Three-phase balanced power supply energy-saving control method and system

Abstract

The application discloses a kind of three-phase balanced power supply energy-saving control method and system, method includes: building three-phase four-wire soft switching topological model, and real-time operating parameter of power supply system is collected and transmitted to complex domain convex optimization decision platform;Parameter characteristics are extracted in the platform, determine key parameters and establish its mapping relationship with three-phase unbalance degree, energy loss;Mapping relationship is input into multi-objective symmetric semi-positive programming algorithm, with three-phase unbalance degree and energy loss minimization as target to construct constraint condition, and the optimal conduction timing of controllable switching element, duty ratio and filter component optimal parameter are solved;Optimal parameter is fed back to topological model, and the working state of component is adjusted to realize control.System contains six units, and each unit works cooperatively.The application can reduce energy consumption, accurately control three-phase balance, adapt to complex working conditions, improve the operating efficiency and stability of power supply system, meet the dual needs of energy saving and balance of power system.

Description

Technical Field

[0001] This invention relates to the field of power supply energy-saving control technology, and in particular to a three-phase balanced power supply energy-saving control method and system. Background Technology

[0002] In the operation of power systems, three-phase power supply is widely used in various fields such as industrial production, commercial operation, and residential electricity consumption. Its power supply stability and energy-saving effect directly affect the overall operating efficiency and economic benefits of the power system. With the increase in the types of electrical equipment and the intensification of power load fluctuations, three-phase power supply systems are prone to three-phase voltage and current imbalances, increased neutral line current, and problems such as increased energy loss and decreased power factor. This not only affects the normal operating life of electrical equipment but also wastes electrical resources. To solve these problems, related fields need to construct technical solutions that combine three-phase balance regulation and energy-saving control functions. By optimizing the power supply system topology and introducing efficient algorithms and decision-making platforms, precise control of power supply parameters can be achieved to meet the dual requirements of power system stability and energy saving. Therefore, the development of a three-phase balanced power supply energy-saving control method and system has become an important direction in the current power technology field.

[0003] Existing technologies for energy-saving control of three-phase balanced power supply have two significant drawbacks. Firstly, the power supply system topologies used in existing technologies are mostly traditional hard-switching topologies. In these topologies, the switching process of controllable switching elements easily generates significant switching losses, and it is difficult to effectively control the neutral current, resulting in high energy consumption during system operation. Furthermore, the three-phase imbalance is difficult to control within an ideal range, failing to simultaneously achieve balance regulation and energy-saving effects. Secondly, the algorithms used for parameter optimization in existing technologies are mostly single-objective optimization algorithms, capable of optimizing only a single indicator of three-phase imbalance or energy loss. They lack multi-objective collaborative optimization capabilities and do not incorporate a complex domain convex optimization decision platform for precise parameter analysis and selection. This results in insufficient practicality and reliability of the optimization results, making it difficult to adapt to the complex and ever-changing operating conditions of power supply systems and failing to provide comprehensive and efficient control support for the power supply system. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of the existing technology, the present invention provides a three-phase balanced power supply energy-saving control method and system.

[0005] The technical solution adopted in this invention is a three-phase balanced power supply energy-saving control method, characterized by comprising: Step S1, constructing a three-phase four-wire soft-switching topology model, which includes three-phase main circuit branches and a neutral line branch, each branch being equipped with controllable switching elements and filtering components, and adjusting the conduction timing and duty cycle of the controllable switching elements to make the output voltage and current waveforms of the topology model meet the preset three-phase electrical characteristics; Step S2, collecting real-time operating parameters of the power supply system based on the three-phase four-wire soft-switching topology model, the real-time operating parameters including the voltage amplitude of each phase, the current amplitude of each phase, the neutral line current amplitude, power factor, and harmonic content, and transmitting the collected real-time operating parameters to a complex domain convex optimization decision platform; Step S3, extracting features from the real-time operating parameters in the complex domain convex optimization decision platform to determine the indicators affecting the three-phase balance and energy-saving effect. Step S4: Determine the parameters and establish a mapping relationship between the calibration parameters and the three-phase imbalance and energy loss; Step S5: Input the mapping relationship into a multi-objective symmetric semidefinite programming algorithm, with the optimization objectives of minimizing the three-phase imbalance and minimizing energy loss, and construct the constraints of the algorithm. The constraints include the maximum conduction current of the controllable switching element, the fluctuation range of the output voltage, and the limit threshold of the neutral line current; Step S6: Solve the optimal solution under the constraints using the multi-objective symmetric semidefinite programming algorithm to obtain the optimal conduction timing, optimal duty cycle of the controllable switching element, and optimal parameter configuration of the filter component; Step S7: Feed back the optimal conduction timing, optimal duty cycle, and optimal parameter configuration to the three-phase four-wire soft-switching topology model, and control the topology model to adjust the working state of the controllable switching element in each branch and the operating parameters of the filter component to perform three-phase balanced power supply and energy-saving control.

[0006] Furthermore, the voltage output expression of the three-phase four-wire soft-switching topology model is as follows: ,in It is a three-phase output voltage matrix. The topological coefficient matrix, This is the conduction state matrix of a controllable switching element. This is the DC-side input voltage. This is a matrix of inductance values ​​for each branch. This is a three-phase output current matrix. The matrix represents the resistance values ​​of each branch; the topology coefficient matrix... The conduction state matrix is ​​determined by the connection method between the three-phase main circuit branch and the neutral line branch. The value of the element is 0 or 1, which corresponds to the off and on states of the controllable switching element, respectively.

[0007] Furthermore, the objective function expression of the multi-objective symmetric semidefinite programming algorithm is: ,in To comprehensively optimize the target value, This is the weighting coefficient for the three-phase unbalance. This refers to the three-phase voltage imbalance. This is the weighting coefficient for energy loss. The active power loss of the power supply system; the three-phase voltage imbalance This is the maximum amplitude value among the three-phase voltages. This is the minimum amplitude of the three-phase voltage. The average amplitude of the three-phase voltage; active power loss. They are respectively Three-phase current amplitude, They are respectively Three-phase branch resistance value This is the amplitude of the neutral current. This represents the resistance value of the neutral branch.

[0008] Furthermore, the parameter optimization constraint expression of the complex domain convex optimization decision platform is as follows: ,in The optimized parameter matrix output by the platform. The initial value matrix for the parameters, It is a 2-norm. The error threshold is optimized for the parameters; the optimized parameter matrix This includes the duty cycle parameters of the controllable switching elements, the capacitance and inductance values ​​of the filter components, and the initial parameter matrix. The error threshold is determined by the rated operating parameters of the three-phase four-wire soft-switching topology model. The settings are based on the accuracy requirements of the power supply system.

[0009] Furthermore, the neutral current control expression for the three-phase four-wire soft-switching topology model is as follows: ,in This is the real-time current of the neutral line. This represents the number of parallel branches in the topology model. The first The real-time currents of phases A, B, and C in a parallel branch are monitored; the conduction state of the controllable switching elements in each parallel branch is adjusted using a multi-objective symmetrical positive definite programming algorithm to ensure the real-time current of the neutral line is controlled. satisfy This represents the maximum allowable current for the neutral line.

[0010] Furthermore, the optimal solution expression of the multi-objective symmetric semidefinite programming algorithm is as follows: ,in This is the optimal parameter vector output by the algorithm. For the feasible region of the parameters, For matrix trace operations, Let be the coefficient matrix of the quadratic term of the objective function. The parameter is the coefficient vector of the first-order term of the objective function; the feasible region of the parameter is... The quadratic term coefficient matrix is ​​determined by the conduction timing constraints of the controllable switching element, the output voltage amplitude range, and the current amplitude constraints. With the coefficient vector of the first term The weighting relationship between three-phase balance and energy-saving targets is set.

