Multi-objective cooperative AI converter control system and method based on complex operating conditions

Abstract

This invention provides a multi-objective collaborative AI converter control system and method based on complex operating conditions, belonging to the field of converter control technology. The system includes a complex operating condition perception module, a multi-objective setting and decomposition module, an AI collaborative control module, and a monitoring and evaluation module. The complex operating condition perception module is used to acquire operating condition characteristic data. The multi-objective setting and decomposition module is used to set control strategies for multiple objectives and decompose the multiple objectives into sub-objectives. The AI ​​collaborative control module is used to acquire control commands for the converter switching frequency, adjust the converter's modulation method, and adjust the heat dissipation strategy. The monitoring and evaluation module is used to evaluate the degree of objective completion. This invention, through the collaborative work of multiple modules, achieves intelligent and refined control and monitoring of the converter, effectively improving the converter's performance and reliability while reducing operating costs.

Description

Technical Field

[0001] This invention relates to the field of converter control technology, and in particular to a multi-objective cooperative AI converter control system and a multi-objective cooperative AI converter control method based on complex operating conditions. Background Technology

[0002] In the field of electric two-wheelers, the converter, as the core device for energy conversion, often operates under extremely complex conditions. In practical applications, the converter faces a variety of uncertainties such as input voltage / current fluctuations, sudden load changes, temperature variations, and electromagnetic interference. These complex conditions place stringent demands on the converter's control performance.

[0003] Traditional converter control systems often employ a single control objective (such as pursuing only output waveform quality or efficiency), and the control strategies are mostly based on fixed-model PID control or classical control algorithms. However, under complex operating conditions, single-objective control struggles to simultaneously address multiple performance indicators of the converter, such as dynamic response speed, stability, efficiency, and harmonic suppression. Furthermore, fixed control strategies cannot adapt to changes in operating conditions, easily leading to decreased control accuracy, system instability, or even equipment damage. Summary of the Invention

[0004] This invention provides a multi-objective cooperative AI converter control system and method based on complex operating conditions, in order to solve the defects existing in the prior art.

[0005] On one hand, the present invention provides a multi-objective cooperative AI converter control system based on complex operating conditions, comprising:

[0006] The complex operating condition sensing module is used to collect the operating condition data of the converter and process the operating condition data to obtain operating condition characteristic data.

[0007] The multi-objective setting and decomposition module is used to set the control strategy for multiple objectives based on the working condition characteristic data and the particle swarm algorithm, and decompose the multiple objectives into sub-objectives.

[0008] The AI ​​collaborative control module is used to obtain control commands for the inverter switching frequency, adjust the inverter modulation method, and adjust the heat dissipation strategy based on sub-objectives using artificial intelligence.

[0009] The monitoring and evaluation module is used to collect converter operating data in real time and compare it with sub-targets to evaluate the degree of target completion.

[0010] According to the multi-objective cooperative AI converter control system based on complex operating conditions provided by the present invention, the complex operating condition perception module includes a data collection unit, a data processing unit, and a feature extraction unit. The data collection unit collects voltage, current, temperature, humidity, and mechanical vibration data of the converter through sensors and performs signal conversion to obtain operating condition data. The data processing unit uses a filtering algorithm to clean the noise from the operating condition data, obtaining preprocessed data. The feature extraction unit uses principal component analysis to extract operating condition features from the preprocessed data, including voltage and current variation trends, environmental characteristics, and mechanical vibration characteristics.

[0011] The multi-objective cooperative AI converter control system based on complex operating conditions provided by this invention includes a multi-objective setting and decomposition module comprising an objective determination unit, an efficiency objective decomposition unit, a power quality objective decomposition unit, and an equipment reliability objective decomposition unit. The objective determination unit is used to set multi-objective control strategies based on operating condition characteristic data using a particle swarm optimization algorithm. These multi-objective control strategies include maximizing efficiency, optimizing power quality, and extending equipment lifespan. The efficiency objective decomposition unit decomposes the efficiency maximization objective into power factor optimization and loss reduction sub-objectives. The power quality objective decomposition unit decomposes the power quality objective into harmonic suppression and voltage stability sub-objectives. The equipment reliability objective decomposition unit transforms the equipment lifespan extension objective into temperature control and stress equalization sub-objectives.

[0012] According to the multi-objective cooperative AI converter control system based on complex operating conditions provided by the present invention, the process of setting the multi-objective control strategy using the particle swarm optimization algorithm includes:

[0013] Initialize the particle swarm parameters, where each particle represents a combination of control strategies.

