Multi-target optimization plasma melt-blown powder spheroidizing equipment parameter analysis method and device and storage medium

Abstract

The invention discloses a multi-target optimization plasma melt-blown powder spheroidizing equipment parameter analysis method and device and a storage medium, and relates to the field of mechanical automation control. The specific scheme comprises the steps of obtaining parameter data; preprocessing the parameter data according to a preset algorithm; performing feature extraction and feature construction on the preprocessed parameter data; constructing a training data set and a verification data set by using the data subjected to feature extraction and feature construction; constructing an unsupervised learning model by using the training data set, and evaluating and optimizing the unsupervised learning model by using the verification data set; determining a multi-target optimization algorithm; and inputting the data subjected to feature extraction into a multi-target optimization algorithm, searching an optimal balance among different targets, and confirming the optimal value range of a plurality of parameter data. According to the invention, the production process is optimized, the cost is reduced, and the control capability of the production process is improved.

Description

Technical Field

[0001] The present invention relates to the field of mechanical automation control technology, and in particular to a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method, device and storage medium. Background Art

[0002] Plasma spheroidization is a technology that uses the characteristics of high-temperature plasma to process materials. In this process, powder materials are fed into a plasma flame, and the high-temperature plasma quickly heats the powder particles to a molten state. In the molten state, the surface tension causes the molten particles to automatically shrink into a spherical shape to reduce surface energy. Subsequently, through rapid cooling, the molten spherical particles solidify into spherical powder particles. Plasma spheroidization technology can produce spherical powders with excellent fluidity and sintering properties, and is widely used in powder metallurgy, 3D printing, new material research and development and other fields.

[0003] In the preparation process of plasma spheroidization, precise control of equipment operating parameters is crucial. These parameters include powder feeding flow, speed, temperature, pressure and argon content. In the prior art, the following methods are usually adopted to determine the normal value range of equipment parameters, each of which has its inherent shortcomings or deficiencies: 1. Rule of thumb: relying on the experience and intuition of technical experts to set parameters. This method is highly subjective, difficult to adapt to changing production conditions, and difficult to ensure consistency and repeatability. 2. Single parameter adjustment: adjusting each parameter one by one and observing the effect, ignoring the interdependence between parameters and the possibility of overall optimization, resulting in the inability to find the global optimal solution. 3. Static numerical analysis: setting the parameter range based on equipment specifications and theoretical models. This method ignores factors such as dynamic changes and equipment aging in the actual production process. 4. Linear model prediction: using a simple linear model to predict the impact of parameter changes. This method cannot accurately capture complex nonlinear relationships and interactions. 5. Heuristic algorithm: although it can quickly provide a solution, it may lack consideration of the overall problem, resulting in the solution found may be only a local optimum rather than a global optimum. 6. Traditional statistical methods: such as analysis of variance (ANOVA). These methods may have computational limitations when processing large amounts of data and are not sensitive enough to non-normally distributed data. Therefore, a more comprehensive, accurate, and automated parameter analysis method is needed. Summary of the invention

[0004] The present invention provides a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method, device and storage medium, which optimize the production process, reduce costs, and improve the control ability of the production process.

[0005] In order to achieve the above object, the present invention adopts the following technical scheme: In a first aspect, the present invention provides a multi-objective optimization method for analyzing parameters of a plasma melt-blown powder spheroidization device, the method comprising: Get parameter data; Preprocessing the parameter data according to a preset algorithm; Perform feature extraction and feature construction on the preprocessed parameter data; Use the data after feature extraction and feature construction to construct training data sets and validation data sets; Using the training data set to construct an unsupervised learning model, and using the validation data set to evaluate and optimize the unsupervised learning model; Determine the multi-objective optimization algorithm; The data after feature extraction is input into the multi-objective optimization algorithm to find the optimal balance between different objectives and confirm the most appropriate value range of several parameter data.

