Model optimization method for heat treatment furnace

Abstract

The invention provides a model optimization method for a heat treatment furnace, and belongs to the technical field of heat treatment furnaces, and the method comprises the steps: carrying out the structural disassembly of the heat treatment furnace according to the internal structure of the heat treatment furnace, and constructing a part model of each disassembly part; according to the operation process of the heat treatment furnace, carrying out analogue simulation operation on all the component models, and obtaining an operation thread meeting a preset operation condition; obtaining a discriminant item under each operation index, and constructing an independent optimization function and an overall optimization function in combination with a standard item; performing primary simulation optimization on the component model in the running thread, and determining to obtain an optimized feature vector; according to the vector difference between the optimized feature vector and the initial feature vector of the preset operation condition and the required precision, the optimized feature vector is optimized again, and simulation optimization of the operation thread is achieved. The simulation optimization effect is improved, and the small error of the heat treatment furnace in the actual operation process is guaranteed.

Description

Technical Field

[0001] The invention relates to the technical field of heat treatment furnaces, and in particular to a model optimization method for heat treatment furnaces. Background Art

[0002] Vacuum heat treatment is a comprehensive technology that combines vacuum technology and heat treatment. It means that all or part of the heat treatment process is carried out under vacuum. The vacuum heat treatment furnace has high thermal efficiency, can achieve rapid heating and cooling, can achieve no oxidation, no decarburization, no carburization, can remove phosphorus chips on the surface of the workpiece, and has the functions of degreasing and degassing, so as to achieve the effect of surface brightness and purification.

[0003] The heat treatment furnace needs to be tested before it is actually put into operation. Production will only continue when it meets the operating standards. However, before the factory test, the heat treatment furnace is generally subjected to a physical test, as well as a simulation test of the overall design and structure of the heat treatment furnace. This method will guarantee certain test results. However, because it is an overall test, there will be excessive simulation errors as a whole, which will lead to even greater errors in the physical test.

[0004] Therefore, the present invention proposes a model optimization method for a heat treatment furnace. Summary of the invention

[0005] The present invention provides a model optimization method for a heat treatment furnace, which is used to obtain individual optimization functions and an overall optimization function of different operating indicators by constructing several component models for simulation, thereby optimizing the model for the first time, and optimizing the model again according to vector differences and required accuracy, thereby improving the simulation optimization effect and ensuring a smaller error in the actual operation of the heat treatment furnace.

[0006] The present invention provides a model optimization method for a heat treatment furnace, comprising:

[0007] Step 1: Disassemble the heat treatment furnace according to its internal structure, and construct a component model of each disassembled component;

[0008] Step 2: According to the operation process of the heat treatment furnace, all component models are simulated and run to obtain the running threads that meet the preset running conditions;

[0009] Step 3: Perform actual heat treatment according to the running thread, obtain the discriminant item under each running index and combine it with the standard item to construct a separate optimization function for each running index and the overall optimization function of the running thread;

[0010] Step 4: Performing initial simulation optimization on the component model in the running thread based on the individual optimization function and the overall optimization function to determine the optimized feature vector;

[0011] Step 5: According to the vector difference between the optimized feature vector and the initial feature vector of the preset operating condition and the required accuracy, the optimized feature vector is optimized again to achieve simulation optimization of the running thread.

[0012] Preferably, constructing a component model of each disassembled component includes:

[0013] The components of each disassembled component are obtained according to the structural disassembly results;

[0014] The components are sequentially assembled in a simulated manner according to the internal structure of the heat treatment furnace to obtain component models corresponding to the disassembled components.

[0015] Preferably, according to the operation process of the heat treatment furnace, all component models are simulated and run, including:

[0016] Arranging all component models according to the working principle of the heat treatment furnace to obtain a model assembly line;

[0017] The operation process of the heat treatment furnace is matched with the model pipeline according to the control program of the execution stage, and all component models are controlled to perform simulation operation.

[0018] Preferably, obtaining a running thread that meets a preset running condition includes:

[0019] Capture all simulation operation parameters involved in each component model separately;

[0020] The captured parameters are compared and analyzed with the preset operating conditions, and the simulation operating parameters are adjusted to meet the preset operating conditions to obtain the running thread.

