Intelligent aided design method for two-unit power supply system

Abstract

The invention provides an intelligent aided design method for a two-unit power supply system, which belongs to the technical field of intelligent aided design and comprises the following steps: constructing a basic database of the two-unit power supply system; acquiring application scene information aiming at the currently designed two-unit power supply system in real time; the obtained application scene information and data in a basic database are matched and analyzed, a plurality of historical design schemes most matched with the current scene are screened out through an intelligent algorithm to serve as candidate schemes, and according to the candidate schemes, through parameter adjustment and structure optimization in combination with preset design rules and optimization objectives, the optimal design scheme of the current scene is obtained. Preliminarily generating a design scheme of the two-unit power supply system; and performing multi-dimensional simulation verification on the generated design scheme, and performing iterative optimization on the design scheme until the design requirements are met. And the design of the two-unit power supply system is promoted to develop towards the directions of high efficiency, precision and intellectualization.

Description

Technical Field

[0001] This invention relates to the field of intelligent auxiliary design technology, and in particular to an intelligent auxiliary design method for a two-unit power supply system. Background Technology

[0002] With the rapid development of new energy, industrial automation, and communication equipment, two-unit power supply systems (such as master-slave redundant power supplies and AC / DC hybrid power supplies) are increasingly widely used in complex operating scenarios due to their high reliability, flexible scalability, and efficient power conversion capabilities. These systems typically consist of two functional units (such as an input rectifier unit and a power conversion unit, or a main power supply unit and a backup power supply unit) working collaboratively. Their design must consider multiple technical indicators, including circuit topology optimization, component selection, thermal management, and electromagnetic compatibility (EMC), while also meeting the personalized needs of different application scenarios (such as heat dissipation requirements in high-temperature environments and immunity to strong electromagnetic interference). Traditional two-unit power supply system design mainly relies on engineers' experience and trial-and-error methods, which presents the following technical bottlenecks: Existing design methods have limited ability to collect data on application environment (such as temperature, humidity, and electromagnetic interference) and dynamic requirements of load devices, and cannot quantify the impact of environmental parameters on power supply performance in real time (such as component parameter drift caused by high temperature and signal noise caused by strong electromagnetic interference), which often leads to performance deviations in the design scheme during actual operation.

[0003] The core parameters of a two-unit power supply system (such as switching frequency, transformer turns ratio, and capacitor value) are strongly coupled. Traditional optimization algorithms (such as simple genetic algorithms and particle swarm optimization algorithms) have difficulty handling the nonlinear interactions between parameters (such as increasing the switching frequency will improve efficiency but increase EMI noise). Furthermore, they lack systematic modeling of design rules (such as safety specifications and industry standards) and engineering constraints (such as component supply cycles and cost limitations), making them prone to getting trapped in local optima.

[0004] In recent years, although some auxiliary design tools have attempted to introduce database management and simple algorithm optimization, they lack in-depth analysis of the parameter matching relationship and dynamic coordination mechanism between units for the unique dual-unit collaborative control logic of two-unit power systems, making it difficult to meet the design requirements under high reliability scenarios.

[0005] Therefore, this invention proposes an intelligent auxiliary design method for two-unit power supply systems. Summary of the Invention

[0006] This invention provides an intelligent auxiliary design method for two-unit power supply systems, which solves the problems of low efficiency, poor robustness and insufficient consideration of engineering constraints in traditional design processes, and promotes the development of two-unit power supply system design towards high efficiency, precision and intelligence.

[0007] This invention proposes an intelligent auxiliary design method for a two-unit power supply system, comprising: Step 1: Construct a basic database for a two-unit power supply system. The basic database stores historical design data for two-unit power supply systems of different types and specifications. The historical design data includes the circuit topology, component parameters, performance indicators, and fault data in actual operation of each unit, as well as the power supply system requirements of the load equipment. The requirements data include voltage stability requirements, current output range, and power factor standards. Step 2: Obtain in real time the application scenario information of the currently designed two-unit power supply system, wherein the application scenario information includes: temperature, humidity, electromagnetic interference of the application environment, and the working characteristics and special requirements of the load equipment; Step 3: Match and analyze the obtained application scenario information with the data in the basic database, use intelligent algorithms to select several historical design schemes that best fit the current scenario as candidate schemes, and generate a preliminary design scheme for the two-unit power supply system based on the candidate schemes, preset design rules and optimization goals, through parameter adjustment and structural optimization. Step 4: Perform multi-dimensional simulation verification on the generated design scheme, including circuit performance simulation, heat dissipation simulation and electromagnetic compatibility simulation. Iteratively optimize the design scheme based on the simulation results until the design requirements are met.

[0008] Preferably, the process of constructing the basic database for a two-unit power supply system includes: Historical design data is categorized and organized, and multi-dimensional labels are applied according to the type of two-unit power supply system, application field, and component type. In-depth analysis of fault data is conducted to establish a correspondence between fault modes and causes. The correspondence includes component failure problems and failure probabilities, circuit design defects and defect probabilities corresponding to different fault phenomena.

[0009] Preferably, the temperature, humidity, and electromagnetic interference of the application environment are obtained based on data collected by a temperature sensor, a humidity sensor, and an electromagnetic interference sensor, respectively.

[0010] Preferably, several historical design schemes that best fit the current scenario are selected as candidate schemes using intelligent algorithms, including: A fuzzy matching algorithm is used to perform a preliminary match between the application scenario information and the historical design schemes in the basic database. The similarity between each historical scheme and the current scenario is calculated based on environmental parameters and load requirements. Based on the similarity, the top N1 schemes with the highest similarity are selected as candidate schemes.