[0014] Define a multi-objective function, which calculates fitness based on efficiency, power quality, and equipment reliability indicators.

[0015] The particle position and velocity are iteratively updated, and the strategy combination is adjusted through global and local optimum searches to output the optimized control strategy.

[0016] The multi-objective collaborative AI converter control system based on complex operating conditions provided by this invention includes an AI collaborative control module comprising an efficiency control unit, a power quality control unit, and an equipment lifespan control unit. The efficiency control unit is used to construct a control model based on a deep neural network, taking power factor optimization and loss reduction sub-objectives as inputs, and outputting control commands for the converter switching frequency. The power quality control unit is used to adjust the converter's modulation method using a convolutional neural network based on harmonic suppression and voltage stability sub-objectives. The equipment lifespan control unit is used to adjust the heat dissipation strategy using an optimized control algorithm based on temperature control and stress equalization sub-objectives.

[0017] According to the multi-objective cooperative AI converter control system based on complex operating conditions provided by the present invention, the process of constructing a control model based on a deep neural network includes:

[0018] The design incorporates a deep neural network-based basic model, comprising an input layer, hidden layers, and an output layer. The input layer receives operating condition feature data and sub-objectives. The hidden layer consists of multiple fully connected layers using ReLU activation. The output layer generates switching frequency control commands.

[0019] Historical operational data is collected, and the mean squared error is used as the loss function to train the basic model. Model parameters that meet the preset accuracy are retained to obtain the control model.

[0020] According to the multi-objective cooperative AI converter control system based on complex operating conditions provided by the present invention, the process of adjusting the modulation mode of the converter using a convolutional neural network includes:

[0021] A fuzzy logic rule base is constructed based on the harmonic suppression target.

[0022] Based on a fuzzy logic rule base, a convolutional neural network is trained to predict the modulation mode according to the voltage stability sub-objective.

[0023] The pulse width modulation signal parameter adjustment command is output according to the modulation method.

[0024] Based on the adjustment instructions, the model parameters are optimized through a feedback loop.

[0025] According to the multi-objective cooperative AI converter control system based on complex operating conditions provided by the present invention, the process of adjusting the heat dissipation strategy using an optimized control algorithm includes:

[0026] Build a temperature prediction model by inputting device thermal distribution data.

[0027] Define an optimization function to minimize temperature fluctuations and stress differences.

[0028] The model predictive control algorithm is used to generate an optimized scheme for heat dissipation parameters.

[0029] Perform real-time adjustments to cooling fan speed and radiator layout.

[0030] According to the multi-objective collaborative AI converter control system based on complex operating conditions provided by the present invention, the monitoring and evaluation module includes a real-time data acquisition unit, an objective comparison unit, a trend prediction unit, and an evaluation feedback unit. The real-time data acquisition unit is used to acquire converter operating data in real time, including electrical performance data, efficiency-related data, and control command execution data. The objective comparison unit is used to compare the operating data with multiple objectives to obtain the objective difference values. The trend prediction unit is used to construct a trend prediction model based on time series analysis, taking the objective difference values ​​as input and outputting the predicted change trends of the multiple objectives. The evaluation feedback unit is used to generate an evaluation report based on the predicted change trends and provide feedback.

[0031] On the other hand, the present invention also provides a multi-objective cooperative AI converter control method based on complex operating conditions, comprising:

[0032] Collect the operating condition data of the converter and process the operating condition data to obtain the operating condition characteristic data.

[0033] Based on the operating condition characteristic data, a multi-objective control strategy is set using the particle swarm optimization algorithm, and the multi-objective is decomposed into sub-objectives.

[0034] Based on the sub-objectives, artificial intelligence is used to obtain control commands for the inverter's switching frequency, adjust the inverter's modulation method, and adjust the heat dissipation strategy.

[0035] Real-time acquisition of converter operating data is used to compare with sub-targets and assess the degree of target achievement.