[0006] The embodiments provided by the present invention improve production efficiency: by determining the optimal operating parameters, production efficiency is improved to ensure that the equipment operates in the optimal state. Product quality is guaranteed: the quality of the finished products produced during the production process is ensured to be stable and meet industry standards and customer needs. Energy consumption is reduced: parameter settings are optimized to reduce unnecessary energy waste and achieve the dual goals of cost-effectiveness and environmental friendliness. Automated control is achieved: through algorithm analysis, data support is provided for the automated control system of the equipment to achieve intelligent production. Multi-objective optimization: considering that there may be multiple conflicting goals in the production process, such as efficiency and quality, the multi-objective optimization algorithm can balance these goals and find the best solution. Abnormality detection and prediction: by analyzing the normal value range, the system's ability to detect abnormal conditions is enhanced, and potential production problems are predicted in advance.

[0007] In a possible implementation, the unsupervised learning model includes: a clustering analysis algorithm and an autoencoder algorithm.

[0008] In a possible implementation, the cluster analysis algorithm includes: Data preprocessing: clean the data, remove missing values ​​and outliers, and standardize or normalize the data; Select clustering algorithm: Determine the clustering algorithm based on data characteristics; Determine the number of clusters: Use any of the methods such as elbow rule and silhouette coefficient to determine the optimal number of clusters; Perform clustering: Perform cluster analysis on the data set based on the clustering algorithm; Evaluate clustering results: Use the silhouette coefficient or Davies-Bouldin index to evaluate clustering effect; Interpret clustering results: Interpret data patterns based on clustering results and identify potential structures in the data.

[0009] In one possible implementation, the autoencoder algorithm includes: Data preprocessing: Data cleaning and standardization through homocluster analysis; Design the autoencoder network structure: determine the size of the input layer; Select activation function: Determine the activation function of the hidden layer of the autoencoder; Training the autoencoder: Use the dataset to train the autoencoder and optimize the network weights through the back-propagation algorithm; Feature learning: Using low-dimensional representations to learn hidden layers and capture the intrinsic structure of the data; Anomaly detection or dimensionality reduction: Anomaly detection is performed based on the reconstruction error of the autoencoder, and the data representation learned by the autoencoder is used as the reduced dimension feature input into other machine learning models; Evaluation and Interpretation: Evaluate performance and interpret learned features to understand the structure of the data.

[0010] In a possible implementation, determining a multi-objective optimization algorithm includes: Objective function definition: define three optimization objectives; Feature value selection: determine the key feature values ​​that affect the target; Eigenvalue modeling: Establish a relationship model between eigenvalues ​​and optimization targets, and use historical data to train the model; Constraint setting: According to process requirements, equipment characteristics and unsupervised learning results, set the operating range and constraints of each feature value; Pareto Optimality Definition: In multi-objective optimization, define the Pareto superiority criterion; Algorithm selection: Determine at least one algorithm from among multi-objective particle swarm optimization, non-dominated sorting genetic algorithm II and multi-objective simulated annealing; Algorithm framework design: Design the algorithm framework, including at least: initialization, iteration process, selection operation, crossover and mutation operation, and Pareto frontier update; Performance evaluation: Design performance evaluation criteria; Algorithm implementation: Implement the algorithm and code it using the Java programming language and tools; Solve the Pareto optimal solution set: run the algorithm to generate the Pareto optimal solution set; Result analysis: Analyze the Pareto optimal solution set, identify feasible solutions, and select them based on process experience and evaluation criteria such as comparison of convergence accuracy, solution set diversity, calculation time and stability; Experimental verification: Combined with the verification set, the effectiveness of the algorithm is verified.

[0011] In a second aspect, the present invention provides a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device, comprising: Parameter acquisition module, used to obtain parameter data; A preprocessing module, used for preprocessing the parameter data according to a preset algorithm; A feature extraction module is used to extract and construct features from the preprocessed parameter data; Unsupervised learning module, used to construct training data set and validation data set from the data after feature extraction and feature construction; It is also used to construct an unsupervised learning model using the training data set, and to evaluate and optimize the unsupervised learning model using the verification data set; Multi-objective optimization module, used to determine the multi-objective optimization algorithm; It is also used to input the feature-extracted data into a multi-objective optimization algorithm to find the optimal balance between different objectives and confirm the most appropriate value range for several parameter data.