[0011] Preferably, based on candidate schemes and pre-defined design rules and optimization objectives, a preliminary design scheme for a two-unit power supply system is generated through parameter adjustment and structural optimization, including: The neural network algorithm is used to conduct in-depth analysis of each candidate solution and predict the performance of each candidate solution in the current scenario. The performance includes power efficiency, output voltage fluctuation range and failure rate. Using preset design rules and optimization objectives as the fitness function, and combining the performance data, a recommended parameter combination for each candidate solution is found. The optimal parameter combination for each candidate solution is determined based on the first difference between the current parameter combination and the recommended parameter combination, and the second difference between the recommended parameter combination and the best setting combination. The role of each parameter and the influence between parameters in the optimal parameter combination under all candidate schemes are determined, and a preliminary design scheme for a two-unit power supply system is generated.

[0012] Preferably, the formulation of design rules includes circuit design rules, component selection rules, and system integration rules. Among them, the circuit design rules cover the design specifications of each unit circuit in the two-unit power supply system, the component selection rules specify the performance index range of components according to different design goals and application scenarios, and the system integration rules clarify the connection method and collaborative working mechanism between the two units.

[0013] Preferably, the generated design scheme is subjected to multi-dimensional simulation verification, including: The circuit of the design scheme is modeled and simulated using simulation software. The input and output characteristics, frequency response and harmonic distortion of the circuit are analyzed to verify whether the circuit meets the electrical performance requirements of the load equipment. Establish a three-dimensional thermal model of the circuit, and calculate the temperature distribution and heat flow path of each component based on the power consumption, heat dissipation method and ambient temperature of each component, and evaluate the effectiveness of the heat dissipation design to see if it meets the heat dissipation requirements. The intensity and propagation path of electromagnetic interference in the circuit are predicted by simulation software, and corresponding electromagnetic shielding and filtering measures are designed to determine whether the electromagnetic cancellation requirements are met. Based on the multi-dimensional simulation verification results, the optimization factor for each parameter in the design scheme is determined.

[0014] Preferably, the fitness function is based on preset design rules and optimization objectives, and the recommended parameter combination for each candidate solution is found by combining the performance characteristics, including: A multi-dimensional constraint model is established, which includes circuit design rules, component selection rules, and system integration rules. The multi-dimensional constraint model transforms the electrical parameter boundaries of each unit circuit, the physical characteristic thresholds of components, and the unit collaborative control logic into computable mathematical constraints. A composite fitness function is constructed, which includes an efficiency optimization objective function, a reliability optimization objective function, and a cost optimization objective function. The efficiency objective function uses the power conversion efficiency η as the core variable and introduces dynamic weight coefficients. The reliability objective function calculates the failure rate based on the failure mode database and combines Monte Carlo simulation for probabilistic constraints. The cost objective function includes a full-cycle cost model of component procurement costs and expected maintenance costs. The weighting factors of each objective function are dynamically adjusted by a fuzzy logic controller. An improved non-dominated sorting genetic algorithm is used for parameter space search. A fitness correction strategy based on constraint violation degree is introduced into the genetic operation. The performance data predicted by the neural network is collected in real time as a feedback signal during the search process. The performance trend of the parameter combination is dynamically estimated by an adaptive Kalman filter. When the performance index of a certain parameter combination is detected to fluctuate beyond the preset confidence interval, the crossover probability and mutation probability of the genetic algorithm are automatically adjusted. In the Pareto optimal solution set, a second screening is performed based on the core objectives of the current design task and the historical decision case library of the expert system to generate recommended parameter combinations for each candidate solution.

[0015] Preferably, the recommended parameter combination includes the core device parameters, control strategy parameters, and system-level configuration parameters of each unit circuit.

[0016] Preliminary design scheme for a two-unit power supply system includes: Establish a mapping matrix F containing m design parameters and n performance indicators, where the elements of the mapping matrix F are... This represents the sensitivity of the i-th design parameter to the j-th performance index; Construct the parameter participation matrix S:

[0017] in, It represents the element in the parameter participation matrix S, and represents the normalized coefficient of the effect of the i-th design parameter; Calculate the correlation coefficients between parameters and establish the parameter coupling matrix C;

[0018]

[0019] in, Let represent the joint probability density of the i-th design parameter and the k-th design parameter; These represent the marginal probability densities; This represents the correlation coefficient between the i-th design parameter and the k-th design parameter; This represents the correlation function between the i-th design parameter and the k-th design parameter; Based on the parameter participation matrix S and the parameter coupling matrix C, a dynamic influence model among parameters is constructed to determine the rolling optimization problem. The core adjustment parameters are determined based on the parameter participation ranking. An adjustment sequence is constructed by combining the dynamic influence model between parameters. The gradient descent method is used to solve the optimization problem with coupling constraints and generate a preliminary design scheme. The preliminary design scheme includes: a parameter importance ranking table, a parameter coupling relationship diagram, and a set of parameter adjustment strategies with constraints.

[0020] Compared with the prior art, the beneficial effects of this application are as follows: By rapidly matching historical design schemes with a basic database, redundant design work is reduced. Multi-dimensional simulation verification and iterative optimization ensure that the design scheme meets requirements in terms of circuit performance, heat dissipation, and electromagnetic compatibility, improving the reliability and stability of the power supply system. Real-time acquisition of application scenario information allows the design scheme to fully consider environmental factors and the special requirements of load devices, avoiding performance problems caused by environmental incompatibility. Integrating historical design data and load demand data forms a data accumulation and reuse mechanism, providing data support and decision-making basis for power supply system design.