[0011] Further, step S3 includes the following sub-steps: S31. The collected real-time operating parameters are converted from time domain to frequency domain. The time domain signals of each phase voltage and current are converted into frequency domain signals through Fourier transform, and the fundamental component and each harmonic component in the frequency domain signal are extracted; S32. The fundamental component and harmonic components are feature-filtered, retaining the components with amplitudes exceeding a preset threshold and removing interference components with amplitudes below the threshold. The preset threshold is determined according to the harmonic standard of the power supply system; S33. A correlation function is established between the filtered components and the three-phase unbalance and energy loss. The correlation coefficient between each component and the three-phase unbalance and energy loss is calculated through correlation analysis, and the components with an absolute value of correlation coefficient greater than 0.8 are selected as calibration parameters; S34. The calibration parameters are classified by type to form a voltage calibration parameter set, a current calibration parameter set, and a power calibration parameter set, which are stored in the parameter database of the complex domain convex optimization decision platform.

[0012] Further, step S4 includes the following sub-steps: S41 transforms the mapping relationship obtained in step S3 into a mathematical expression for a multi-objective symmetric semidefinite programming algorithm, clarifying that the optimization variables of the algorithm are the conduction timing, duty cycle, and parameters of the filter component of the controllable switching element; S42 sets the inequality constraints of the algorithm, including the maximum conduction current constraint of the controllable switching element, the maximum fluctuation range constraint of the output voltage, and the upper limit constraint of the neutral line current, transforming each constraint into a mathematical inequality; S43 sets the equality constraints of the algorithm, including the constraint of equal three-phase voltage amplitude and the constraint of achieving the three-phase power factor standard, transforming each constraint into a mathematical equation; S44 integrates the inequality constraints and the equality constraints to form a complete constraint system for the multi-objective symmetric semidefinite programming algorithm, ensuring that the optimization variables take values ​​within the constraint system.

[0013] Further, step S5 includes the following sub-steps: S51 Initialize the iteration parameters of the multi-objective symmetric semidefinite programming algorithm, including the number of iterations, the iteration step size, and the convergence threshold. Set the initial value of the number of iterations to 100, the initial value of the iteration step size to 0.01, and the initial value of the convergence threshold to 0.001; S52 Input the constraint system constructed in step S4 into the algorithm, perform the first iteration calculation according to the set iteration parameters, and obtain the preliminary optimized parameter values; S53 Calculate the three-phase imbalance and energy loss corresponding to the preliminary optimized parameter values, compare the difference with the result of the previous iteration, and if the difference is greater than the convergence threshold, adjust the iteration step size and continue iterating; S54 Repeat the iteration process until the difference in the three-phase imbalance and the difference in energy loss between two adjacent iterations are both less than the convergence threshold, and output the optimized parameters at this time as the optimal solution.

[0014] A three-phase balanced power supply energy-saving control system includes: a three-phase four-wire soft-switching topology construction unit, which is connected to the main circuit of the power supply system and is used to build a topology structure including the three-phase main circuit branch and the neutral line branch, and output the real-time operating parameters of the topology model; a parameter acquisition and transmission unit, whose input terminal is connected to the output terminal of the three-phase four-wire soft-switching topology construction unit, for acquiring real-time operating parameters and transmitting them to a complex domain convex optimization decision unit; a complex domain convex optimization decision unit, whose input terminal is connected to the output terminal of the parameter acquisition and transmission unit, for extracting features from the real-time operating parameters and establishing a mapping relationship between calibration parameters and three-phase imbalance and energy loss; and a multi-objective symmetric semidefinite programming calculation unit. The input of this unit is connected to the output of the complex domain convex optimization decision unit, and is used to solve for the optimal parameters under constraints with the objectives of minimizing three-phase imbalance and energy loss. The optimal parameter feedback control unit is connected to the output of the multi-objective symmetric semidefinite programming calculation unit, and its output is connected to the control end of the three-phase four-wire soft-switching topology construction unit, and is used to feed back the optimal parameters to the topology and control its adjustment. The system operation monitoring unit is connected to the outputs of the three-phase four-wire soft-switching topology construction unit and the multi-objective symmetric semidefinite programming calculation unit, respectively, and is used to monitor the three-phase balance state and energy consumption state of the system, and transmit the monitoring results to the complex domain convex optimization decision unit.

[0015] Beneficial Effects: This invention proposes a three-phase balanced power supply energy-saving control method and system. By constructing a three-phase four-wire soft-switching topology model to replace the traditional hard-switching topology, it optimizes the conduction timing and duty cycle of controllable switching elements, significantly reducing switching losses. Simultaneously, combined with parameter control of the neutral branch, it effectively controls the neutral current, solving the problems of high energy consumption and difficulty in controlling three-phase imbalance, achieving a balance between balance regulation and energy saving. A multi-objective symmetric semi-positive definite programming algorithm is introduced, with minimizing three-phase imbalance and minimizing energy loss as dual optimization objectives, replacing the single-objective optimization algorithm and possessing multi-objective collaborative optimization capabilities. The system possesses the capability to extract features and perform precise analysis and screening of real-time operating parameters using a complex domain convex optimization decision-making platform. It establishes a mapping relationship between key parameters and balance and energy-saving indicators, improves constraints, and solves for optimal parameters. This addresses the issues of insufficient practicality and reliability of optimization results and their inability to adapt to complex operating conditions. Through a complete process including parameter acquisition and transmission, and optimal solution feedback control, it achieves dynamic and precise regulation of power supply parameters, meeting the dual requirements of power system for operational stability and energy saving. It provides comprehensive and efficient control support for three-phase power supply scenarios in different fields, further improving the overall operating efficiency and economic benefits of the power system. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method steps of the present invention;

[0017] Figure 2 This is a diagram showing the system unit composition of the present invention. Detailed Implementation

[0018] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] like Figure 1 As shown, a three-phase balanced power supply energy-saving control method includes:

[0020] Step S1: Build a three-phase four-wire soft-switching topology model. The model includes a three-phase main circuit branch and a neutral line branch. Each branch is equipped with controllable switching elements and filter components. By adjusting the conduction timing and duty cycle of the controllable switching elements, the output voltage and current waveforms of the topology model meet the preset three-phase electrical characteristics.

[0021] Specifically, step S1 constructs the hardware foundation for three-phase balanced power supply energy-saving control. By building a three-phase four-wire soft-switching topology model, a stable circuit structure is provided for subsequent parameter acquisition, optimization and control. The rationality of this model directly affects the accuracy of three-phase balance regulation and energy-saving effect. In the specific implementation process, the topology model includes three-phase main circuit branches (A, B, and C) and one neutral branch. Each main circuit branch is equipped with IGBT-type controllable switching elements and LC-type filter components. The neutral branch is also equipped with controllable switching elements and filter components. The rated on-state current of the controllable switching elements is set to 50A, and the maximum off-state voltage is set to 1200V. The inductance value of the filter components is selected as 10mH, and the capacitance value is selected as 220μF. During implementation, the components of each branch are connected according to the circuit design drawings to ensure that the three-phase main circuit branches are symmetrically distributed. The neutral branch is connected to the common node of the three-phase main circuit. After the hardware assembly is completed, the initial on-time of the controllable switching elements is preset to a 10kHz PWM signal and the initial duty cycle is set to 50% through the debugging equipment. This ensures that the topology model can output a three-phase voltage waveform that meets the industrial power standards after being powered on, creating a stable circuit environment for subsequent acquisition of operating parameters.