[0036] This invention provides a multi-objective collaborative AI converter control system and method based on complex operating conditions. Through data collection, processing, and feature extraction units, it can accurately collect operating condition data such as voltage, current, and temperature of the converter, effectively remove noise interference, and extract key operating condition features. The multi-objective setting and decomposition module scientifically sets multi-objective control strategies based on the operating condition feature data using a particle swarm optimization algorithm, and decomposes these strategies into sub-objectives. This not only comprehensively considers multiple objectives such as maximizing efficiency, optimizing power quality, and extending equipment lifespan, but also refines macro-objectives into specific and operable sub-objectives, improving the targeting and executability of the control strategy. The AI ​​collaborative control module utilizes artificial intelligence technology to perform precise control for different sub-objectives. The efficiency control unit optimizes the switching frequency through a deep neural network, the power quality control unit adjusts the modulation method using a convolutional neural network, and the equipment lifespan control unit adjusts the heat dissipation strategy using an optimized control algorithm. This significantly improves the converter's operating efficiency, enhances power quality, effectively extends equipment lifespan, and reduces equipment maintenance costs. The monitoring and evaluation module collects the converter's operating data in real time, compares it with the sub-objectives, and predicts the changing trends of the multiple objectives through a trend prediction model. The generated assessment report can provide timely feedback on the system's operating status, helping operators to identify problems and adjust control strategies in a timely manner, ensuring that the converter is always in a stable and efficient operating state. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0038] Figure 1 This is a schematic diagram of the structure of a multi-objective cooperative AI converter control system based on complex operating conditions provided in an embodiment of the present invention;

[0039] Figure 2 This is a flowchart illustrating the multi-objective collaborative AI converter control method based on complex operating conditions provided in this embodiment of the invention. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0041] The following is combined with Figures 1-2 This invention describes a multi-objective cooperative AI converter control system and method based on complex operating conditions.

[0042] Figure 1 This is a schematic diagram of the structure of a multi-objective cooperative AI converter control system based on complex operating conditions provided in an embodiment of the present invention.

[0043] like Figure 1 As shown in the embodiments of the present invention, the multi-objective collaborative AI converter control system and method based on complex operating conditions can be executed by a multi-objective collaborative AI converter control system based on complex operating conditions. The system includes a complex operating condition perception module, a multi-objective setting decomposition module, an AI collaborative control module, and a monitoring and evaluation module.

[0044] The complex operating condition sensing module is used to collect the operating condition data of the converter and process the operating condition data to obtain operating condition characteristic data.

[0045] The complex working condition perception module includes a data collection unit, a data processing unit, and a feature extraction unit.

[0046] The data acquisition unit is used to collect voltage, current, temperature, humidity and mechanical vibration of the converter through sensors and convert the signals to obtain operating condition data.

[0047] This unit uses various sensors to collect key physical quantities of the converter, including voltage, current, temperature, humidity, and mechanical vibration. Different physical quantities require different types of sensors; for example, voltage sensors measure voltage, current sensors measure current, temperature sensors, such as thermocouples and thermistors, measure temperature, humidity sensors measure humidity, and acceleration sensors measure mechanical vibration.

[0048] The signals collected by sensors are usually analog signals, while subsequent data processing equipment can generally only process digital signals, so signal conversion is required. This conversion process is completed by an analog-to-digital converter, which converts the analog signal into a digital signal, ultimately obtaining the operating condition data that can be processed later.

[0049] The data processing unit is used to clean the noise from the operating data using a filtering algorithm to obtain preprocessed data.

[0050] The collected operating data may contain various types of noise, which can affect subsequent data analysis and feature extraction. Therefore, noise cleaning is necessary. This unit uses a filtering algorithm to achieve this purpose.

[0051] Common filtering algorithms include mean filtering, median filtering, and Kalman filtering. Mean filtering calculates the average of data within a given window and replaces the data at the center of the window with this average to smooth the data. The formula is as follows:

[0052]

[0053] In the formula, This represents the data at the center of the filtered window, and 2N+1 represents the window size. This represents the k+i original data points within the window.

[0054] Median filtering sorts the data within a window and takes the median value as the center of the window, effectively removing impulse noise. Kalman filtering is an optimal estimation algorithm that uses two steps—prediction and update—to filter and estimate data based on the system's state and observation equations. This includes state prediction and state update, with the state prediction formula expressed as:

[0055]

[0056] The formula for state update is expressed as:

[0057]

[0058] In the formula, Let A be the predicted state at time k, and let A be the state transition matrix. Let B be the optimal estimated state at time k-1, and let B be the control input matrix. This is the control input at time k-1. For Kalman gain, Let H be the observation value at time k, and H be the observation matrix.

[0059] After filtering, preprocessed data is obtained.

[0060] The feature extraction unit is used to extract operating condition features from the preprocessed data using principal component analysis. These operating condition features include the changing trends of voltage and current, environmental features, and mechanical vibration features.

[0061] This unit uses principal component analysis (PCA) to extract operating condition features from preprocessed data. PCA is an unsupervised dimensionality reduction technique. Its basic idea is to project the original data onto a new set of orthogonal coordinate axes, maximizing the variance of the data on these axes.