[0021] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0022] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0023] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of an intelligent auxiliary design method for a two-unit power supply system according to an embodiment of the present invention. Detailed Implementation

[0024] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0025] This invention proposes an intelligent auxiliary design method for two-unit power supply systems, such as... Figure 1 As shown, it includes: Step 1: Construct a basic database for a two-unit power supply system. The basic database stores historical design data for two-unit power supply systems of different types and specifications. The historical design data includes the circuit topology, component parameters, performance indicators, and fault data in actual operation of each unit, as well as the power supply system requirements of the load equipment. The requirements data include voltage stability requirements, current output range, and power factor standards. Step 2: Obtain in real time the application scenario information of the currently designed two-unit power supply system, wherein the application scenario information includes: temperature, humidity, electromagnetic interference of the application environment, and the working characteristics and special requirements of the load equipment; Step 3: Match and analyze the obtained application scenario information with the data in the basic database, use intelligent algorithms to select several historical design schemes that best fit the current scenario as candidate schemes, and generate a preliminary design scheme for the two-unit power supply system based on the candidate schemes, preset design rules and optimization goals, through parameter adjustment and structural optimization. Step 4: Perform multi-dimensional simulation verification on the generated design scheme, including circuit performance simulation, heat dissipation simulation and electromagnetic compatibility simulation. Iteratively optimize the design scheme based on the simulation results until the design requirements are met.

[0026] Preferably, the temperature, humidity, and electromagnetic interference of the application environment are obtained based on data collected by a temperature sensor, a humidity sensor, and an electromagnetic interference sensor, respectively.

[0027] In this embodiment, the two-unit power system is a power system in which two functional units work together. Common types include master-slave redundant power supplies (one master power supply unit and one backup power supply unit, with the backup power supply automatically switching on when the master power supply fails) and AC / DC hybrid power supplies (a combination of an AC input rectifier unit and a DC power conversion unit). For example, master-slave redundant power supplies used in data centers can ensure continuous power supply to servers and improve system reliability. The basic database is used to store and manage data related to the two-unit power system. It adopts a structured storage method to facilitate data querying, retrieval, and analysis. Historical design data refers to the relevant data of previous two-unit power supply systems. Circuit topology refers to the connection method and composition structure of circuits in a power supply system, such as flyback topology, forward topology, and half-bridge topology. For example, low-power switching power supplies often use flyback topology to achieve isolation and voltage reduction functions.

[0028] Component parameters are the specific performance parameters of the components that make up a power supply system. For example, the capacitance and voltage rating of a capacitor; the inductance and saturation current of an inductor; and the voltage rating, current rating, and on-resistance of a power transistor.

[0029] Performance metrics: Parameters that measure the performance of a power supply system, including conversion efficiency (the ratio of output power to input power, such as a power supply with a conversion efficiency of 90%), ripple voltage (the fluctuation range of the output voltage, which is generally required to be within tens of millivolts), dynamic response time (the time it takes for the output voltage to return to a stable state when the load changes), etc.

[0030] Fault data records faults that occur in the power supply system during actual operation, such as input overvoltage faults, output short-circuit faults, and component thermal failures. For example, a record of a power supply's power transistor overheating and burning out due to poor heat dissipation.

[0031] Load devices are devices that are powered by a power supply system.

[0032] Demand data refers to the power system requirements of the load devices.

[0033] Voltage stability requirements refer to the degree of stability of the power supply output voltage when the load device is operating normally. For example, precision instruments may require the power supply output voltage fluctuation range to be within ±1%.

[0034] The current output range is the range of current required by the load device to operate normally. For example, the power requirement of a server might be a current output range of 0-20A.

[0035] The power factor is a metric that measures the efficiency of a power system in utilizing electrical energy from the grid; a higher value indicates higher energy utilization. For example, some industrial equipment requires a power factor of ≥0.95 for the power system.

[0036] In this embodiment, the application scenario information is information related to the actual usage environment of the power system and the load device.

[0037] The application environment is the physical environment in which the power supply system is located.

[0038] The operating characteristics of a load device refer to its working patterns and features, such as operating voltage, current variations, and starting methods. For example, motor-type loads generate a large inrush current when starting.

[0039] Special requirements are the specific needs of the load equipment beyond its basic power supply requirements. For example, medical equipment requires power supply systems with extremely high electromagnetic compatibility to avoid interference with the signals of medical instruments; aerospace equipment requires power supplies to operate stably in wide-temperature and high-vibration environments.

[0040] The candidate solutions are selected from the basic database and are historical design solutions that may be applicable to the current design scenario.

[0041] The preset design rules are pre-defined power system design specifications and guidelines, including circuit design rules (such as design requirements for input filter circuits and topology selection principles for power conversion circuits), component selection rules (such as selecting appropriate transformers and capacitors based on power levels), and system integration rules (such as communication protocols and power distribution strategies between master and slave units).

[0042] The optimization goal is the desired outcome of the design, such as improving the efficiency of the power system, reducing costs, enhancing reliability, and reducing size.

[0043] Parameter adjustment involves changing the parameters of components and circuits in a power supply system design, such as adjusting the switching frequency, transformer turns ratio, and capacitor value.

[0044] Structural optimization involves improving the circuit topology and physical layout of a power supply system. For example, changing the topology of a power supply from a single-ended flyback to a two-transistor forward converter can improve power output capability; optimizing the layout and wiring of circuit boards can enhance electromagnetic compatibility.

[0045] In this embodiment, multi-dimensional simulation verification is a computer simulation verification of the design scheme from multiple aspects.

[0046] Circuit performance simulation uses professional circuit simulation software (such as PSpice and LTspice) to model and simulate power supply circuits. Heat dissipation simulation uses thermal simulation software (such as ANSYS Icepak) to build a three-dimensional thermal model of the power supply system. Electromagnetic compatibility simulation uses electromagnetic simulation software (such as CST) to analyze the electromagnetic radiation generated by the power supply system during operation and its immunity to external electromagnetic interference.