[0022] Step S2: Collect real-time operating parameters of the power supply system based on the three-phase four-wire soft-switching topology model. The real-time operating parameters include the voltage amplitude of each phase, the current amplitude of each phase, the neutral line current amplitude, the power factor and the harmonic content. Transmit the collected real-time operating parameters to the complex domain convex optimization decision platform.

[0023] Specifically, step S2 acquires real-time operating status data of the power supply system, providing accurate analytical basis for the complex domain convex optimization decision-making platform. Only by accurately collecting key parameters can we ensure that subsequent optimization objectives match actual operating conditions. In the specific implementation process, based on the established three-phase four-wire soft-switching topology model, voltage and current sensors are installed at the output terminals of each phase's main circuit branch and neutral branch. The voltage sensor's measurement range is set to 0-500V, and the measurement accuracy is set to ±0.5%. The current sensor's measurement range is set to 0-100A, and the measurement accuracy is set to ±0.2%. Simultaneously, a power factor meter and a harmonic analyzer are installed at the output terminals of the topology model. The power factor meter's measurement range is set to 0.5-1.0, and the harmonic analyzer's measurement frequency covers the 1st to 30th harmonics. During implementation, the sensors are first... The measuring instruments were calibrated to ensure accurate measurement data. Then, the topology model was started. The sensors and instruments collected data in real time at 10ms intervals, including the voltage amplitude of each phase, the current amplitude of each phase, the neutral current amplitude, the power factor, and the harmonic content. The voltage amplitude of each phase was collected within the commonly used industrial power range of 380V±10%, and the current amplitude of each phase was collected to cover the load fluctuation range of 0-80A. The collected data was transmitted to the database of the complex domain convex optimization decision platform via RS485 communication protocol at a transmission rate of 9600bps to ensure the real-time performance and integrity of the data.

[0024] Step S3: In the complex domain convex optimization decision platform, feature extraction is performed on the real-time operating parameters to determine the calibration parameters that affect the three-phase balance and energy-saving effect, and a mapping relationship is established between the calibration parameters and the three-phase imbalance and energy loss.

[0025] Specifically, step S3 filters key information from massive real-time operating parameters, establishes the correlation between parameters and control objectives, avoids irrelevant parameters interfering with the optimization process, and improves the efficiency and accuracy of subsequent algorithm optimization. In the specific implementation process, the parameter processing module is activated in the complex domain convex optimization decision platform. First, the real-time operating parameters transmitted to the platform are cleaned to remove outliers caused by sensor interference. The outlier criterion is set as data exceeding the normal measurement range by ±15%. Then, feature extraction is performed. Using the platform's built-in feature analysis algorithm, the correlation between each parameter and the three-phase imbalance and energy loss is calculated. The correlation calculation uses the Pearson correlation coefficient method, setting parameters with an absolute correlation coefficient greater than 0.8 as key parameters. During implementation, historical operating data is first imported as reference samples. The threshold parameters for feature extraction are determined through sample training. Then, the real-time collected data for each phase... Substituting voltage amplitude, phase current amplitude, neutral current amplitude, power factor, and harmonic content into the analysis, the key parameters finally selected include phase A voltage amplitude, phase B voltage amplitude, phase C voltage amplitude, phase A current amplitude, phase B current amplitude, phase C current amplitude, neutral current amplitude, and third harmonic content. Then, a mapping relationship between the key parameters and three-phase imbalance and energy loss is established. Through the platform's mathematical modeling module, the key parameters are used as input variables, and three-phase imbalance and energy loss are used as output variables to construct a multivariate linear mapping model. The model's fit is required to reach above 0.95 to ensure that the mapping relationship accurately reflects the impact of parameter changes on the control objective.

[0026] Step S4: Input the mapping relationship into the multi-objective symmetric semidefinite programming algorithm, with the optimization objectives of minimizing the three-phase imbalance and minimizing energy loss, and construct the constraints of the algorithm. The constraints include the maximum conduction current of the controllable switching element, the fluctuation range of the output voltage, and the limit threshold of the neutral line current.

[0027] Specifically, step S4 clarifies the algorithm optimization direction and boundary conditions. By constructing a multi-objective optimization framework, the algorithm can simultaneously consider three-phase balance and energy-saving requirements, avoiding control deviations caused by single-objective optimization. In the specific implementation process, the mapping relationship established in step S3 is imported into the input interface of the multi-objective symmetric semi-definite programming algorithm. First, the optimization objectives of the algorithm are set as minimizing three-phase imbalance and minimizing energy loss, with the target value for three-phase imbalance set at ≤2% and the target value for energy loss set at ≤5%. Then, the constraints of the algorithm are constructed, including operational constraints of controllable switching elements, stability constraints of output voltage, and safety constraints of neutral line current. During implementation, the constraints of the controllable switching elements are set as follows: maximum conduction current ≤50A and conduction frequency ≤20kHz. Hz, turn-off time ≥1μs; the output voltage constraint is set as follows: the voltage amplitude fluctuation range of each phase ≤380V±5%, and the voltage amplitude difference of the three phases ≤5V; the neutral line current constraint is set as the maximum current ≤10A; the constraint conditions are converted into mathematical restriction rules through the constraint configuration module of the algorithm, and the priority of the constraint conditions is set. Among them, the output voltage stability constraint and the neutral line current safety constraint have higher priority than the operation constraint of the controllable switching element, ensuring that the power supply safety and stability are satisfied first during the optimization process, and defining a reasonable parameter range for the subsequent algorithm to solve the optimal solution.

[0028] Step S5: Solve the optimal solution under the constraints using a multi-objective symmetric semidefinite programming algorithm to obtain the optimal on-time of the controllable switching element, the optimal duty cycle, and the optimal parameter configuration of the filter component.

[0029] Specifically, step S5 calculates the optimal control parameters through algorithmic calculation, providing a concrete basis for adjusting the topology model. This is the core link connecting parameter analysis and actual control; the accuracy of the algorithm's solution directly determines the final realization of three-phase balance and energy-saving effects. In the specific implementation process, the solution module of the multi-objective symmetric semi-positive definite programming algorithm is started. First, the algorithm's solution parameters are initialized, setting the iteration count to 200, the iteration step size to 0.001, and the convergence threshold to 0.0001. Then, based on the optimization objectives and constraints set in step S4, the algorithm searches the parameter space using the interior-point method. In each iteration, the algorithm calculates the three-phase imbalance and energy loss corresponding to the current parameter combination and compares it with the result of the previous iteration. During implementation, if the difference between two iteration results is greater than the convergence threshold, the algorithm will adjust according to the gradient descent principle. The parameter search direction gradually reduces the parameter error. When the number of iterations reaches 200 or the difference between two adjacent iterations is less than the convergence threshold, the algorithm stops iterating and outputs the optimal turn-on timing, optimal duty cycle, and optimal parameters of the filter component for the controllable switching element. The optimal turn-on timing is set to a 15kHz PWM signal, the optimal duty cycle is adjusted according to the load differences of each phase to 48% for phase A, 52% for phase B, and 50% for phase C, and the optimal parameters of the filter component are adjusted to an inductance value of 12mH and a capacitance value of 200μF to ensure that these parameters meet the dual requirements of the optimization objective and the constraints.

[0030] Step S6: Feed back the optimal conduction timing, optimal duty cycle and optimal parameter configuration to the three-phase four-wire soft switch topology model, and control the topology model to adjust the working state of the controllable switching elements of each branch and the operating parameters of the filter components to perform three-phase balanced power supply and energy-saving control.