[0021] Preferably, constructing a separate optimization function for each running indicator and an overall optimization function for the running thread includes:

[0022] Acquire processing data during the actual heat treatment process, and at the same time, obtain the operating index of each actual component in the heat treatment furnace from the component-index mapping table;

[0023] Extracting the processed data according to the operation index to obtain first data matching the operation index, and decomposing the index structure of the operation index to determine a discriminant item;

[0024] Mapping the discriminant item with the first data and comparing them with the standard item to construct a first comparison function for each operation index;

[0025] Determining a separate optimization function according to the comparison difference of each function parameter in the first comparison function and the parameter weight;

[0026] Based on all the individual optimization functions and the operational relationships between different actual components, an overall optimization function of the running thread is constructed.

[0027] Preferably, based on all the individual optimization functions and the operational relationships between different actual components, an overall optimization function of the running thread is constructed, including:

[0028] Determine the operation stage of each actual component in the heat treatment furnace according to the thread mode of the operation thread, and construct a first heat treatment efficiency of each actual component involved in the operation stage at different actual heat treatment moments to construct a heat treatment diagram of the corresponding actual component, and then combine the operation relationship between different actual components in the corresponding operation stage to calculate the overall efficiency at each actual heat treatment moment in the corresponding operation stage;

[0029]

[0030] Among them, XZ t represents the overall efficiency at the tth actual heat treatment moment in the corresponding operation stage; n01 represents the total number of actual components involved in the corresponding operation stage; y1 i1 represents the first heat treatment efficiency of the i1th actual component at the tth actual heat treatment moment in the corresponding operation stage; n01 i1 y2 represents the number of components that have an operating relationship with the i1th actual component, determined based on the operating relationship between different actual components in the corresponding operating stage; i2 represents the first heat treatment efficiency of the i2th component that has an operating relationship with the i1th actual component at the tth actual heat treatment moment in the corresponding operation stage; δ1 i2 It represents the operation factor between the i1th actual component and the i2th component that have an operation relationship with the ith actual component at the tth actual heat treatment moment in the corresponding operation stage; δ1 i1 It represents the factor average value of the operating factor that satisfies the normal distribution probability with the ith actual component at the tth actual heat treatment time in the corresponding operation stage; σi1 2 represents all δ1 i2 Variance of 2 represents all y1 i1 The variance of

[0031] Based on the overall efficiency at different actual heat treatment moments, an overall efficiency diagram of the corresponding operation stage is constructed;

[0032] According to all the thermal treatment diagrams and the overall efficiency diagram in the same operation stage, a processing array in the corresponding operation stage is constructed;

[0033] According to the phase pipeline order of different running phases in the running thread and in combination with the processing array, a first efficiency graph and a second efficiency graph are constructed;

[0034] Comparing and analyzing the first efficiency graph and the second efficiency graph to obtain a first difference value in each operation stage;

[0035] At the same time, the first efficiency values ​​of adjacent operating stages in the first efficiency diagram are compared and analyzed to obtain second difference values ​​in adjacent stages;

[0036] Comparing and analyzing the second efficiency values ​​of adjacent operating stages in the second efficiency diagram, and obtaining third difference values ​​in adjacent stages;

[0037] Compare and analyze the second difference value and the third difference value in the same adjacent stage to obtain the fifth difference value in the corresponding adjacent stage;

[0038] Inserting the fifth difference value in the adjacent stage after the first difference value in the operation stage consistent with the previous stage according to the time of the previous stage, to obtain an analysis diagram;

[0039] Determine the distribution probability of the difference values ​​contained in all moments of each two operation stages in the analysis diagram at different difference levels to obtain a distribution structure;

[0040] All distribution structures are mapped to all individual optimization functions as a whole, and the overall optimization function of the running thread is constructed.

[0041] Preferably, performing an initial simulation optimization on the component model in the running thread based on the individual optimization function and the overall optimization function to determine the optimized feature vector includes:

[0042] Performing a first optimization on the corresponding component models in the running thread according to the separate optimization function, and recording a first optimization log of each component model;

[0043] Performing a second optimization on all component models in the running thread according to the overall optimization function, and recording a second optimization log of all component models;

[0044] Arrange all first optimization logs in the optimization order and align them with the second optimization logs to determine optimization conflicts;

[0045] The optimization contradiction is eliminated, and the optimization feature vector is obtained by combining the optimization log without contradiction.