[0047] Temperature sensors are used to measure ambient temperature. Common types include thermocouples, resistance temperature detectors (RTDs), and semiconductor temperature sensors. For example, a type K thermocouple can measure a temperature range of -200℃ to 1300℃ and is often used in industrial high-temperature environments.

[0048] Humidity sensors are sensors that detect ambient humidity, such as capacitive humidity sensors and resistive humidity sensors. Capacitive humidity sensors reflect humidity changes by detecting changes in capacitance.

[0049] Electromagnetic interference (EMI) sensors are sensors used to detect electromagnetic interference signals, measuring the intensity and frequency range of the interference. For example, a loop antenna sensor can be used to detect electromagnetic radiation interference in space.

[0050] The beneficial effects of the above technical solution are as follows: It rapidly matches historical design schemes through a basic database, reducing repetitive design work; multi-dimensional simulation verification and iterative optimization ensure that the design scheme meets requirements in terms of circuit performance, heat dissipation, and electromagnetic compatibility, improving the reliability and stability of the power supply system; real-time collection of application scenario information allows the design scheme to fully consider environmental factors and the special requirements of the load equipment, avoiding performance problems caused by environmental incompatibility; and the integration of historical design data and load demand data forms a data accumulation and reuse mechanism, providing data support and decision-making basis for power supply system design.

[0051] This invention proposes an intelligent auxiliary design method for two-unit power supply systems, which includes the following steps in constructing a basic database for two-unit power supply systems: Historical design data is categorized and organized, and multi-dimensional labels are applied according to the type of two-unit power supply system, application field, and component type. In-depth analysis of fault data is conducted to establish a correspondence between fault modes and causes. The correspondence includes component failure problems and failure probabilities, circuit design defects and defect probabilities corresponding to different fault phenomena.

[0052] In this embodiment, historical design data is a collection of various data generated from previously completed two-unit power system designs, covering information from the entire process from design conception to actual application, and is a digital carrier of design experience and knowledge. Categorization involves dividing complex and disordered historical design data into different categories according to specific logic and rules, making it easier for subsequent retrieval, management, and analysis. Two-unit power supply systems are categorized based on their functional architecture, energy conversion methods, and other characteristics. Common types include: The types of components refer to the various electronic components and devices that make up a power supply system, and are a key factor affecting power supply performance and cost. Semiconductor devices, including power MOSFETs and IGBTs, directly affect power supply efficiency due to their switching speed and conduction losses. For example, silicon carbide MOSFETs are commonly used in high-frequency, high-voltage power supply designs. Magnetic components such as transformers and inductors determine the power transmission and filtering performance of a power supply. High-frequency transformers in switching power supplies require careful design of the turns ratio and core material to achieve efficient energy conversion. Tagging adds descriptive identifiers to historical design data, building a data index system to facilitate quick location and matching of relevant data. For example, a master-slave redundant power supply design scheme for a communication base station could be tagged as "Master-slave redundant power supply | Communication field | Power MOSFET + high-frequency transformer". Fault data is information recorded when a fault occurs in the two-unit power supply system during actual operation, including the time of the fault, the phenomenon, and the scope of impact. Failure mode is a specific manifestation or state that occurs when a power system fails; it is an abstract classification of the fault phenomena. Component failures, such as bulging and leakage of capacitors or broken magnetic cores of inductors, directly lead to abnormal circuit function.

[0053] Circuit design flaws are unreasonable aspects present during the design phase, such as insufficient loop stability or improper heat dissipation path planning. For example, a power supply experienced output voltage oscillations due to an error in the design of its feedback loop parameters. The cause of failure is the fundamental factor that leads to the occurrence of the failure mode, which can be divided into component problems, design errors, and usage environment. The correspondence relationship establishes a mapping connection between failure modes and failure causes, forming a structured knowledge network. For example, "capacitor bulging and leakage (failure mode) → insufficient capacitor withstand voltage (failure cause)". Failure probability is the likelihood of a component failing under specific conditions and within a given time period, typically derived from statistical analysis of extensive historical data. For example, statistics show that a certain type of electrolytic capacitor has a 3% failure probability per year under high-temperature conditions. Defect probability is the probability that a power supply system will fail due to a circuit design flaw. It can be assessed through methods such as fault tree analysis and simulation verification. A certain power supply topology has a common-mode interference design flaw, and its probability of causing a failure is assessed to be 5%. The beneficial effects of the above technical solution are: multi-dimensional labeling enables accurate classification of historical design data; by establishing the correspondence between fault modes and causes and probability analysis, potential risk points in power supply design can be identified in advance; scattered historical design and fault handling experience can be transformed into structured knowledge, avoiding knowledge loss due to personnel turnover; and the classified and organized historical data and fault analysis knowledge provide high-quality training samples for intelligent algorithms, promoting the intelligent and precise development of two-unit power supply system design.

[0054] This invention proposes an intelligent auxiliary design method for two-unit power supply systems. It uses intelligent algorithms to select several historical design schemes that best fit the current scenario as candidate schemes, including: A fuzzy matching algorithm is used to perform a preliminary match between the application scenario information and the historical design schemes in the basic database. The similarity between each historical scheme and the current scenario is calculated based on environmental parameters and load requirements. Based on the similarity, the top N1 schemes with the highest similarity are selected as candidate schemes.

[0055] The beneficial effect of the above technical solution is that it facilitates the screening of candidate solutions based on similarity comparison, providing a basis for subsequent analysis.