[0012] In a third aspect, the present invention provides a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device comprising: a processor and a memory. The memory is used to store computer program code, and the computer program code includes computer instructions. When the processor executes the computer instructions, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device performs the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method as described in the first aspect and any possible implementation thereof.

[0013] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon. When the computer instructions are run on a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device performs a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method such as the first aspect or any one of the possible implementation methods of the first aspect.

[0014] In a fifth aspect, the present invention provides a computer program product, which includes computer instructions. When the computer instructions are run on a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device performs a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method as in the first aspect and any possible implementation method thereof.

[0015] For the specific description of the second to fifth aspects and their various implementations in the present invention, reference can be made to the detailed description in the first aspect and its various implementations; and for the beneficial effects of the second to fifth aspects and their various implementations, reference can be made to the analysis of the beneficial effects in the first aspect and its various implementations, which will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of the structure of a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device provided by an embodiment of the present invention; Figure 2 One of the flow charts of a multi-objective optimization method for analyzing parameters of a plasma melt-blown powder spheroidization equipment provided by an embodiment of the present invention; Figure 3 A second flowchart of a multi-objective optimization method for analyzing parameters of a plasma melt-blown powder spheroidization equipment provided by an embodiment of the present invention; Figure 4 A flow chart of a cluster analysis algorithm provided by an embodiment of the present invention; Figure 5 A flowchart of an autoencoder algorithm provided by an embodiment of the present invention; Figure 6 A flowchart of a multi-objective optimization algorithm provided by an embodiment of the present invention; Figure 7 A schematic diagram of the composition of a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present disclosure, unless otherwise specified, "multiple" means two or more. In addition, the use of "based on" or "according to" means openness and inclusiveness, because the process, steps, calculations or other actions "based on" or "according to" one or more of the conditions or values ​​may be based on additional conditions or values ​​beyond the described values ​​in practice.

[0019] The execution subject of the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method provided in the embodiment of the present invention is a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device. The multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device can be a computer device, which can be an electronic device or a server, or a central processing unit (CPU) in the electronic device or server.

[0020] Figure 1 A schematic diagram of the composition of a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device provided by an embodiment of the present invention. Figure 1 As shown, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device may include: at least one processor 11, a memory 12, a communication interface 13 and a communication bus 14.

[0021] Among them, the processor 11 is the control center of the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device, which can be a CPU, a microprocessing unit, or one or more integrated circuits for controlling the execution of the program of the embodiment of the present invention.

[0022] As an embodiment, the processor 11 may include one or more CPUs, such as Figure 1 In addition, as an embodiment, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device may include multiple processors, such as Figure 1 The processor 11 and the processor 15 shown in FIG. Each of these processors can be a single-core processor (Single-CPU) or a multi-core processor (Multi-CPU).

[0023] The memory 12 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 12 may exist independently and be connected to the processor 11 via the communication bus 14. The memory 12 may also be integrated with the processor 11.

[0024] In a specific implementation, the memory 12 is used to store the data in the present invention and to execute the software program of the present invention. The processor 11 can execute various functions of the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device by running or executing the software program stored in the memory 12 and calling the data stored in the memory 12.

[0025] The communication interface 13 uses any transceiver or other device for communicating with other devices or communication networks, such as radio access networks (RAN), wireless local area networks (WLAN), etc. The communication interface 13 may include a receiving unit to implement a receiving function, and a sending unit to implement a sending function.

[0026] The communication bus 14 may include a pathway for transmitting information between the above-mentioned components.

[0027] It should be pointed out that Figure 1 The structure shown in the figure does not constitute a limitation on the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device, except Figure 1 In addition to the components shown, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0028] Based on the above introduction to the structure of the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device, the embodiment of the present invention provides a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method. Figure 2 As shown, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method may include the following steps S1-S7: S1. Obtain parameter data; It is understandable that by installing and calibrating the necessary sensors, key parameters such as the powder delivery flow, speed, temperature, pressure and argon content of the equipment are monitored in real time. Collect the operating data of the equipment under different operating conditions to ensure the comprehensiveness and diversity of the data. The purpose of this step is to collect the historical data of the equipment production parameters of the equipment under different production conditions. Before using the historical production data for model training, it is necessary to select features that have a significant impact on the output and quality of qualified spherical powder from the historical production data. Among them, the equipment production parameters are used to characterize the equipment parameters in the preparation process of titanium zinc spherical powder. The equipment production parameters include pressure, temperature, material flow, reaction time, etc. The corresponding production result parameters include the output, qualified rate and energy consumption of the preparation of spherical powder. The above data will serve as the basis for machine learning model training.