[0062] The specific steps are as follows: First, the preprocessed data is standardized to eliminate the influence of different dimensions between variables. The standardization formula is:

[0063]

[0064] In the formula, For standardized data, This is the original data. Let be the mean of the j-th variable. Let be the standard deviation of the j-th variable.

[0065] Then calculate the covariance matrix of the data, expressed by the formula:

[0066]

[0067] In the formula, n is the sample size. , For standardized data, , Let be the mean of the i-th and j-th variables after standardization.

[0068] Next, we solve for the eigenvalues ​​and eigenvectors of the covariance matrix, i.e., satisfying... ,in Let v be the eigenvalue and v be the corresponding eigenvector.

[0069] Based on the magnitude of the eigenvalues, select the eigenvectors corresponding to the k largest eigenvalues ​​to form the projection matrix P; finally, project the preprocessed data onto the projection matrix to obtain the dimensionality-reduced principal components, calculated using the following formula:

[0070]

[0071] In the formula, Y represents the principal component data after dimensionality reduction, and X′ represents the standardized preprocessed data matrix.

[0072] These principal components are the extracted operating condition characteristics, including the trends of voltage and current changes, environmental characteristics, and mechanical vibration characteristics.

[0073] The multi-objective setting and decomposition module is used to set the control strategy for multiple objectives based on the working condition characteristic data and the particle swarm algorithm, and decompose the multiple objectives into sub-objectives.

[0074] The multi-objective setting and decomposition module includes an objective determination unit, an efficiency objective decomposition unit, a power quality objective decomposition unit, and an equipment reliability objective decomposition unit.

[0075] The objective determination unit is used to set multi-objective control strategies based on operating condition characteristic data and using particle swarm optimization. The multi-objective control strategies include maximizing efficiency, optimizing power quality, and extending equipment life.

[0076] The process of setting a multi-objective control strategy using the particle swarm optimization algorithm includes:

[0077] Initialize the particle swarm parameters, where each particle represents a combination of control strategies.

[0078] Define a multi-objective function, which calculates fitness based on efficiency, power quality, and equipment reliability indicators.

[0079] The particle position and velocity are iteratively updated, and the strategy combination is adjusted through global and local optimum searches to output the optimized control strategy.

[0080] Based on operating condition data, this unit employs a particle swarm optimization (PSO) algorithm to set a multi-objective control strategy. PSO is a swarm intelligence-based optimization algorithm that finds the optimal solution by simulating the foraging behavior of flocks of birds or schools of fish.

[0081] Initialize particle swarm parameters: Randomly initialize a swarm of particles, each representing a combination of control strategies. Each particle has its own position and velocity; position represents the specific value of the control strategy, and velocity represents the particle's direction and speed of movement in the search space.

[0082] Define a multi-objective function: The multi-objective function calculates fitness based on efficiency indicators, power quality indicators, and equipment reliability indicators.

[0083] Fitness values ​​are used to evaluate the quality of the control strategy combination represented by each particle. For example, an efficiency metric could be the ratio of a converter's output power to its input power, i.e., efficiency. ,in For output power, This refers to the input power.

[0084] Power quality indicators can include harmonic content, voltage fluctuations, etc., while equipment reliability indicators can include equipment failure rate, lifespan, etc. A multi-objective function can be expressed as:

[0085]

[0086] In the formula, X represents the combination of control strategies. For the efficiency objective function, Let the power quality objective function be... The objective function for equipment reliability is given, and the fitness is calculated comprehensively based on these objective functions.

[0087] Iterative updates to particle position and velocity: In each iteration, the global best particle is updated based on the fitness value of each particle, representing the particle with the best fitness value in the entire particle swarm, and the locally best particle, representing the position of each particle with the best fitness value historically. Then, the velocity and position of each particle are updated based on the information of the global best particle and the locally best particle. The particle velocity update formula is:

[0088]

[0089] The particle position update formula is:

[0090]

[0091] In the formula, Let be the velocity of the i-th particle in the d-th dimension at time t. For inertial weights, , As a learning factor, , A random number in the range [0, 1] Let be the optimal position of the i-th particle in the d-th dimension at time t. Let be the position of the i-th particle in the d-th dimension at time t. Let be the globally optimal position in the d-th dimension at time t.

[0092] Through continuous iterative searching and adjustment of strategy combinations, an optimized control strategy is ultimately output. Multi-objective control strategies include maximizing efficiency, optimizing power quality, and extending equipment lifespan.