[0056] This invention proposes an intelligent auxiliary design method for two-unit power supply systems. Based on candidate schemes combined with preset design rules and optimization objectives, and through parameter adjustment and structural optimization, a preliminary design scheme for the two-unit power supply system is generated, including: The neural network algorithm is used to conduct in-depth analysis of each candidate solution and predict the performance of each candidate solution in the current scenario. The performance includes power efficiency, output voltage fluctuation range and failure rate. Using preset design rules and optimization objectives as the fitness function, and combining the performance data, a recommended parameter combination for each candidate solution is found. The optimal parameter combination for each candidate solution is determined based on the first difference between the current parameter combination and the recommended parameter combination, and the second difference between the recommended parameter combination and the best setting combination. The role of each parameter and the influence between parameters in the optimal parameter combination under all candidate schemes are determined, and a preliminary design scheme for a two-unit power supply system is generated.

[0057] Preferably, the formulation of design rules includes circuit design rules, component selection rules, and system integration rules. Among them, the circuit design rules cover the design specifications of each unit circuit in the two-unit power supply system, the component selection rules specify the performance index range of components according to different design goals and application scenarios, and the system integration rules clarify the connection method and collaborative working mechanism between the two units.

[0058] In this embodiment, performance is an indicator reflecting the working status and quality of the two-unit power supply system under specific scenarios, specifically including: Power efficiency: The ratio of power output to power input, it is a key indicator of a power supply's energy conversion capability, usually expressed as a percentage. For example, if a power supply system has an input power of 1000W and an output power of 920W, then its power efficiency is 92%. Output voltage fluctuation range: The degree to which the power supply output voltage deviates from the nominal value under certain conditions (such as load changes, input voltage fluctuations). For example, if a power supply has a nominal output voltage of 24V, its output voltage fluctuation range should be controlled within ±0.5V. Failure rate: The probability of a power supply system failing under certain time periods and operating conditions, usually expressed as a probability value. For example, statistical analysis shows that the annual failure rate of a certain type of power supply under normal operating conditions is 2%. Pre-defined design rules, such as a series of specifications and guidelines formulated based on industry standards, engineering experience and actual needs before designing a two-unit power supply system, including circuit design rules (such as the calculation method of heat dissipation area of ​​power devices), component selection rules (such as the capacitor withstand voltage value must be greater than 1.5 times the maximum operating voltage), system integration rules (such as master-slave unit communication protocol and synchronization mechanism), etc. Optimization goals are the performance or cost objectives that designers expect the power system to achieve, such as improving power efficiency to over 95%, keeping costs within budget, and enhancing system reliability to reduce the failure rate to below 1%. The fitness function is a mathematical function used in optimization algorithms to evaluate the quality of candidate solutions, transforming pre-defined design rules and optimization objectives into quantifiable metrics.

[0059] The recommended parameter combinations are based on the evaluation of candidate scheme performance using a fitness function, combined with design rules and optimization objectives, to provide a set of optimal component and circuit parameter combinations for each candidate scheme. For example, for a certain candidate scheme, it is recommended to adjust the switching frequency from 50kHz to 60kHz and the transformer turns ratio from 10:1 to 12:1 to improve power supply efficiency and meet output voltage stability requirements. The current parameter combination is the specific set of values ​​for the component and circuit parameters currently used in the candidate solution. For example, the current parameter combination of a candidate solution is a switching frequency of 50kHz, a transformer turns ratio of 10:1, and a filter capacitor value of 100μF. The first difference is the gap between the current parameter combination of the candidate solution and the recommended parameter combination, which can be quantified by calculating the sum of squares of the differences between each parameter or the Euclidean distance. For example, if the switching frequency is 50kHz in the current parameter combination and 60kHz in the recommended parameter combination, the difference in the switching frequency parameter is 10kHz. The optimal parameter combination is the ideal combination that best satisfies the preset design rules and optimization objectives among all possible parameter combinations. It is usually obtained through theoretical analysis, extensive simulation or actual testing. The second difference is the gap between the recommended parameter combination and the optimal setting combination, which is used to further evaluate the optimization potential of the recommended parameter combination. The optimal parameter combination is obtained by comprehensively considering the first and second differences and adjusting the parameters of the candidate schemes. It is the parameter combination that best meets the design requirements under the current conditions. For example, through continuous adjustment and evaluation, the optimal parameter combination of a certain candidate scheme is determined to be a switching frequency of 62kHz, a transformer turns ratio of 12.5:1, and a filter capacitor value of 120μF. The role of parameters in a power supply system refers to the degree and manner in which each parameter contributes to achieving design goals (such as power efficiency and output stability). For example, the switching frequency parameter has a significant impact on both power efficiency and electromagnetic interference levels. Its role is reflected in the fact that adjusting the switching frequency can optimize power efficiency, but may also increase electromagnetic interference. Inter-parameter influence refers to the interrelationship and mutual constraint between different parameters in a power supply system. For example, the transformer turns ratio is directly related to the output voltage and current, and it also affects the voltage stress and current magnitude of the power transistor. In other words, changes in the transformer turns ratio will affect the selection of power transistor parameters and performance. The initial design scheme is generated based on the analysis of the optimal parameter combination of all candidate schemes. Combining the role of parameters and the influence relationship between parameters, a two-unit power supply system design scheme that meets the design requirements is formed, including circuit schematics, component list, parameter setting instructions, etc. The beneficial effects of the above technical solution are as follows: by utilizing the powerful nonlinear fitting capability of neural network algorithms, it is possible to accurately predict the performance of candidate solutions in complex scenarios; by constructing a fitness function based on preset design rules and optimization objectives, and combining the performance prediction results, it is possible to quickly find the optimal parameter combination for each candidate solution; by analyzing the participation and mutual influence of parameters, it is possible to gain a deep understanding of the intrinsic relationship between power system parameters; and by performing intelligent analysis of the entire process from candidate solution selection to parameter optimization and design solution generation, it is possible to improve the success rate of two-unit power system design. This invention proposes an intelligent auxiliary design method for two-unit power supply systems, which performs multi-dimensional simulation verification on the generated design scheme, including: The circuit of the design scheme is modeled and simulated using simulation software. The input and output characteristics, frequency response and harmonic distortion of the circuit are analyzed to verify whether the circuit meets the electrical performance requirements of the load equipment. Establish a three-dimensional thermal model of the circuit, and calculate the temperature distribution and heat flow path of each component based on the power consumption, heat dissipation method and ambient temperature of each component, and evaluate the effectiveness of the heat dissipation design to see if it meets the heat dissipation requirements. The intensity and propagation path of electromagnetic interference in the circuit are predicted by simulation software, and corresponding electromagnetic shielding and filtering measures are designed to determine whether the electromagnetic cancellation requirements are met. Based on the multi-dimensional simulation verification results, the optimization factor for each parameter in the design scheme is determined.

