation device are, in order, the reactive power of the load, the node voltage, and the power factor. According to the expert control strategy, the range of input values ​​is divided into several regions, and the membership degree of each region is determined to obtain satisfactory control results. For example, the system load reactive power is divided into 8 levels proportionally, the system voltage reference voltage qualification rate is divided into high, medium, and low levels, and the system load power factor is also divided into high, medium, and low levels. The input capacity of the parallel capacitor compensation device is a three-dimensional function of the load reactive power, node voltage and power factor. Its fuzzy controller has 3 inputs and 1 output. Generally, a set of multi-dimensional fuzzy conditional statements can be used to describe its control strategy. That is, IF {X1=A1i;X2=A2j;…Xn=Ank} i=1,…,mi,j=1,…,mj,…,k=1,…,mk THEN Y=Cl l=1,…,ml (5) where X is the input variable; A is the value range of each input variable; Y is the output variable; C is the value range of the output variable. Online calculation of the fuzzy control strategy according to equation (5) requires a lot of computation time and is not convenient for programming. In order to improve the response speed of the controller and simplify programming, based on the characteristics of discontinuous control of the capacitor compensation control device, the fuzzy regions of each input variable are classified. Then, based on the analysis of the capacitor compensation control law, the three-dimensional fuzzy control condition statement, equation (5), is transformed into a one-dimensional fuzzy condition statement in the following form: Please log in to: Power Transmission and Distribution Equipment Network to browse more information IF {X1=A1i} THEN IF {X2=A2j} THEN IF {X3=A3k} THEN Y=C1 Since the control result of capacitor grouping only has two states, ON and OFF, the fuzzy algorithm represented by equation (6) can be easily implemented using computer language. To avoid oscillation caused by frequent switching actions, the hysteresis characteristic of reactive power classification is added when performing fuzzy classification. The compensation device first classifies according to the reactive power of the load, and then considers the power factor and system voltage for compensation and correction. The reactive power compensation device designed in this paper has been tested and run for 18 months at Baoding Fengfan Motorcycle Battery Co., Ltd. References: [1] He Zhongxiong. Fuzzy Mathematics and Its Applications [M]. Tianjin: Tianjin Science and Technology Press, 1983: 295-312. [2] Yang Rengang. Microcomputer-based Voltage and Reactive Power Integrated Controller [C]. Proc, CUS-EPSA, 1992. [3] Ma Ruijun. Parallel Capacitor Switching Method Using Fuzzy Control Theory [J]. Journal of North China Electric Power University, 1998, (7). Author Introduction: Shi Xinchun (1950-), male, professor, engaged in research and teaching in power electronics and power system automation; Yang Meiling (1950-), female, senior engineer, engaged in research and management in power engineering and its automation; Yu Dezhong (1976-), male, master's student; Peng Wei (1970-), male, teaching assistant, master of engineering.