[0093] The efficiency target decomposition unit is used to decompose the efficiency maximization target into power factor optimization and loss reduction sub-targets.

[0094] This unit decomposes the efficiency maximization objective into power factor optimization and loss reduction sub-objectives. Power factor is a crucial indicator of the efficiency of electrical equipment in utilizing electrical energy; a lower power factor indicates lower energy utilization and greater reactive power consumption. Therefore, optimizing the power factor can improve converter efficiency. Loss reduction, on the other hand, improves efficiency by minimizing various losses during converter operation, such as copper and iron losses.

[0095] The power quality target decomposition unit is used to decompose the power quality target into harmonic suppression and voltage stability sub-targets.

[0096] This unit decomposes the power quality objective into harmonic suppression and voltage stability sub-objectives. Harmonics refer to sinusoidal waves with frequencies that are integer multiples of the fundamental frequency. The presence of harmonics can affect the normal operation of the power system, leading to problems such as equipment overheating, increased losses, and malfunctions of protection devices. Therefore, it is necessary to suppress harmonic content and improve power quality. Voltage stability refers to ensuring that the amplitude and frequency of the converter output voltage fluctuate within a certain range to meet the load requirements.

[0097] The equipment reliability target decomposition unit is used to transform the equipment life extension target into temperature control and stress equalization sub-targets.

[0098] This unit translates the goal of extending equipment lifespan into sub-goals of temperature control and stress equalization. Temperature is a crucial factor affecting equipment lifespan; excessively high temperatures accelerate aging and damage, thus requiring temperature control within a reasonable range. Stress equalization refers to ensuring uniform stress distribution across all parts of the equipment during operation, preventing excessive localized stress that could lead to equipment damage.

[0099] The AI ​​collaborative control module is used to obtain control commands for the inverter switching frequency, adjust the inverter modulation method, and adjust the heat dissipation strategy based on sub-objectives using artificial intelligence.

[0100] The AI ​​collaborative control module includes an efficiency control unit, a power quality control unit, and an equipment lifespan control unit.

[0101] The efficiency control unit is used to build a control model based on a deep neural network, taking power factor optimization and loss reduction sub-objectives as inputs, and outputting control commands for the converter switching frequency.

[0102] The process of constructing a control model based on a deep neural network includes:

[0103] The design incorporates a deep neural network-based basic model, comprising an input layer, hidden layers, and an output layer. The input layer receives operating condition feature data and sub-objectives. The hidden layer consists of multiple fully connected layers using ReLU activation. The output layer generates switching frequency control commands.

[0104] Historical operational data is collected, and the mean squared error is used as the loss function to train the basic model. Model parameters that meet the preset accuracy are retained to obtain the control model.

[0105] The design incorporates a deep neural network-based basic model, consisting of an input layer, hidden layers, and an output layer. The input layer receives condition feature data and sub-objectives, using this data as input to the model. The hidden layer comprises multiple fully connected layers, where neurons in each layer are connected to all neurons in the previous layer. The activation function used is ReLU (Modified Linear Unit), which introduces non-linearity and enhances the model's expressive power. The output layer generates switching frequency control commands, outputting appropriate commands based on the input data and the model's learning results.

[0106] Collect historical operating data: Collect historical operating data of the converter, including operating condition characteristics, sub-objectives, and corresponding switching frequency control commands. This data is used to train the base model.

[0107] Training the base model: Mean squared error is used as the loss function. The model parameters are continuously adjusted to minimize the loss function. During training, the data is divided into a training set and a validation set. The training set is used to train the model, and the validation set is used to evaluate the model's performance. When the model's performance on the validation set meets the preset accuracy, the model parameters at this point are retained, resulting in the control model.

[0108] The power quality control unit is used to adjust the modulation mode of the converter based on the sub-targets of harmonic suppression and voltage stability using a convolutional neural network.

[0109] The process of adjusting the modulation method of a converter using a convolutional neural network includes:

[0110] A fuzzy logic rule base is constructed based on the harmonic suppression target.

[0111] Based on a fuzzy logic rule base, a convolutional neural network is trained to predict the modulation mode according to the voltage stability sub-objective.

[0112] The pulse width modulation signal parameter adjustment command is output according to the modulation method.

[0113] Based on the adjustment instructions, the model parameters are optimized through a feedback loop.

[0114] A fuzzy logic rule base is constructed based on the harmonic suppression objective: Fuzzy logic is a method for handling uncertainty and fuzzy information. Based on the harmonic suppression objective, a series of fuzzy rules are formulated; for example, if the harmonic content is high, the complexity of the modulation method is increased. These rules constitute the fuzzy logic rule base.