[0060] In this embodiment, the temperature distribution refers to the temperature levels and variations in various parts of the power supply system, calculated using a thermal model. As the thermal simulation results show, the temperature near the power transistors on the power circuit board is the highest, reaching 80°C, while other areas have relatively lower temperatures. Heat flow paths are the routes through which heat is transferred within a power supply system. Understanding these paths helps optimize heat dissipation design. For example, thermal simulations may reveal that heat is primarily conducted from the power transistors through the circuit board to the heatsink. An improperly designed heatsink can lead to heat buildup.

[0061] In this embodiment, the simulation verification results serve as the quantitative basis for adjusting various parameters in the design scheme, guiding the direction and magnitude of parameter optimization. For example, if the simulation reveals low power supply efficiency, analysis may determine that adjusting the switching frequency and transformer turns ratio can improve efficiency; in this case, the adjustment ratio of the switching frequency and transformer turns ratio becomes the optimization factor.

[0062] The beneficial effects of the above technical solution are: through multi-dimensional simulation verification, the performance of the design scheme in terms of electrical performance, heat dissipation and electromagnetic compatibility can be comprehensively evaluated, potential problems can be discovered in advance, and simulation technology can be used for optimization in the design stage to reduce the number of physical prototypes and tests. For the electrical performance requirements of different load devices, heat dissipation and electromagnetic compatibility requirements of complex application environments, through simulation and optimization, the design scheme can be ensured to meet diverse actual needs.

[0063] This invention proposes an intelligent auxiliary design method for two-unit power supply systems. It uses preset design rules and optimization objectives as fitness functions, and combines these with performance data to find recommended parameter combinations for each candidate solution, including: A multi-dimensional constraint model is established, which includes circuit design rules, component selection rules, and system integration rules. The multi-dimensional constraint model transforms the electrical parameter boundaries of each unit circuit, the physical characteristic thresholds of components, and the unit collaborative control logic into computable mathematical constraints. A composite fitness function is constructed, which includes an efficiency optimization objective function, a reliability optimization objective function, and a cost optimization objective function. The efficiency objective function uses the power conversion efficiency η as the core variable and introduces dynamic weight coefficients. The reliability objective function calculates the failure rate based on the failure mode database and combines Monte Carlo simulation for probabilistic constraints. The cost objective function includes a full-cycle cost model of component procurement costs and expected maintenance costs. The weighting factors of each objective function are dynamically adjusted by a fuzzy logic controller. An improved non-dominated sorting genetic algorithm is used for parameter space search. A fitness correction strategy based on constraint violation degree is introduced into the genetic operation. The performance data predicted by the neural network is collected in real time as a feedback signal during the search process. The performance trend of the parameter combination is dynamically estimated by an adaptive Kalman filter. When the performance index of a certain parameter combination is detected to fluctuate beyond the preset confidence interval, the crossover probability and mutation probability of the genetic algorithm are automatically adjusted. In the Pareto optimal solution set, a second screening is performed based on the core objectives of the current design task and the historical decision case library of the expert system to generate recommended parameter combinations for each candidate solution.

[0064] The beneficial effects of the above technical solutions are as follows: the multi-dimensional constraint model transforms circuit design, component selection, and system integration rules into mathematical constraints, ensuring that the design scheme strictly follows engineering specifications and standards, reducing potential risks caused by design oversights; the combination of composite fitness function and fuzzy logic controller can dynamically adjust the design target weights according to different scenarios, achieving an optimal balance between efficiency, reliability, and cost, meeting diverse application needs; the improved non-dominated sorting genetic algorithm combined with an adaptive feedback mechanism effectively avoids getting trapped in local optima, improving the efficiency and accuracy of searching for global optimal solutions in complex parameter spaces, shortening the design optimization cycle; and the secondary screening based on the expert system's historical case library provides a scientific basis and practical experience reference for the final determination of the design scheme, making the recommended parameter combinations more engineering feasible and practical, and improving the overall design quality of the two-unit power supply system.

[0065] This invention proposes an intelligent auxiliary design method for a two-unit power supply system. The recommended parameter combination includes the core component parameters, control strategy parameters, and system-level configuration parameters of each unit circuit.