[0031] Specifically, step S6 applies the optimal parameters obtained from the algorithm to actual control. By adjusting the operating state of the topology model, the goal of three-phase balanced power supply and energy-saving control is ultimately achieved, which is the final execution stage of the entire control process. In the specific implementation process, the optimal conduction timing, optimal duty cycle, and optimal parameters of the filter components obtained in step S5 are fed back to the control unit of the three-phase four-wire soft-switching topology model via the data transmission module. The control unit uses a PLC controller, which has the functions of parameter reception, parsing, and execution. During implementation, the PLC controller first parses the received optimal parameters, converting the optimal conduction timing into drive signals for the controllable switching elements. The voltage amplitude of the drive signal is set to 15V, and the current amplitude is set to 2A. Subsequently, the conduction time of the controllable switching elements in each phase's main circuit branch is adjusted according to the optimal duty cycle. The A-phase element conducts... The on-time is adjusted to 24μs for phase B, 26μs for phase C, and 25μs for phase C. Simultaneously, by controlling the adjustment knobs of the filter components, the inductance value is adjusted from the initial 10mH to 12mH, and the capacitance value from the initial 220μF to 200μF. During the adjustment process, the control unit monitors the output parameters of the topology model in real time. If the parameters deviate from the optimal values, a secondary fine-tuning is immediately performed to ensure the topology model operates stably at the optimal parameter state. Ultimately, the three-phase voltage imbalance is controlled within 1.5%, and energy loss is controlled within 4%, achieving the dual effects of balanced three-phase power supply and energy saving.

[0032] Preferably, the voltage output expression of the three-phase four-wire soft-switching topology model is: ,in It is a three-phase output voltage matrix. The topological coefficient matrix, This is the conduction state matrix of a controllable switching element. This is the DC-side input voltage. This is a matrix of inductance values ​​for each branch. This is a three-phase output current matrix. The matrix represents the resistance values ​​of each branch; the topology coefficient matrix... The conduction state matrix is ​​determined by the connection method between the three-phase main circuit branch and the neutral line branch. The value of the element is 0 or 1, which corresponds to the off and on states of the controllable switching element, respectively.

[0033] Specifically, this paper clarifies the voltage output law and key parameter correlation of a three-phase four-wire soft-switching topology model, providing a quantitative basis for voltage regulation of the topology model. This ensures that the output voltage waveform meets the requirements of three-phase balance and energy-saving control. Furthermore, defining the meaning of parameters makes model parameter adjustments more targeted. In the specific implementation process, the voltage output-related parameters involved in this paper need to be determined in conjunction with the hardware configuration of the topology model. The topology coefficient matrix is ​​determined by the connection method of the three-phase main circuit branches and the neutral line branch. When the three-phase main circuit adopts a star connection and the neutral line is directly connected to the common node, the coefficient matrix needs to reflect the symmetrical relationship of each phase branch. In the conduction state matrix of the controllable switching elements, the element values ​​correspond to the element's off and on states. During implementation, this matrix needs to be adjusted according to real-time load changes. For example, when the load on phase A increases, the conduction state element corresponding to phase A needs to have its value of 1 for an extended period. The DC side input voltage is set to 500V. The system matches the DC bus voltage commonly used in industrial power supplies. In the inductance matrix of each branch, the inductance value of the three-phase main circuit branch is uniformly set to 10mH, and the inductance value of the neutral line branch is set to 8mH. In the resistance matrix of each branch, the resistance value of the main circuit branch is set to 0.5Ω, and the resistance value of the neutral line branch is set to 0.3Ω. By configuring these parameters and combining them with the voltage output law, the three-phase output voltage under different operating conditions can be calculated, thereby determining whether it meets the preset industrial power standard of 380V±5%. If the voltage exceeds the range, it is corrected by adjusting the conduction state and duty cycle of the controllable switching element to ensure that the topology model outputs a stable voltage.

[0034] Preferably, the objective function expression of the multi-objective symmetric semidefinite programming algorithm is: ,in To comprehensively optimize the target value, This is the weighting coefficient for the three-phase unbalance. This refers to the three-phase voltage imbalance. This is the weighting coefficient for energy loss. The active power loss of the power supply system; the three-phase voltage imbalance This is the maximum amplitude value among the three-phase voltages. This is the minimum amplitude of the three-phase voltage. The average amplitude of the three-phase voltage; active power loss. They are respectively Three-phase current amplitude, They are respectively Three-phase branch resistance value This is the amplitude of the neutral current. This represents the resistance value of the neutral branch.

[0035] Specifically, a quantitative model of the optimization objective of a multi-objective symmetric semidefinite programming algorithm is established to clarify the relationship between the comprehensive optimization objective and three-phase imbalance and energy loss. By defining the meaning and calculation method of parameters, the optimization direction of the algorithm is made clearer, and a reference is provided for setting subsequent constraints. In the specific implementation process, the calculation of the comprehensive optimization objective value requires first determining the weighting coefficients. Based on the priority requirements of the power supply system for balance and energy saving, the weighting coefficient of three-phase imbalance is set to 0.6, and the weighting coefficient of energy loss is set to 0.4 to ensure priority control of three-phase imbalance. When calculating the three-phase voltage imbalance, the voltage amplitude of each phase needs to be collected first. For example, in actual operation, the voltage of phase A is 382V, phase B is 378V, and phase C is 380V. At this time, the maximum amplitude is 382V, the minimum is 378V, and the average is 380V. Substituting these values ​​into the calculation, the imbalance is found to be 1.05%, which needs to be controlled within the target value of ≤2%. When calculating the active power loss, the current of each phase is... The amplitude needs to be collected by a current sensor. For example, if the current of phase A is 45A, phase B is 42A, phase C is 43A, the neutral current is 8A, the resistance of each phase branch is set to 0.5Ω, and the neutral resistance is 0.3Ω, the power loss can be calculated as 45²×0.5+42²×0.5+43²×0.5+8²×0.3=1012.5+882+924.5+19.2=2838.2W, which needs to be controlled within the target range. During implementation, the algorithm will calculate the comprehensive optimization target value in real time. If the target value exceeds the preset threshold, the optimization parameters will be adjusted until the imbalance and power loss meet the requirements, thus achieving multi-objective collaborative optimization.

[0036] Preferably, the parameter optimization constraint expression of the complex domain convex optimization decision platform is: ,in The optimized parameter matrix output by the platform. The initial value matrix for the parameters, It is a 2-norm. The error threshold is optimized for the parameters; the optimized parameter matrix This includes the duty cycle parameters of the controllable switching elements, the capacitance and inductance values ​​of the filter components, and the initial parameter matrix. The error threshold is determined by the rated operating parameters of the three-phase four-wire soft-switching topology model. The settings are based on the accuracy requirements of the power supply system.