[0029] S2. Preprocess the parameter data according to a preset algorithm; It is understandable that the parameter data obtained in S1 needs to be cleaned to remove outliers and noise. It is also necessary to perform necessary data conversion and standardization to eliminate the influence of different dimensions.

[0030] For example, the parameter data obtained in S1 may have different frequencies, and in this step, it needs to be processed into data of the same frequency. In this case, the processing methods of mean resampling, interpolation and aggregation can be used. The details are as follows: Mean resampling: Some original data has high frequency, while the target frequency is low frequency. The data frequency can be reduced by calculating the average value within the time window.

[0031]

[0032] in, is the value at time point t after resampling, is the value of the original data at time point t+i, n is the size of the time window.

[0033] Interpolation: For some relatively stable parameters, linear interpolation can be used to transform the parameters, that is, to estimate the value of the new data point by connecting two known data points with a straight line.

[0034] Aggregation: If there are enough observations in the specified time period, these observations can be added together to reduce the data frequency; or, if the amount of data in the specified time period is small, the maximum value in the time period is selected as the aggregation result.

[0035] S3, extracting and constructing features from the preprocessed parameter data; It is understandable that useful features, such as statistics, time series features, etc., need to be extracted from the raw data; through feature selection methods, key features that have a significant impact on production result parameters are screened out. The goal of this step is to extract features that can effectively represent data characteristics and help improve model performance. Through feature selection and dimensionality reduction, redundant and irrelevant features can be removed, the complexity of the model can be reduced, and the interpretability and predictive power of the model can be improved.

[0036] Exemplarily, feature extraction includes: calculating descriptive statistics such as mean, median, standard deviation, minimum, maximum, range, skewness, kurtosis, etc. from the original data, trying to discover patterns, trends and interesting phenomena.

[0037] Feature construction involves combining existing features to create new features, such as response time.

[0038] S4, constructing a training data set and a validation data set using the data after feature extraction and feature construction; It can be understood that the data after feature extraction and feature construction is split into a training data set and a validation data set in a ratio of 8:2. Among them, the training data set is used to train the machine learning model, while the validation data set is used to evaluate the performance of the model and optimize it.

[0039] S5. Use the training data set to construct an unsupervised learning model, and use the verification data set to evaluate and optimize the unsupervised learning model; It can be understood that unsupervised learning algorithms are first applied to perform pattern recognition and structure learning on the data. Then, the unsupervised learning results are used to identify the conditions for normal operation of the equipment and potential abnormal patterns.

[0040] Furthermore, the unsupervised learning model includes: a clustering analysis algorithm and an autoencoder algorithm.

[0041] Among them, refer to Figure 4 As shown, the clustering analysis algorithm includes: S511. Data preprocessing: clean the data, remove missing values ​​and outliers, and standardize or normalize the data; the purpose is to eliminate the influence of different dimensions.

[0042] S512, select a clustering algorithm: determine a clustering algorithm according to data characteristics; illustratively, it may be K-means, hierarchical clustering, DBSCAN, etc.

[0043] S513, determining the number of clusters: using any method of the elbow rule and the silhouette coefficient to determine the optimal number of clusters; S514, executing clustering: performing cluster analysis on the data set according to a clustering algorithm.

[0044] S515. Evaluate clustering results: Use the silhouette coefficient or Davies-Bouldin index to evaluate the clustering effect. The purpose is to visualize the clustering results and check whether the clustering is reasonable.

[0045] S516. Explain clustering results: Explain the data pattern based on the clustering results and identify the potential structure in the data.