[0066] Preliminary design scheme for a two-unit power supply system includes: Establish a mapping matrix F containing m design parameters and n performance indicators, where the elements of the mapping matrix F are... This represents the sensitivity of the i-th design parameter to the j-th performance index; Construct the parameter participation matrix S:

[0067] in, It represents the element in the parameter participation matrix S, and represents the normalized coefficient of the effect of the i-th design parameter; Calculate the correlation coefficients between parameters and establish the parameter coupling matrix C;

[0068]

[0069] in, Let represent the joint probability density of the i-th design parameter and the k-th design parameter; These represent the marginal probability densities; This represents the correlation coefficient between the i-th design parameter and the k-th design parameter; This represents the correlation function between the i-th design parameter and the k-th design parameter; Based on the parameter participation matrix S and the parameter coupling matrix C, a dynamic influence model among parameters is constructed to determine the rolling optimization problem. The core adjustment parameters are determined based on the parameter participation ranking. An adjustment sequence is constructed by combining the dynamic influence model between parameters. The gradient descent method is used to solve the optimization problem with coupling constraints and generate a preliminary design scheme. The preliminary design scheme includes: a parameter importance ranking table, a parameter coupling relationship diagram, and a set of parameter adjustment strategies with constraints.

[0070] In this embodiment, design parameters are key variables that determine the performance and structure of a two-unit power supply system, encompassing circuit parameters (such as switching frequency and transformer turns ratio) and component parameters (such as capacitor values ​​and power transistor models). For example, in flyback power supply design, the transformer turns ratio and the MOSFET on-resistance are important design parameters. Performance metrics are quantitative standards used to measure the operating quality of a power supply system, including power efficiency, output voltage ripple, and electromagnetic radiation intensity. For example, a communication power supply may require an output voltage ripple of no more than 50mV to ensure signal transmission stability.

[0071] Rolling optimization is an iterative optimization task based on parameter participation and coupling relationships, aiming at short-term performance improvements. For example, in scenarios with sudden changes in power load, the parameter adjustment scheme is recalculated every 10ms to ensure rapid system response. A dynamic influence model between parameters is a mathematical model that describes the dynamic impact of parameter adjustments on other parameters and performance indicators. It is usually constructed based on transfer functions or state-space equations. For example, in a power supply closed-loop control system, after adjusting the parameters of a PI controller, the transient process of the output voltage is predicted through a dynamic model.

[0072] Parameter participation ranking is based on the parameter participation matrix S, which arranges all design parameters from highest to lowest contribution. For example, the ranking results show that switching frequency and transformer turns ratio are the top two core parameters, and prioritizing their optimization can significantly improve power supply efficiency. Core tuning parameters are a subset of parameters that play a decisive role in achieving the design goals; typically, these are the top 30% of parameters in terms of their importance. For example, in low-ripple power supply design, the output filter capacitor and inductor parameters are considered core tuning parameters. The adjustment sequence is a strategy for determining the order and synergistic relationships of parameter optimization. For example, the switching frequency, which has the greatest impact on efficiency, is adjusted first, and then the power transistor drive resistor is optimized synchronously based on the coupling relationship. The parameter importance ranking table presents the participation ranking, sensitivity value, and optimization priority of design parameters in tabular form, providing designers with an intuitive reference. Parameter coupling diagrams use nodes and connections to visualize the coupling strength and direction of influence between parameters, helping to identify the chain reaction of parameter adjustments. For example, the arrow in the diagram points from the inductor to the output ripple, indicating that changes in inductance will affect the ripple magnitude.

[0073] A set of constrained parameter adjustment strategies is a parameter optimization scheme formulated in conjunction with design rule constraints (such as safety specifications and cost limits), including adjustment range, priority, and risk assessment. For example, it may stipulate that the withstand voltage value of the power transistor must have a 20% margin, thus limiting the parameter adjustment range.

[0074] The beneficial effects of the above technical solutions are: revealing the implicit correlation between parameters based on the parameter coupling matrix and dynamic influence model, preventing system performance degradation caused by local optimization, reducing the number of design iterations, automatically constructing parameter adjustment sequences based on mathematical models and algorithms, generating feasible solutions in combination with engineering constraints, and graphically presenting complex parameter relationships through parameter importance ranking tables and coupling relationship diagrams, thereby lowering the design threshold and improving team collaboration efficiency.

[0075] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. An intelligent auxiliary design method for a two-unit power supply system, characterized in that, include: Step 1: Construct a basic database for a two-unit power supply system. The basic database stores historical design data for two-unit power supply systems of different types and specifications. The historical design data includes the circuit topology, component parameters, performance indicators, and fault data in actual operation of each unit, as well as the power supply system requirements of the load equipment. The requirements data include voltage stability requirements, current output range, and power factor standards. Step 2: Obtain in real time the application scenario information of the currently designed two-unit power supply system, wherein the application scenario information includes: temperature, humidity, electromagnetic interference of the application environment, and the working characteristics and special requirements of the load equipment; Step 3: Match and analyze the obtained application scenario information with the data in the basic database, use intelligent algorithms to select several historical design schemes that best fit the current scenario as candidate schemes, and generate a preliminary design scheme for the two-unit power supply system based on the candidate schemes, preset design rules and optimization goals, through parameter adjustment and structural optimization. Step 4: Perform multi-dimensional simulation verification on the generated design scheme, including circuit performance simulation, heat dissipation simulation and electromagnetic compatibility simulation. Iteratively optimize the design scheme based on the simulation results until the design requirements are met.

2. The intelligent auxiliary design method for a two-unit power supply system according to claim 1, characterized in that, The process of constructing the basic database for a two-unit power supply system includes: Historical design data is categorized and organized, and multi-dimensional labels are applied according to the type of two-unit power supply system, application field, and component type. In-depth analysis of fault data is conducted to establish a correspondence between fault modes and causes. The correspondence includes component failure problems and failure probabilities, circuit design defects and defect probabilities corresponding to different fault phenomena.

3. The intelligent auxiliary design method for a two-unit power supply system according to claim 1, characterized in that, The temperature, humidity, and electromagnetic interference of the application environment are obtained based on data collected by temperature sensors, humidity sensors, and electromagnetic interference sensors, respectively.