[0037] Specifically, a reasonable range for parameter optimization is defined for the complex domain convex optimization decision platform. By clarifying the parameter optimization constraints and their meanings, the optimization process avoids parameters exceeding hardware and system accuracy limits, ensuring the feasibility and reliability of the optimization results. In the specific implementation, the optimization parameter matrix includes the duty cycle of controllable switching elements and the capacitance and inductance values ​​of filter components. The duty cycle parameter range is set to 40%-60%, the filter component capacitance range is 180μF-250μF, and the inductance range is 8mH-15mH. The initial parameter value matrix is ​​determined by the rated operating parameters of the topology model. Under rated operating conditions, the initial duty cycle is set to 50%, the capacitance to 220μF, and the inductance to 10mH. The parameter optimization error threshold is set to 0.02 based on the power supply system accuracy requirements, ensuring that the deviation between the optimized parameters and the initial values ​​is within acceptable limits. Within the specified range; during implementation, the platform first obtains the current optimization parameter matrix, such as a duty cycle of 48%, a capacitance value of 200μF, and an inductance value of 12mH at a certain moment, and compares it with the initial values ​​to calculate the L2 norm. If the L2 norm is less than 0.02, the parameter optimization is determined to meet the constraint requirements; if the L2 norm exceeds the threshold, for example, if the duty cycle is adjusted to 38%, and the deviation from the initial value of 50% is large, and the L2 norm exceeds 0.02, the platform will pull the parameter back to the range of 40%-60% and recalculate the optimization to ensure that all optimization parameters are within a reasonable range and to avoid system instability due to abnormal parameters.

[0038] Preferably, the neutral current control expression of the three-phase four-wire soft-switching topology model is: ,in This is the real-time current of the neutral line. This represents the number of parallel branches in the topology model. The first The real-time currents of phases A, B, and C in a parallel branch are monitored; the conduction state of the controllable switching elements in each parallel branch is adjusted using a multi-objective symmetrical positive definite programming algorithm to ensure the real-time current of the neutral line is controlled. satisfy This represents the maximum allowable current for the neutral line.

[0039] Specifically, the control law of neutral line current in a three-phase four-wire soft-switching topology model is clarified. By defining the meaning of parameters and control thresholds, the neutral line current regulation is made more targeted, avoiding line losses and safety hazards caused by excessive neutral line current. At the same time, it is coordinated with a multi-objective symmetric semi-definite programming algorithm to improve the overall control effect. In the specific implementation process, the calculation of the real-time neutral current requires first determining the number of parallel branches. Based on the power capacity requirements of the topology model, the number of parallel branches is set to 3. The real-time A, B, and C phase currents of each branch are collected by current sensors. For example, the first branch has phase A current of 15A, phase B current of 14A, and phase C current of 14.5A; the second branch has phase A current of 16A, phase B current of 15A, and phase C current of 15.2A; and the third branch has phase A current of 14A, phase B current of 13A, and phase C current of 13.3A. Substituting these values ​​into the calculation, the real-time neutral current can be obtained as (15+14+14.5)+(16+15+15.2)+(14+13+13.3)=43.5+46 0.2 + 40.3 = 129A, which needs to be compared with the maximum allowable current of the neutral line. The maximum allowable current of the neutral line is determined based on the cross-sectional area and material of the neutral line conductor. When using copper conductor with a cross-sectional area of ​​10mm², the maximum allowable current is set to 150A. During implementation, the multi-objective symmetric semi-definite programming algorithm will monitor the neutral line current in real time. If the current is close to 150A, for example, reaching 145A, the algorithm will adjust the conduction state of the controllable switching elements of each parallel branch. For example, it will reduce the conduction time of the third branch and reduce the current of that branch, so that the neutral line current drops to about 120A, ensuring that the current is always less than the maximum allowable value, avoiding overheating and damage to the neutral line, and reducing line losses.

[0040] Preferably, the optimal solution expression of the multi-objective symmetric semidefinite programming algorithm is: ,in This is the optimal parameter vector output by the algorithm. For the feasible region of the parameters, For matrix trace operations, Let be the coefficient matrix of the quadratic term of the objective function. The parameter is the coefficient vector of the first-order term of the objective function; the feasible region of the parameter is... The quadratic term coefficient matrix is ​​determined by the conduction timing constraints of the controllable switching element, the output voltage amplitude range, and the current amplitude constraints. With the coefficient vector of the first term The weighting relationship between three-phase balance and energy-saving targets is set.

[0041] Specifically, this paper provides a quantitative method for solving the optimal solution using a multi-objective symmetric semidefinite programming algorithm. It clarifies the rules for obtaining the optimal parameter vector and its relationship with the feasible region and coefficient matrix. By defining the meaning of the parameters, the algorithm's solution process is made more standardized, ensuring that the output optimal parameters meet the system control requirements. In the specific implementation, the optimal parameter vector includes the conduction timing and duty cycle of the controllable switching elements, and the capacitance and inductance values ​​of the filter components. These parameters must be taken within the feasible region. The feasible region is determined by multiple constraints: the conduction timing constraint of the controllable switching elements is set to 5kHz-20kHz, the duty cycle constraint is 40%-60%, the output voltage amplitude constraint is 380V±5%, and the current amplitude constraint is 0-50A. The coefficient matrix of the quadratic term of the objective function is set according to the degree of influence of the parameters on the optimization objective, with the coefficient corresponding to the duty cycle set to 0.3, the coefficient corresponding to the conduction timing set to 0.2, the coefficient corresponding to the capacitance value set to 0.25, and the coefficient corresponding to the inductance value set to 0.2. 5; The coefficient vector of the first-order term is set according to the initial values ​​of the parameters. The coefficient corresponding to the initial value of the duty cycle of 50% is -0.5, the coefficient corresponding to the initial value of the conduction timing of 10kHz is -0.2, the coefficient corresponding to the initial value of the capacitance of 220μF is -0.25, and the coefficient corresponding to the initial value of the inductance of 10mH is -0.25. During implementation, the algorithm calculates the objective function value through matrix trace operation and searches for the parameter vector that minimizes the objective function within the feasible region of the parameters. For example, the optimal parameters are finally obtained as follows: conduction timing of 15kHz, duty cycle of A phase 48%, B phase 52%, C phase 50%, capacitance of 200μF, and inductance of 12mH. Substituting and verifying, it can be seen that the parameters satisfy all constraints and can achieve three-phase balance and energy-saving control.

[0042] Preferably, step S3 includes the following sub-steps: S31. The collected real-time operating parameters are converted from time domain to frequency domain. The time domain signals of each phase voltage and current are converted into frequency domain signals through Fourier transform, and the fundamental component and each harmonic component in the frequency domain signal are extracted; S32. The fundamental component and harmonic components are feature-filtered, retaining the components with amplitudes exceeding a preset threshold and removing interference components with amplitudes below the threshold. The preset threshold is determined according to the harmonic standard of the power supply system; S33. A correlation function is established between the filtered components and the three-phase unbalance and energy loss. The correlation coefficient between each component and the three-phase unbalance and energy loss is calculated through correlation analysis, and the components with an absolute value of correlation coefficient greater than 0.8 are selected as calibration parameters; S34. The calibration parameters are classified by type to form a voltage calibration parameter set, a current calibration parameter set, and a power calibration parameter set, which are stored in the parameter database of the complex domain convex optimization decision platform.

[0043] Specifically, step S3's parameter processing flow ensures the accurate selection of key parameters from real-time operating parameters, providing high-quality input for subsequent algorithm optimization, avoiding interference from irrelevant parameters, and improving the accuracy of the correlation between control targets and parameters. In the specific implementation process, step S31 requires time-domain and frequency-domain conversion of the acquired real-time parameters. Fourier transform is used to convert the time-domain signals of each phase voltage and current into frequency-domain signals. During the transformation, the sampling frequency is set to 50kHz, and the number of sampling points is set to 1024 points to ensure complete extraction of the fundamental component and the 1st to 30th harmonic components. In step S32, during feature filtering, a preset threshold is set to 5% of the fundamental component amplitude. Interference components with amplitudes below this threshold are removed. For example, if the amplitude of the 5th harmonic component in a certain acquisition is only 3% of the fundamental, it is removed, retaining the fundamental and the 3rd, 7th, and other harmonic components with amplitudes exceeding the threshold. Step S33 establishes the correlation. When performing function calculations, correlation analysis is used to calculate the correlation coefficients between each component and the three-phase unbalance and energy loss. A correlation coefficient with an absolute value greater than 0.8 is set as the screening criterion. For example, if the correlation coefficient between the fundamental voltage component and the unbalance is 0.92 and the correlation coefficient between the third harmonic current component and the energy loss is 0.85, then both are included in the key parameters. When storing the S34 categories, the voltage category key parameter set includes the fundamental voltage, the third harmonic voltage, etc., the current category includes the fundamental current, the seventh harmonic current, etc., and the power category includes the power factor. These are stored in the corresponding partitions of the platform parameter database, and the database storage rate is set to 10MB / s to ensure fast data retrieval.