[0115] Based on a fuzzy logic rule base, a convolutional neural network (CNN) is trained to predict modulation schemes according to the voltage stability sub-target. A CNN is a type of neural network specifically designed for processing data with a grid structure, and it has wide applications in image recognition, signal processing, and other fields. Using the voltage stability sub-target and information from the fuzzy logic rule base as inputs, the CNN model is trained to predict appropriate modulation schemes.

[0116] Based on the modulation method, output pulse width modulation signal parameter adjustment instructions: Based on the predicted modulation method, determine the parameters of the pulse width modulation signal, such as pulse width and frequency, and output adjustment instructions.

[0117] Based on the adjustment instructions, the model parameters are optimized through a feedback loop: the adjustment instructions are applied to the converter, the actual operating data of the converter is collected, the actual data is compared with the target data, and the parameters of the convolutional neural network model are adjusted according to the comparison results. Through continuous feedback loops, the performance of the model is optimized.

[0118] The equipment life control unit is used to adjust the heat dissipation strategy using an optimized control algorithm based on the sub-objectives of temperature control and stress equalization.

[0119] The process of adjusting the heat dissipation strategy using optimized control algorithms includes:

[0120] Build a temperature prediction model by inputting device thermal distribution data.

[0121] Define an optimization function to minimize temperature fluctuations and stress differences.

[0122] The model predictive control algorithm is used to generate an optimized scheme for heat dissipation parameters.

[0123] Perform real-time adjustments to cooling fan speed and radiator layout.

[0124] Building a temperature prediction model: Input the device's thermal distribution data, which can be collected by temperature sensors. The temperature prediction model can be built using machine learning or deep learning algorithms, such as recurrent neural networks or long short-term memory networks, to predict future temperature changes of the device based on historical temperature data and the device's operating status.

[0125] Define the optimization function: The objective of the optimization function is to minimize temperature fluctuations and stress differences, expressed by the formula:

[0126]

[0127] In the formula, , These are the weighting coefficients. For the temperature at each point, The average temperature. For the stress at each point, The average stress is given.

[0128] Excessive temperature fluctuations can lead to thermal fatigue of equipment, while excessive stress differences can cause localized damage to the equipment. Therefore, it is necessary to balance these two factors through optimization functions.

[0129] Model predictive control (MMCC) is a model-based optimization control algorithm that solves an optimization problem at each sampling time based on the system's predictive model and optimization function, obtaining a control sequence for a future period. Applying MMCC, based on the temperature prediction model and optimization function, we can generate optimized heat dissipation parameter schemes, including the speed of the cooling fan and the layout of the heat sink.

[0130] Perform real-time adjustments to cooling fan speed and radiator layout: Based on the generated heat dissipation parameter optimization scheme, adjust the cooling fan speed and radiator layout in real time to ensure that the equipment temperature is within a reasonable range and extend the equipment life.

[0131] The monitoring and evaluation module is used to collect converter operating data in real time and compare it with sub-targets to evaluate the degree of target completion.

[0132] The monitoring and evaluation module includes a real-time data acquisition unit, a target comparison unit, a trend prediction unit, and an evaluation feedback unit.

[0133] The real-time data acquisition unit is used to acquire converter operating data in real time, including electrical performance data, efficiency-related data, and control command execution data.

[0134] This unit collects real-time operating data from the converter, including electrical performance data such as voltage, current, and power; efficiency-related data such as power factor and efficiency; and control command execution data, such as the execution status of switching frequency control commands and the adjustment status of modulation methods. This data is collected through various sensors and monitoring devices and transmitted to the data processing equipment.

[0135] The target comparison unit is used to compare the running data with multiple targets and obtain the target difference value.

[0136] The collected operational data is compared with multiple objectives to calculate the target difference value. For example, the actual efficiency value is compared with the efficiency maximization objective to calculate the efficiency difference value; the actual harmonic content is compared with the harmonic suppression objective to calculate the harmonic difference value, and so on. Through comparison, the gap between the actual operation of the converter and the target can be intuitively understood.

[0137] The trend prediction unit is used to build a trend prediction model based on time series analysis. It takes the target difference value as input and outputs the predicted change trend of multiple targets.

[0138] Trend forecasting models based on time series analysis are constructed. Commonly used time series analysis models include the autoregressive integral moving average model and the seasonal autoregressive integral moving average model. Using the target variance value as input, the model analyzes the changing patterns of historical target variance value data to predict the future trends of multiple targets.