[0046] Among them, refer to Figure 5 As shown, the autoencoder algorithm includes: S521, data preprocessing: data cleaning and standardization through homocluster analysis; S522. Design the autoencoder network structure: determine the size of the input layer; illustratively, the input layer is usually the same as the number of data features, for example, 6, and the size and number of hidden layers are designed at the same time. Here, the hidden layer size is 2, the first layer has 12 neurons, and the second layer has 6 neurons.

[0047] S523, select activation function: determine the activation function of the hidden layer of the autoencoder; ReLU can be used.

[0048] S524, training the autoencoder: using the data set to train the autoencoder, and optimizing the network weights through the back propagation algorithm; S525, Feature Learning: Use low-dimensional representation of hidden layers to learn data and capture the intrinsic structure of the data; S526, anomaly detection or dimensionality reduction: perform anomaly detection based on the reconstruction error of the autoencoder, and use the data representation learned by the autoencoder as a feature after dimensionality reduction to input into other machine learning models; S527. Evaluation and Interpretation: Evaluate performance and interpret learned features to understand the data structure.

[0049] S6. Determine a multi-objective optimization algorithm; It is understandable that three optimization goals are first defined based on process requirements and equipment characteristics. For example, the optimization goals are: improving production efficiency, ensuring product quality, and reducing energy consumption. Then, a multi-objective optimization algorithm framework is designed to integrate different goals and solve the Pareto optimal solution set.

[0050] Specifically, refer to Figure 6 As shown, a multi-objective optimization algorithm is determined, including: S61. Objective function definition: define three optimization objectives; exemplary ones include: optimal finished product qualification rate, maximum output, and minimization of energy consumption per unit of raw material.

[0051] S62. Characteristic value selection: determine the key characteristic values ​​that affect the target; for example, powder feeding flow rate, speed, temperature, pressure, argon content and reaction time.

[0052] S63, Eigenvalue modeling: Establish a relationship model between eigenvalues ​​and optimization objectives, and use historical data to train the model; S64, constraint setting: according to the process requirements, equipment characteristics and the results of unsupervised learning, set the operating range and constraints of each characteristic value; S65. Definition of Pareto optimality: In multi-objective optimization, define the Pareto superiority criterion; the meaning of the Pareto superiority criterion is that a solution is considered to be Pareto optimal if no other solution is better in all objectives.

[0053] S66. Algorithm selection: determine at least one algorithm from among multi-objective particle swarm optimization (MOPSO), non-dominated sorting genetic algorithm II (NSGA-II), and multi-objective simulated annealing (MOGSA); S67, Algorithm framework design: Design the algorithm framework, including at least: initialization, iteration process, selection operation, crossover and mutation operation, Pareto frontier update; S68, Performance evaluation: Design performance evaluation criteria; S69, Algorithm Implementation: Implement the algorithm and code it using the Java programming language and tools; S610, solving the Pareto optimal solution set: running the algorithm to generate the Pareto optimal solution set; these solutions provide different trade-offs between multiple objectives.

[0054] S611, Result Analysis: Analyze the Pareto optimal solution set, identify feasible solutions, and select them based on process experience and evaluation criteria such as comparative convergence accuracy, solution set diversity, calculation time and stability; S612. Experimental verification: Verify the effectiveness of the algorithm by combining the verification set.

[0055] Further, the multi-objective particle swarm optimization (MOPSO) in step S66 can be implemented by the following steps: S6611. Objective function definition: Define the objective functions of production efficiency, product quality and energy consumption. For example: Production efficiency objective function: maximize the product output per unit time; Product quality objective function: minimize the product defect rate or proportion of defective products; Energy consumption objective function: minimize the energy consumption per unit product production.

[0056] S6612. Eigenvalue selection: Determine the eigenvalues ​​that affect the objective function: powder delivery flow rate, speed, temperature, pressure, argon content and reaction time.

[0057] S6613. Search space definition: Define a suitable search range for each eigenvalue based on process, equipment characteristics and unsupervised learning results.

[0058] S6614, Initial particle swarm generation: Randomly generate an initial particle swarm, each particle represents a possible solution and contains all eigenvalues.