[0044] Preferably, step S4 includes the following sub-steps: S41 transforming the mapping relationship obtained in step S3 into a mathematical expression of a multi-objective symmetric semidefinite programming algorithm, clarifying that the optimization variables of the algorithm are the conduction timing, duty cycle, and parameters of the filter component of the controllable switching element; S42 setting the inequality constraints of the algorithm, including the maximum conduction current constraint of the controllable switching element, the maximum fluctuation range constraint of the output voltage, and the upper limit constraint of the neutral line current, and transforming each constraint into a mathematical inequality; S43 setting the equality constraints of the algorithm, including the constraint of equal three-phase voltage amplitude and the constraint of achieving the three-phase power factor standard, and transforming each constraint into a mathematical equation; S44 integrating the inequality constraints and the equality constraints to form a complete constraint system of the multi-objective symmetric semidefinite programming algorithm, ensuring that the optimization variables take values ​​within the constraint system.

[0045] Specifically, in step S4, the algorithm construction process clarifies the mathematical transformation and constraint system construction logic of the multi-objective symmetric semidefinite programming algorithm, ensuring that the optimization objective and constraint conditions match the actual working conditions, and providing a clear framework for the algorithm solution. In the specific implementation process, when transforming the mathematical expression in S41, key parameters in the mapping relationship are used as algorithm input variables, while the conduction timing, duty cycle, and filter component parameters of the controllable switching element are used as optimization variables. Linear interpolation is used during the transformation to ensure the continuity of the mathematical relationship between variables and parameters. In step S42, when setting inequality constraints, the maximum conduction current constraint of the controllable switching element is set to 50A, the maximum output voltage fluctuation range constraint is 380V±5%, and the upper limit constraint of the neutral line current is 10A. These constraints are transformed into mathematical inequalities, such as "the conduction current corresponding to the optimization variable..." "Current ≤ 50A"; When setting equality constraints in S43, the constraint of equal three-phase voltage amplitude is transformed into "Ua=Ub=Uc", and the constraint of achieving the power factor standard is transformed into "Power factor ≥ 0.95", ensuring that the equality constraints directly correspond to the control target; When integrating the constraint system in S44, the constraints are sorted by priority, with the output voltage and neutral line current constraints having higher priority than the switching element constraints. After integration, a constraint system containing 12 inequalities and 2 equalities is formed. The constraint configuration module of the algorithm transforms it into machine-recognizable rules. The rule verification pass rate must reach 100% to avoid constraint conflicts.

[0046] Preferably, step S5 includes the following sub-steps: S51 Initialize the iteration parameters of the multi-objective symmetric semidefinite programming algorithm, including the number of iterations, the iteration step size, and the convergence threshold. Set the initial value of the number of iterations to 100, the initial value of the iteration step size to 0.01, and the initial value of the convergence threshold to 0.001; S52 Input the constraint system constructed in step S4 into the algorithm, perform the first iteration calculation according to the set iteration parameters, and obtain the preliminary optimized parameter values; S53 Calculate the three-phase imbalance and energy loss corresponding to the preliminary optimized parameter values, compare the difference with the result of the previous iteration, and if the difference is greater than the convergence threshold, adjust the iteration step size and continue iterating; S54 Repeat the iteration process until the difference in the three-phase imbalance and the difference in energy loss between two adjacent iterations are both less than the convergence threshold, and output the optimized parameters at this time as the optimal solution.

[0047] Specifically, in step S5, the algorithm iterative solution process clarifies the iterative parameter settings, calculation logic, and convergence criteria to ensure that the algorithm can stably output the optimal solution that meets the requirements, thereby improving the solution accuracy and efficiency. In the specific implementation process, when initializing the iteration parameters in S51, the initial number of iterations is set to 100, the iteration step size to 0.01, and the convergence threshold to 0.001. These parameters are set based on historical solution data; for example, in the past 100 solutions, 95% convergence occurred within 100 iterations, hence this initial value is set. In step S52, during the first iteration calculation, the constraint system is input into the algorithm, and the interior point method is used for parameter search. In the first iteration, 10 candidate points near the initial parameter values ​​are selected, and the three-phase imbalance and energy loss corresponding to each candidate point are calculated. The candidate point with the best overall performance is selected as the result of the first iteration. In S53, when comparing the difference, if the imbalance is 2.5% in the first iteration and 2.3% in the second, with a difference of 0.2 > 0.001, the iteration step size is adjusted to 0.008, and the iteration continues. In S54, when iterating repeatedly, the candidate point range is updated after each iteration to narrow the search interval. When the imbalance is 1.5% in the 85th iteration and 1.5005% in the 86th iteration, with a difference of 0.0005 < 0.001, and the energy loss difference also meets the condition, the iteration stops, and the optimized parameters at this time are output. The total iteration time is controlled within 500ms to ensure real-time control requirements.

[0048] The three-phase four-wire soft-switching topology model in this invention is the hardware foundation supporting energy-saving control of three-phase balanced power supply. Essentially, it is a circuit structure comprising three-phase main circuit branches (A, B, and C) and one neutral branch. Each branch is equipped with controllable switching elements (such as IGBTs) and LC-type filter components. Unlike traditional hard-switching topologies, its core feature is reducing switching losses through soft-switching technology. In implementation, the three-phase main circuit branches are first assembled in a star connection to ensure symmetrical distribution. The neutral branch is connected to the common node of the three-phase main circuit. The rated on-time of the controllable switching elements is set to 50A, and the maximum off-time voltage to 1200V. The filter component has an inductance of 10-12mH and a capacitance of 200-220μF. After assembly, a 10kHz initial PWM turn-on sequence and a 50% initial duty cycle are preset, and hardware debugging is completed. Its function is to output three-phase voltage waveforms conforming to industrial standards, collect real-time operating parameters (such as the voltage and current amplitude of each phase), and adjust the operating state of the components based on subsequent optimization parameters. The significance of this model lies in solving the problems of high switching losses and difficult control of neutral line current in traditional topologies, providing a stable circuit environment for subsequent parameter optimization and precise control, and serving as a hardware prerequisite for achieving three-phase balance and energy saving.

[0049] The multi-objective symmetric semidefinite programming algorithm in this invention is the core computational tool for parameter optimization. Essentially, it's a multi-objective optimization algorithm that simultaneously minimizes three-phase imbalance and energy loss. By constructing a mathematical model using symmetric semidefinite programming theory, it differs from single-objective algorithms in that it can simultaneously meet dual control requirements. In implementation, the key parameter mapping relationships output by the complex domain convex optimization decision platform are first transformed into mathematical expressions. Optimization objectives are set as follows: three-phase imbalance ≤2% and energy loss ≤5%. A constraint system is constructed (e.g., maximum conduction current of controllable switching elements 50A, output voltage fluctuation ±5%, neutral line current ≤10A). The initial iteration count is 200, step size 0.001, and convergence threshold 0.0001. An interior-point method is used for iterative search within the parameter feasible region. Each iteration calculates the target value corresponding to the current parameter, comparing the difference between adjacent iterations until the convergence condition is met, outputting the optimal parameters (e.g., 15kHz conduction timing, 48%-52% duty cycle). Its function is to solve for the optimal parameters of the controllable switching elements and filter components, providing a basis for topology model adjustment. The significance of this algorithm lies in overcoming the limitations of traditional single-objective optimization, achieving balanced and energy-saving synergistic optimization, improving the practicality and reliability of parameter optimization, and serving as a core link connecting parameter analysis and actual control.