4. The intelligent auxiliary design method for a two-unit power supply system according to claim 1, characterized in that, Intelligent algorithms were used to select several historical design schemes that best fit the current scenario as candidate schemes, including: A fuzzy matching algorithm is used to perform a preliminary match between the application scenario information and the historical design schemes in the basic database. The similarity between each historical scheme and the current scenario is calculated based on environmental parameters and load requirements. Based on the similarity, the top N1 schemes with the highest similarity are selected as candidate schemes.

5. The intelligent auxiliary design method for a two-unit power supply system according to claim 1, characterized in that, Based on the candidate schemes and the preset design rules and optimization objectives, a preliminary design scheme for a two-unit power supply system is generated through parameter adjustment and structural optimization, including: The neural network algorithm is used to conduct in-depth analysis of each candidate solution and predict the performance of each candidate solution in the current scenario. The performance includes power efficiency, output voltage fluctuation range and failure rate. Using preset design rules and optimization objectives as the fitness function, and combining the performance data, a recommended parameter combination for each candidate solution is found. The optimal parameter combination for each candidate solution is determined based on the first difference between the current parameter combination and the recommended parameter combination, and the second difference between the recommended parameter combination and the best setting combination. The role of each parameter and the influence between parameters in the optimal parameter combination under all candidate schemes are determined, and a preliminary design scheme for a two-unit power supply system is generated.

6. The intelligent auxiliary design method for a two-unit power supply system according to claim 5, characterized in that, The formulation of design rules includes circuit design rules, component selection rules, and system integration rules. Among them, the circuit design rules cover the design specifications of each unit circuit in the two-unit power supply system, the component selection rules specify the performance index range of components according to different design goals and application scenarios, and the system integration rules clarify the connection method and collaborative working mechanism between the two units.

7. The intelligent auxiliary design method for a two-unit power supply system according to claim 6, characterized in that, The generated design scheme is verified through multi-dimensional simulation, including: The circuit of the design scheme is modeled and simulated using simulation software. The input and output characteristics, frequency response and harmonic distortion of the circuit are analyzed to verify whether the circuit meets the electrical performance requirements of the load equipment. Establish a three-dimensional thermal model of the circuit, and calculate the temperature distribution and heat flow path of each component based on the power consumption, heat dissipation method and ambient temperature of each component, and evaluate the effectiveness of the heat dissipation design to see if it meets the heat dissipation requirements. The intensity and propagation path of electromagnetic interference in the circuit are predicted by simulation software, and corresponding electromagnetic shielding and filtering measures are designed to determine whether the electromagnetic cancellation requirements are met. Based on the multi-dimensional simulation verification results, the optimization factor for each parameter in the design scheme is determined.

8. The intelligent auxiliary design method for a two-unit power supply system according to claim 5, characterized in that, Using preset design rules and optimization objectives as the fitness function, and combining the performance data, a recommended parameter combination for each candidate solution is found, including: A multi-dimensional constraint model is established, which includes circuit design rules, component selection rules, and system integration rules. The multi-dimensional constraint model transforms the electrical parameter boundaries of each unit circuit, the physical characteristic thresholds of components, and the unit collaborative control logic into computable mathematical constraints. A composite fitness function is constructed, which includes an efficiency optimization objective function, a reliability optimization objective function, and a cost optimization objective function. The efficiency objective function uses the power conversion efficiency η as the core variable and introduces dynamic weight coefficients. The reliability objective function calculates the failure rate based on the failure mode database and combines Monte Carlo simulation for probabilistic constraints. The cost objective function includes a full-cycle cost model of component procurement costs and expected maintenance costs. The weighting factors of each objective function are dynamically adjusted by a fuzzy logic controller. An improved non-dominated sorting genetic algorithm is used for parameter space search. A fitness correction strategy based on constraint violation degree is introduced into the genetic operation. The performance data predicted by the neural network is collected in real time as a feedback signal during the search process. The performance trend of the parameter combination is dynamically estimated by an adaptive Kalman filter. When the performance index of a certain parameter combination is detected to fluctuate beyond the preset confidence interval, the crossover probability and mutation probability of the genetic algorithm are automatically adjusted. In the Pareto optimal solution set, a second screening is performed based on the core objectives of the current design task and the historical decision case library of the expert system to generate recommended parameter combinations for each candidate solution.

9. The intelligent auxiliary design method for a two-unit power supply system according to claim 8, characterized in that, The recommended parameter combination includes the core component parameters, control strategy parameters, and system-level configuration parameters for each unit circuit.

10. The intelligent auxiliary design method for a two-unit power supply system according to claim 5, characterized in that, A preliminary design scheme for a two-unit power supply system has been generated, including: Establish a mapping matrix F containing m design parameters and n performance indicators, where the elements of the mapping matrix F are... This represents the sensitivity of the i-th design parameter to the j-th performance index; Construct the parameter participation matrix S: in, It represents the element in the parameter participation matrix S, and represents the normalized coefficient of the effect of the i-th design parameter; Calculate the correlation coefficients between parameters and establish the parameter coupling matrix C; in, Let represent the joint probability density of the i-th design parameter and the k-th design parameter; These represent the marginal probability densities; This represents the correlation coefficient between the i-th design parameter and the k-th design parameter; This represents the correlation function between the i-th design parameter and the k-th design parameter; Based on the parameter participation matrix S and the parameter coupling matrix C, a dynamic influence model among parameters is constructed to determine the rolling optimization problem. The core adjustment parameters are determined based on the parameter participation ranking. An adjustment sequence is constructed by combining the dynamic influence model between parameters. The gradient descent method is used to solve the optimization problem with coupling constraints and generate a preliminary design scheme. The preliminary design scheme includes: a parameter importance ranking table, a parameter coupling relationship diagram, and a set of parameter adjustment strategies with constraints.

Images