[0059] S6615, Fitness evaluation: Calculate the values ​​of three objective functions for each particle and evaluate its fitness.

[0060] S6616. Leader particle determination: According to the fitness evaluation result, a non-dominated particle is selected as the leader particle.

[0061] S6617, Particle velocity and position update: Update the velocity and position of each particle, taking into account the personal best, the leader particle position and random perturbations.

[0062] S6618, Diversity and leader particle update: Ensure the diversity of the particle swarm to avoid premature convergence; update the leader particle to guide the particle swarm to approach the Pareto frontier.

[0063] S6619, Iteration termination condition: Stop iteration when the number of iterations or fitness no longer improves significantly.

[0064] S66110, Pareto optimal solution set extraction: Extract the Pareto optimal solution set from the leading particles.

[0065] Further, the non-dominated sorting genetic algorithm II (NSGA-II) in step S66 can be implemented by the following steps: S6621. Initial population generation: Randomly generate the initial population, where each individual represents a potential solution.

[0066] S6622. Fitness evaluation: Calculate the values ​​of the three objective functions for each individual in the initial population and evaluate its fitness.

[0067] S6623, Non-dominated sorting: Perform non-dominated sorting on the individuals in the population and divide the individuals into different frontiers (Pareto frontiers).

[0068] S6624. Crowding distance calculation: Calculate the crowding distance of each individual in the non-dominated front to measure the distribution density of solutions around the individual.

[0069] S6625, selection operation: select excellent individuals based on non-dominated sorting and crowding distance for subsequent crossover and mutation operations.

[0070] S6626, Crossover and mutation: Perform a crossover operation on selected individuals to produce new offspring, while applying a mutation operation to introduce new genetic diversity.

[0071] S6627. Formation of a new generation of population: merge the parent and offspring generations, and select the new generation of population based on non-dominated sorting and crowding distance.

[0072] S6628, Iteration termination condition: Set the number of iterations or specific optimization stopping conditions, such as when fitness no longer improves significantly.

[0073] S6629, Pareto optimal solution set extraction: Extract the Pareto optimal solution set from the final population, which represents the best trade-off between different objectives.

[0074] Further, the multi-objective simulated annealing (MOGSA) in step S66 can be implemented by the following steps: S6631, Initial solution generation: Randomly generate an initial solution, representing a possible configuration of process parameters.

[0075] S6632. Energy function construction: Construct an energy function (or the negative value of the objective function) to evaluate the quality of the current solution.

[0076] S6633. Temperature Scheduling Strategy: Design temperature reduction strategies, such as geometric or exponential reduction, to control the search process.

[0077] S6634, Neighborhood Search: Define the neighborhood structure, i.e. how to generate new candidate solutions from the current solution. This may involve minor adjustments to process parameters.

[0078] S6635, Acceptance Criteria: Implement the acceptance criteria to decide whether to accept a new solution. In MOGSA, solutions that are better than the current solution are usually accepted, or worse solutions are accepted with a certain probability to avoid falling into a local optimum.

[0079] S6636, update operation: If the new solution is accepted, update the current solution and adjust the temperature according to the temperature reduction strategy.

[0080] S6637, Iterative process: Repeat neighborhood search, evaluation, acceptance criteria and update operations until the stopping condition is met, such as reaching the maximum number of iterations or the temperature drops to a certain threshold.

[0081] S6637, Pareto solution set construction: At the end of the algorithm, a Pareto optimal solution set is extracted from all accepted solutions, which provide different trade-offs between multiple objectives.

[0082] S7, input the feature-extracted data into a multi-objective optimization algorithm to find the optimal balance between different objectives and determine the optimal value range of several parameter data. For example, the optimal value range of parameters such as powder feeding flow rate, speed, temperature, pressure and argon content is determined by non-dominated sorting genetic algorithm II.

[0083] Further, refer to Figure 2 As shown, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method also includes executing steps S8 to S10 after step S7: S8. System integration and implementation: Integrate the multi-objective optimization algorithm into the equipment control system to realize automatic parameter adjustment; on the premise that one of the three different objectives is selected, update the equipment's operating procedures according to the optimization results to guide actual operations.