[0050] The complex domain convex optimization decision platform of this invention is a data analysis and decision-making hub for processing operating parameters and establishing parameter correlations. Essentially, it is a software system with data acquisition, feature extraction, parameter storage, and mapping modeling functions. Based on the complex domain convex optimization theory, it achieves precise parameter analysis. Unlike ordinary data processing platforms, it can screen key parameters and establish correlations with control objectives. In implementation, real-time data (such as voltage, current, and power factor) transmitted from the parameter acquisition unit is first received via RS485 protocol (9600bps transmission rate). Fourier transform is performed using a 50kHz sampling frequency and 1024 sampling points to convert the time-domain signal to a frequency-domain signal. A screening threshold of 5% of the fundamental amplitude is set to eliminate interference components. Key parameters are screened through correlation analysis (absolute correlation coefficient > 0.8). A multivariate linear mapping model (fit ≥ 0.95) is established between key parameters and three-phase imbalance and energy loss. The categorized key parameters (voltage, current, and power) are stored in the corresponding database partitions (storage rate 10MB / s). Its function is to provide high-quality input for multi-objective symmetric semidefinite programming algorithms, filter invalid data, and clarify the relationship between parameters and control objectives. The significance of this platform lies in solving the problems of inaccurate parameter selection and unclear relationships in traditional data processing, ensuring that algorithm optimization is based on effective data, improving the overall accuracy and efficiency of control, and serving as a crucial bridge connecting parameter acquisition and algorithm optimization.

[0051] like Figure 2As shown, a three-phase balanced power supply energy-saving control system includes: a three-phase four-wire soft-switching topology construction unit, which is connected to the main circuit of the power supply system and is used to build a topology structure including the three-phase main circuit branch and the neutral line branch, and output the real-time operating parameters of the topology model; a parameter acquisition and transmission unit, whose input end is connected to the output end of the three-phase four-wire soft-switching topology construction unit, for acquiring real-time operating parameters and transmitting them to a complex domain convex optimization decision unit; a complex domain convex optimization decision unit, whose input end is connected to the output end of the parameter acquisition and transmission unit, for extracting features from the real-time operating parameters and establishing a mapping relationship between calibration parameters and three-phase imbalance and energy loss; and a multi-objective symmetric semidefinite programming calculation unit. The system comprises four sub-units: a first sub-unit, whose input is connected to the output of a complex domain convex optimization decision unit, used to solve for the optimal parameters under constraints with the objectives of minimizing three-phase imbalance and energy loss; an optimal parameter feedback control unit, whose input is connected to the output of a multi-objective symmetric semidefinite programming calculation unit, and whose output is connected to the control unit of a three-phase four-wire soft-switching topology construction unit, used to feed back the optimal parameters to the topology and control its adjustment; and a system operation monitoring unit, whose input is connected to the outputs of both the three-phase four-wire soft-switching topology construction unit and the multi-objective symmetric semidefinite programming calculation unit, used to monitor the three-phase balance state and energy consumption state of the system, and transmit the monitoring results to the complex domain convex optimization decision unit.

[0052] A three-phase balanced power supply energy-saving control method and system, by rationally setting controllable switching elements and filtering components in the three-phase main circuit branches and neutral line branches, can flexibly adjust the conduction sequence and duty cycle of the controllable switching elements, significantly reducing energy loss during switching operation. Simultaneously, relying on the independent controllability of the neutral line branch, it precisely controls neutral line current fluctuations, effectively avoiding additional energy consumption caused by excessive neutral line current. This design directly overcomes the shortcomings of high energy consumption and difficulty in controlling three-phase imbalance in existing technologies, achieving a synergistic improvement in three-phase power supply balance regulation and energy-saving effect, laying a structural foundation for subsequent parameter optimization and precise control.

[0053] At the algorithm and decision support level, this method introduces a multi-objective symmetric semidefinite programming algorithm, coupled with a complex domain convex optimization decision platform, overcoming the limitations of single-objective optimization algorithms in the background technology. The multi-objective symmetric semidefinite programming algorithm can simultaneously optimize for minimizing three-phase imbalance and energy loss, unlike traditional algorithms that only optimize a single index, thus comprehensively considering the balance and energy-saving needs of the power supply system. The complex domain convex optimization decision platform can perform deep feature extraction on the collected real-time operating parameters, accurately screen key parameters affecting balance and energy saving, establish a mapping relationship between key parameters and control objectives, and improve optimization constraints to obtain the optimal solution that better fits the actual operating conditions. This combination of technologies effectively overcomes the shortcomings of existing technologies, such as poor practicality and difficulty in adapting to complex power consumption scenarios, improving the reliability and adaptability of parameter optimization.

[0054] At the overall control process level, this system constructs a complete control link of "parameter acquisition - feature extraction - algorithm solution - optimization feedback," possessing dynamic adjustment capabilities. This represents an optimization and upgrade of the static control mode in the background technology. The system acquires power supply system operating parameters in real time through a parameter acquisition unit. After processing by a complex domain convex optimization decision platform, the optimal parameters are solved using a multi-objective symmetric semi-positive definite programming algorithm. The results are then fed back to the topology model through the optimal parameter feedback control unit, adjusting the operating status of each component. The entire process can adjust the control strategy in real time according to power load fluctuations and parameter changes, rather than the fixed control method of traditional technology. This allows for continuous adaptation to the complex and ever-changing operating conditions of the power supply system, meeting the long-term requirements of the power system for stability and energy efficiency. It further overcomes the problems of existing technologies being unable to cope with changes in operating conditions and having low control effectiveness sustainability, comprehensively improving the operating efficiency of the three-phase power supply system.