[0139] The assessment feedback unit is used to generate assessment reports based on predicted trends and to provide feedback.

[0140] An evaluation report is generated based on the predicted trends. This report includes information such as the converter's current operating status, target achievement, and future development trends. The evaluation report is then fed back to operators and the control system to allow for timely adjustments to control strategies, ensuring stable converter operation and the achievement of multiple objectives.

[0141] In summary, this embodiment provides a multi-objective collaborative AI converter control system based on complex operating conditions. Through data collection, processing, and feature extraction units, it can accurately collect operating condition data such as voltage, current, and temperature of the converter, effectively remove noise interference, and extract key operating condition features. The multi-objective setting and decomposition module scientifically sets multi-objective control strategies based on the operating condition feature data using a particle swarm optimization algorithm, and decomposes them into sub-objectives. This not only comprehensively considers multiple objectives such as maximizing efficiency, optimizing power quality, and extending equipment lifespan, but also refines macro-objectives into specific and operable sub-objectives, improving the targeting and executability of the control strategy. The AI ​​collaborative control module utilizes artificial intelligence technology to perform precise control for different sub-objectives. The efficiency control unit optimizes the switching frequency through a deep neural network, the power quality control unit adjusts the modulation method using a convolutional neural network, and the equipment lifespan control unit adjusts the heat dissipation strategy using an optimized control algorithm. This significantly improves the converter's operating efficiency, enhances power quality, effectively extends equipment lifespan, and reduces equipment maintenance costs. The monitoring and evaluation module collects the converter's operating data in real time, compares it with the sub-objectives, and predicts the changing trends of multiple objectives through a trend prediction model. The generated assessment report can provide timely feedback on the system's operating status, helping operators to identify problems and adjust control strategies in a timely manner, ensuring that the converter is always in a stable and efficient operating state.

[0142] Based on the same general inventive concept, this invention also protects a multi-objective cooperative AI converter control method based on complex operating conditions. The multi-objective cooperative AI converter control method based on complex operating conditions provided by this invention will be described below. The multi-objective cooperative AI converter control method based on complex operating conditions described below can be referred to in correspondence with the multi-objective cooperative AI converter control system based on complex operating conditions described above.

[0143] Figure 2 This is a flowchart illustrating the multi-objective collaborative AI converter control method based on complex operating conditions provided in this embodiment of the invention.

[0144] like Figure 2 As shown, the multi-objective cooperative AI converter control method based on complex operating conditions includes:

[0145] Collect the operating condition data of the converter and process the operating condition data to obtain the operating condition characteristic data.

[0146] Based on the operating condition characteristic data, a multi-objective control strategy is set using the particle swarm optimization algorithm, and the multi-objective is decomposed into sub-objectives.

[0147] Based on the sub-objectives, artificial intelligence is used to obtain control commands for the inverter's switching frequency, adjust the inverter's modulation method, and adjust the heat dissipation strategy.

[0148] Real-time acquisition of converter operating data is used to compare with sub-targets and assess the degree of target achievement.

[0149] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0150] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-objective cooperative AI converter control system based on complex operating conditions, characterized in that, include: The complex operating condition sensing module is used to collect the operating condition data of the converter and process the operating condition data to obtain operating condition feature data. The multi-objective setting and decomposition module is used to set a control strategy for multiple objectives based on the working condition characteristic data using a particle swarm optimization algorithm, and decompose the multiple objectives into sub-objectives; The multi-objective setting and decomposition module includes an objective determination unit, an efficiency objective decomposition unit, a power quality objective decomposition unit, and an equipment reliability objective decomposition unit. The objective determination unit is used to set a multi-objective control strategy based on the operating condition characteristic data using a particle swarm optimization algorithm. The multi-objective control strategy includes maximizing efficiency, optimizing power quality, and extending equipment lifespan. The efficiency objective decomposition unit decomposes the efficiency maximization objective into power factor optimization and loss reduction sub-objectives. The power quality objective decomposition unit decomposes the power quality objective into harmonic suppression and voltage stability sub-objectives. The equipment reliability objective decomposition unit transforms the equipment lifespan extension objective into temperature control and stress equalization sub-objectives. The AI ​​collaborative control module is used to obtain control commands for the inverter switching frequency, adjust the inverter modulation mode, and adjust the heat dissipation strategy based on the sub-objectives using artificial intelligence. The AI ​​collaborative control module includes an efficiency control unit, a power quality control unit, and an equipment life control unit. The efficiency control unit is used to construct a control model based on a deep neural network, taking the power factor optimization and loss reduction sub-objectives as inputs, and outputting control commands for the converter switching frequency. The power quality control unit is used to adjust the converter modulation mode using a convolutional neural network according to the harmonic suppression and voltage stability sub-objectives. The equipment life control unit is used to adjust the heat dissipation strategy using an optimized control algorithm according to the temperature control and stress equalization sub-objectives. The monitoring and evaluation module is used to collect converter operating data in real time and compare it with the sub-targets to evaluate the target completion rate.