[0084] S9. Real-time monitoring and adaptive adjustment: Implement a real-time monitoring system to continuously collect equipment operation data and monitor whether the parameters are within the optimal range; develop an adaptive adjustment mechanism so that when the equipment operation deviates from the optimal state, the system can automatically adjust the parameters.

[0085] S10. Continuous optimization and iteration: Based on actual production feedback, continuously optimize the multi-objective optimization algorithm and unsupervised learning model; regularly review and update the normal value range of parameters to adapt to changes in production conditions.

[0086] The embodiments provided by the present invention have the following technical effects: 1. Improve production efficiency: Improve production efficiency by determining the optimal operating parameters and ensure that the equipment operates in the optimal state.

[0087] 2. Ensure product quality: Ensure that the quality of finished products produced during the production process is stable and meets industry standards and customer needs.

[0088] 3. Reduce energy consumption: Optimize parameter settings, reduce unnecessary energy waste, and achieve the dual goals of cost-effectiveness and environmental friendliness.

[0089] 4. Realize automated control: Through algorithm analysis, provide data support for the equipment's automated control system to achieve intelligent production.

[0090] 5. Multi-objective optimization: Considering that there may be multiple conflicting objectives in the production process, such as efficiency and quality, the multi-objective optimization algorithm can balance these objectives and find the best solution.

[0091] 6. Anomaly detection and prediction: By analyzing the normal value range, the system's ability to detect abnormal conditions is enhanced and potential production problems are predicted in advance.

[0092] The above mainly introduces the solution provided by the embodiment of the present invention from the perspective of the device. It is understandable that in order to realize the above functions, the device includes a hardware structure and / or software module corresponding to each function. Those skilled in the art should easily realize that, in combination with the algorithm steps of each example described in the embodiment disclosed in this article, the present invention can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0093] Figure 7 A possible schematic diagram of the composition of the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device involved in the above embodiment is shown, Figure 7 As shown, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device may include: a parameter acquisition module 10, a preprocessing module 20, a feature extraction module 30, an unsupervised learning module 40 and a multi-objective optimization module 50; A parameter acquisition module 10, used to acquire parameter data; A preprocessing module 20, used to preprocess parameter data according to a preset algorithm; A feature extraction module 30 is used to extract and construct features from the preprocessed parameter data; An unsupervised learning module 40 is used to construct a training data set and a validation data set from the data after feature extraction and feature construction; It is also used to build unsupervised learning models with training data sets, and to evaluate and optimize unsupervised learning models with validation data sets; A multi-objective optimization module 50, used to determine a multi-objective optimization algorithm; It is also used to input the feature-extracted data into a multi-objective optimization algorithm to find the optimal balance between different objectives and confirm the most appropriate value range for several parameter data.

[0094] Of course, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device provided in the embodiment of the present invention includes but is not limited to the above modules.

[0095] In an embodiment of the present invention, a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device is also provided, and the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device includes: a processor and a memory. The memory is used to store computer program code, and the computer program code includes computer instructions. When the processor executes the computer instructions, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device performs each step performed by the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device in the method flow shown in the above method embodiment.

[0096] An embodiment of the present invention also provides a computer-readable storage medium on which computer instructions are stored. When the computer instructions are run on a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device executes each step executed by the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device in the method flow shown in the above method embodiment.

[0097] An embodiment of the present invention also provides a computer program product, which includes computer instructions. When the computer instructions are run on a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device executes each step executed by the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device in the method flow shown in the above method embodiment.

[0098] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A multi-objective optimization method for plasma melt-blown powder spheroidization equipment parameter analysis, characterized in that: include: Get parameter data; Preprocessing the parameter data according to a preset algorithm; Perform feature extraction and feature construction on the preprocessed parameter data; Use the data after feature extraction and feature construction to construct training data sets and validation data sets; Using the training data set to construct an unsupervised learning model, and using the validation data set to evaluate and optimize the unsupervised learning model; Determine the multi-objective optimization algorithm; The data after feature extraction is input into the multi-objective optimization algorithm to find the optimal balance between different objectives and confirm the most appropriate value range of several parameter data.