[0055] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," "link," and "fix" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0056] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A three-phase balanced power supply energy-saving control method, characterized in that, include: Step S1: Construct a three-phase four-wire soft-switching topology model. This model includes three-phase main circuit branches and a neutral branch. Each branch is equipped with controllable switching elements and filtering components. By adjusting the conduction timing and duty cycle of the controllable switching elements, the output voltage and current waveforms of the topology model meet the preset three-phase electrical characteristics. Step S2: Based on the three-phase four-wire soft-switching topology model, collect real-time operating parameters of the power supply system. These real-time operating parameters include the voltage amplitude of each phase, the current amplitude of each phase, the neutral current amplitude, power factor, and harmonic content. Transmit the collected real-time operating parameters to a complex domain convex optimization decision platform. Step S3: In the complex domain convex optimization decision platform, extract features from the real-time operating parameters to determine the calibration parameters that affect three-phase balance and energy saving effect. Establish a relationship between the calibration parameters and three-phase imbalance, energy saving, and energy efficiency. The mapping relationship between energy loss and power consumption is established. Step S4: The mapping relationship is input into a multi-objective symmetric semidefinite programming algorithm. The algorithm's constraints are constructed with the optimization objectives of minimizing three-phase imbalance and minimizing energy loss. The constraints include the maximum conduction current of the controllable switching element, the fluctuation range of the output voltage, and the limit threshold of the neutral line current. Step S5: The optimal solution under the constraints is obtained by solving the multi-objective symmetric semidefinite programming algorithm. The optimal conduction timing, optimal duty cycle, and optimal parameter configuration of the filter component are obtained. Step S6: The optimal conduction timing, optimal duty cycle, and optimal parameter configuration are fed back to the three-phase four-wire soft-switching topology model. The topology model is controlled to adjust the working state of the controllable switching element in each branch and the operating parameters of the filter component to achieve three-phase balanced power supply and energy-saving control. The voltage output expression of the three-phase four-wire soft-switching topology model is as follows: ,in It is a three-phase output voltage matrix. The topological coefficient matrix, This is the conduction state matrix of a controllable switching element. This is the DC-side input voltage. Here is the inductance matrix for each branch. This is a three-phase output current matrix. The matrix represents the resistance values ​​of each branch; the topology coefficient matrix... The conduction state matrix is ​​determined by the connection method between the three-phase main circuit branch and the neutral line branch. The value of the element is 0 or 1, which corresponds to the off and on states of the controllable switching element, respectively. The objective function expression of the multi-objective symmetric semidefinite programming algorithm is: ,in To comprehensively optimize the target value, This is the weighting coefficient for the three-phase unbalance. This refers to the three-phase voltage imbalance. This is the weighting coefficient for energy loss. The active power loss of the power supply system; the three-phase voltage imbalance This is the maximum amplitude value among the three-phase voltages. This is the minimum amplitude of the three-phase voltage. The average amplitude of the three-phase voltage; active power loss. , They are respectively Three-phase current amplitude, They are respectively Three-phase branch resistance value This is the amplitude of the neutral current. This is the resistance value of the neutral branch; The parameter optimization constraint expression of the complex domain convex optimization decision platform is as follows: ,in The optimized parameter matrix output by the platform. The initial value matrix for the parameters, It is a norm 2. The error threshold is optimized for the parameters; the optimized parameter matrix This includes the duty cycle parameters of the controllable switching elements, the capacitance and inductance values ​​of the filter components, and the initial parameter matrix. The error threshold is determined by the rated operating parameters of the three-phase four-wire soft-switching topology model. Set according to the accuracy requirements of the power supply system; The neutral current control expression for the three-phase four-wire soft-switching topology model is as follows: ,in This is the real-time current of the neutral line. This represents the number of parallel branches in the topology model. The first The real-time currents of phases A, B, and C in a parallel branch are monitored; the conduction state of the controllable switching elements in each parallel branch is adjusted using a multi-objective symmetrical positive definite programming algorithm to ensure the real-time current of the neutral line is controlled. satisfy This is the maximum allowable current for the neutral line; The optimal solution expression for the multi-objective symmetric semi-positive definite programming algorithm is as follows: ,in This is the optimal parameter vector output by the algorithm. For the feasible region of the parameters, For matrix trace operations, Let be the coefficient matrix of the quadratic term of the objective function. The parameter is the coefficient vector of the first-order term of the objective function; the feasible region of the parameter is... The quadratic term coefficient matrix is ​​determined by the conduction timing constraints of the controllable switching element, the output voltage amplitude range, and the current amplitude constraints. With the coefficient vector of the first term The weighting relationship between three-phase balance and energy-saving targets is set accordingly; Step S3 includes the following sub-steps: S31. The collected real-time operating parameters are converted from time domain to frequency domain. The time domain signals of each phase voltage and current are converted into frequency domain signals through Fourier transform, and the fundamental component and each harmonic component in the frequency domain signal are extracted; S32. The fundamental component and harmonic components are feature-filtered, retaining the components with amplitudes exceeding a preset threshold and removing interference components with amplitudes below the threshold. The preset threshold is determined according to the harmonic standard of the power supply system; S33. A correlation function is established between the filtered components and the three-phase unbalance and energy loss. The correlation coefficient between each component and the three-phase unbalance and energy loss is calculated through correlation analysis, and the components with an absolute value of correlation coefficient greater than 0.8 are selected as calibration parameters; S34. The calibration parameters are classified by type to form a voltage calibration parameter set, a current calibration parameter set, and a power calibration parameter set, which are stored in the parameter database of the complex domain convex optimization decision platform, respectively. Step S4 includes the following sub-steps: S41. The mapping relationship obtained in step S3 is transformed into a mathematical expression of a multi-objective symmetric semidefinite programming algorithm, and the optimization variables of the algorithm are defined as the conduction timing, duty cycle, and parameters of the filter component of the controllable switching element; S42. The inequality constraints of the algorithm are set, including the maximum conduction current constraint of the controllable switching element, the maximum fluctuation range constraint of the output voltage, and the upper limit constraint of the neutral line current, and each constraint is transformed into a mathematical inequality; S43. The equality constraints of the algorithm are set, including the three-phase voltage amplitude equality constraint and the three-phase power factor compliance constraint, and each constraint is transformed into a mathematical equation; S44. The inequality constraints and equality constraints are integrated to form a complete constraint system of the multi-objective symmetric semidefinite programming algorithm, ensuring that the optimization variables take values ​​within the constraint system; Step S5 includes the following sub-steps: S51 Initialize the iteration parameters of the multi-objective symmetric semidefinite programming algorithm, including the number of iterations, the iteration step size, and the convergence threshold. Set the initial value of the number of iterations to 100, the initial value of the iteration step size to 0.01, and the initial value of the convergence threshold to 0.001; S52 Input the constraint system constructed in step S4 into the algorithm, perform the first iteration calculation according to the set iteration parameters, and obtain the preliminary optimized parameter values; S53 Calculate the three-phase imbalance and energy loss corresponding to the preliminary optimized parameter values, compare the difference with the result of the previous iteration, and if the difference is greater than the convergence threshold, adjust the iteration step size and continue iterating; S54 Repeat the iteration process until the difference in the three-phase imbalance and the difference in energy loss between two adjacent iterations are both less than the convergence threshold, and output the optimized parameters at this time as the optimal solution.

2. A three-phase balanced power supply energy-saving control system, characterized in that, This system is applied to a three-phase balanced power supply energy-saving control method as described in claim 1, comprising: a three-phase four-wire soft-switching topology construction unit, which is connected to the main circuit of the power supply system and is used to build a topology structure including three-phase main circuit branches and neutral line branches, and output the real-time operating parameters of the topology model; a parameter acquisition and transmission unit, whose input end is connected to the output end of the three-phase four-wire soft-switching topology construction unit, for acquiring real-time operating parameters and transmitting them to a complex domain convex optimization decision unit; a complex domain convex optimization decision unit, whose input end is connected to the output end of the parameter acquisition and transmission unit, for extracting features from the real-time operating parameters and establishing a mapping relationship between calibration parameters and three-phase imbalance and energy loss; and a multi-objective symmetric positive semi-definite... The system comprises the following components: a planning calculation unit, whose input is connected to the output of the complex domain convex optimization decision unit, and a system operation monitoring unit. The planning calculation unit's input is connected to the output of the multi-objective symmetric semi-definite programming calculation unit, and its output is connected to the control unit of the three-phase four-wire soft-switching topology construction unit. This unit feeds back the optimal parameters to the topology and controls its adjustment. The system operation monitoring unit's input is connected to the outputs of both the three-phase four-wire soft-switching topology construction unit and the multi-objective symmetric semi-definite programming calculation unit. This unit monitors the system's three-phase balance and energy consumption status and transmits the monitoring results to the complex domain convex optimization decision unit.

Images