2. The multi-objective cooperative AI converter control system based on complex operating conditions according to claim 1, characterized in that, The complex operating condition sensing module includes a data collection unit, a data processing unit, and a feature extraction unit. The data collection unit is used to collect voltage, current, temperature, humidity, and mechanical vibration of the converter through sensors and perform signal conversion to obtain operating condition data. The data processing unit is used to clean the noise of the operating condition data using a filtering algorithm to obtain preprocessed data. The feature extraction unit is used to extract operating condition features from the preprocessed data using principal component analysis, and the operating condition features include the changing trends of voltage and current, environmental characteristics, and mechanical vibration characteristics.

3. The multi-objective cooperative AI converter control system based on complex operating conditions according to claim 1, characterized in that, The process of setting a multi-objective control strategy using the particle swarm optimization algorithm includes: Initialize the particle swarm parameters, where each particle represents a combination of control strategies; Define a multi-objective function, which calculates fitness based on efficiency indicators, power quality indicators, and equipment reliability indicators; The particle position and velocity are iteratively updated, and the strategy combination is adjusted through global and local optimum searches to output the optimized control strategy.

4. The multi-objective cooperative AI converter control system based on complex operating conditions according to claim 1, characterized in that, The process of constructing a control model based on a deep neural network includes: The design incorporates a deep neural network-based basic model, comprising an input layer, hidden layers, and an output layer. The input layer receives operating condition feature data and sub-objectives. The hidden layer comprises multiple fully connected layers, using ReLU activation. The output layer generates switching frequency control commands. Historical operational data is collected, and the base model is trained using mean squared error as the loss function. Model parameters that meet the preset accuracy are retained to obtain the control model.

5. The multi-objective cooperative AI converter control system based on complex operating conditions according to claim 1, characterized in that, The process of adjusting the modulation method of a converter using a convolutional neural network includes: A fuzzy logic rule base is constructed based on the harmonic suppression target; Based on the aforementioned fuzzy logic rule base, a convolutional neural network is trained to predict the modulation mode according to the voltage stability sub-target. According to the modulation method, output pulse width modulation signal parameter adjustment instructions; Based on the adjustment instructions, the model parameters are optimized through a feedback loop.

6. The multi-objective cooperative AI converter control system based on complex operating conditions according to claim 1, characterized in that, The process of adjusting the heat dissipation strategy using optimized control algorithms includes: Build a temperature prediction model by inputting device thermal distribution data; Define an optimization function to minimize temperature fluctuations and stress differences; The model predictive control algorithm is used to generate an optimized solution for heat dissipation parameters. Perform real-time adjustments to cooling fan speed and radiator layout.

7. The multi-objective cooperative AI converter control system based on complex operating conditions according to claim 1, characterized in that, The monitoring and evaluation module includes a real-time data acquisition unit, a target comparison unit, a trend prediction unit, and an evaluation feedback unit. The real-time data acquisition unit is used to acquire converter operating data in real time, including electrical performance data, efficiency-related data, and control command execution data. The target comparison unit is used to compare the running data with the multiple targets and obtain the target difference value; the trend prediction unit is used to construct a trend prediction model based on time series analysis, taking the target difference value as input and outputting the predicted change trend of the multiple targets. The evaluation feedback unit is used to generate an evaluation report based on the predicted change trend and provide feedback.

8. A multi-objective cooperative AI converter control method based on complex operating conditions, wherein the multi-objective cooperative AI converter control method based on complex operating conditions is used to implement the multi-objective cooperative AI converter control system based on complex operating conditions as described in any one of claims 1 to 7, characterized in that, The methods include: Collect the operating condition data of the converter and process the operating condition data to obtain operating condition characteristic data; Based on the operating condition characteristic data, a multi-objective control strategy is set using the particle swarm optimization algorithm, and the multi-objective is decomposed into sub-objectives; Based on the sub-objectives, artificial intelligence is used to obtain control commands for the inverter switching frequency, adjust the inverter modulation method, and adjust the heat dissipation strategy. Real-time converter operation data is collected and compared with the sub-targets to assess the target completion rate.

Images