2. The multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method according to claim 1 is characterized in that: The unsupervised learning model includes: a cluster analysis algorithm and an autoencoder algorithm.

3. The multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method according to claim 2 is characterized in that: The cluster analysis algorithm includes: Data preprocessing: clean the data, remove missing values ​​and outliers, and standardize or normalize the data; Select clustering algorithm: Determine the clustering algorithm based on data characteristics; Determine the number of clusters: Use any of the methods such as elbow rule and silhouette coefficient to determine the optimal number of clusters; Perform clustering: Perform cluster analysis on the data set based on the clustering algorithm; Evaluate clustering results: Use the silhouette coefficient or Davies-Bouldin index to evaluate clustering effect; Interpret clustering results: Interpret data patterns based on clustering results and identify potential structures in the data.

4. The multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method according to claim 2 is characterized in that: The autoencoder algorithm includes: Data preprocessing: Data cleaning and standardization through homocluster analysis; Design the autoencoder network structure: determine the size of the input layer; Select activation function: Determine the activation function of the hidden layer of the autoencoder; Training the autoencoder: Use the dataset to train the autoencoder and optimize the network weights through the back-propagation algorithm; Feature learning: Using low-dimensional representations to learn hidden layers and capture the intrinsic structure of the data; Anomaly detection or dimensionality reduction: Anomaly detection is performed based on the reconstruction error of the autoencoder, and the data representation learned by the autoencoder is used as the reduced dimension feature input into other machine learning models; Evaluation and Interpretation: Evaluate performance and interpret learned features to understand the structure of the data.

5. The multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method according to claim 1 is characterized in that: The determining of the multi-objective optimization algorithm comprises: Objective function definition: define three optimization objectives; Feature value selection: determine the key feature values ​​that affect the target; Eigenvalue modeling: Establish a relationship model between eigenvalues ​​and optimization targets, and use historical data to train the model; Constraint setting: According to process requirements, equipment characteristics and unsupervised learning results, set the operating range and constraints of each feature value; Pareto Optimality Definition: In multi-objective optimization, define the Pareto superiority criterion; Algorithm selection: Determine at least one algorithm from among multi-objective particle swarm optimization, non-dominated sorting genetic algorithm II and multi-objective simulated annealing; Algorithm framework design: Design the algorithm framework, including at least: initialization, iteration process, selection operation, crossover and mutation operation, and Pareto frontier update; Performance evaluation: Design performance evaluation criteria; Algorithm implementation: Implement the algorithm and code it using the Java programming language and tools; Solve the Pareto optimal solution set: run the algorithm to generate the Pareto optimal solution set; Result analysis: Analyze the Pareto optimal solution set, identify feasible solutions, and select them based on process experience and evaluation criteria such as comparison of convergence accuracy, solution set diversity, calculation time and stability; Experimental verification: Combined with the verification set, the effectiveness of the algorithm is verified.

6. A multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device, characterized in that: include: Parameter acquisition module, used to obtain parameter data; A preprocessing module, used for preprocessing the parameter data according to a preset algorithm; A feature extraction module is used to extract and construct features from the preprocessed parameter data; Unsupervised learning module, used to construct training data set and validation data set from the data after feature extraction and feature construction; It is also used to construct an unsupervised learning model using the training data set, and to evaluate and optimize the unsupervised learning model using the verification data set; Multi-objective optimization module, used to determine the multi-objective optimization algorithm; It is also used to input the feature-extracted data into a multi-objective optimization algorithm to find the optimal balance between different objectives and confirm the most appropriate value range for several parameter data.

7. A multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device, characterized in that: It includes a processor and a memory, the memory is used to store computer program code, the computer program code includes computer instructions, when the processor executes the computer instructions, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device executes the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that: It includes computer instructions. When the computer instructions are run on a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device executes the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method as described in any one of claims 1 to 5.

9. A computer program product, which includes computer instructions. When the computer instructions are run on a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device, the multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis device performs a multi-objective optimization plasma melt-blown powder spheroidization equipment parameter analysis method as described in any one of claims 1 to 5.

Images