[0046] Preferably, the optimized feature vector is optimized again according to the vector difference between the optimized feature vector and the initial feature vector of the preset operating condition and the required accuracy, including:

[0047] Comparing the optimized feature vector with the initial feature vector to determine a first distance between the same elements;

[0048] extracting second distances that do not meet the consistency criterion and third distances that meet the consistency criterion from all first distances, and determining the distance variation coefficient of each second distance according to the distance criteria of all third distances;

[0049]

[0050] Among them, L2 j1 represents the j1th second distance; XL L3 L3 represents the fitting distance after fitting based on all third distances; j2 represents the j2th third distance; m1 represents the number of third distances; Δ1 represents the absolute difference between the first point and the last point on the fitting line obtained after fitting all third distances; σL 2 represents the variance of all third distances; Δ2 represents the absolute difference between any point on the left straight line and any point on the right straight line on the fitted straight line obtained after fitting all third distances; b2 j1 represents the distance variation coefficient of the j1th second distance;

[0051] According to the following formula, each second distance is corrected separately:

[0052]

[0053] Among them, L0 j1 represents the distance after correction of the j1th second distance; y01 represents the accuracy conversion factor of the coefficient of variation; yJ represents the required accuracy; ∝0 represents the accuracy conversion factor; In represents the logarithmic function symbol;

[0054] According to the corrected distance, a re-optimized vector is constructed.

[0055] Compared with the prior art solutions, the beneficial effects of this application are as follows:

[0056] By constructing several component models for simulation, individual optimization functions and overall optimization functions of different operating indicators are obtained, and the model is optimized for the first time. Then, the model is optimized again according to the vector difference and the required accuracy, which improves the simulation optimization effect and ensures a smaller error in the actual operation of the heat treatment furnace.

[0057] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0058] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0060] Figure 1 The present invention is a flow chart of a method for optimizing a model of a heat treatment furnace according to an embodiment of the present invention. DETAILED DESCRIPTION

[0061] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0062] The present invention provides a model optimization method for a heat treatment furnace, such as Figure 1 As shown, including:

[0063] Step 1: Disassemble the heat treatment furnace according to its internal structure, and construct a component model of each disassembled component;

[0064] Step 2: According to the operation process of the heat treatment furnace, all component models are simulated and run to obtain the running threads that meet the preset running conditions;

[0065] Step 3: Perform actual heat treatment according to the running thread, obtain the discriminant item under each running index and combine it with the standard item to construct a separate optimization function for each running index and the overall optimization function of the running thread;

[0066] Step 4: Performing initial simulation optimization on the component model in the running thread based on the individual optimization function and the overall optimization function to determine the optimized feature vector;

[0067] Step 5: According to the vector difference between the optimized feature vector and the initial feature vector of the preset operating condition and the required accuracy, the optimized feature vector is optimized again to achieve simulation optimization of the running thread.

[0068] In this embodiment, the internal structure is a structural diagram predetermined by experts before the manufacture of the heat treatment furnace. Different disassembled parts are obtained by structurally disassembling the diagram. Because the disassembled parts need to be complete during the simulation, a component model is constructed to facilitate the simulation operation.

[0069] In this embodiment, for example, the component models in the heat treatment furnace include: model 01, model 02 and model 03, and the operation process refers to the work flow of the heat treatment furnace determined by experts, for example: operation stage 1 corresponds to model 01, operation stage 2 corresponds to model 02, and operation stage corresponds to model 03, or operation stage 2 corresponds to model 02 and model 03, and operation stage 3 corresponds to model 01, model 02 and model 03. Finally, it is found that the preset operating conditions are met by model 01 (after adjusting the original parameters in model 01), model 02, model 03-model 01 (after adjusting the original parameters in model 01), model 02, model 03.

[0070] In this embodiment, the preset running condition is pre-set. Because it is uncertain which process the simulation operation under can meet the preset running condition, simulation will be performed according to each process to obtain a satisfactory running thread.

[0071] In this embodiment, actual heat treatment refers to actual application of a heat treatment furnace.

[0072] In this embodiment, the running index is pre-set based on the running thread. For example, the running thread 01 includes the judgment item a1 and the judgment item a3, and the judgment item a1 and the judgment item a3 both have corresponding set standard items, in order to facilitate comparison and determine the priority of their existence.

[0073] In this embodiment, one operating index corresponds to a separate optimization function, and the operating threads corresponding to all operating indexes correspond to an overall optimization function, the main purpose of which is to achieve optimization simulation of the component model from a single aspect and an overall aspect.

[0074] In this embodiment, after the initial simulation optimization of the model, an optimized feature vector is obtained, and then the vector difference and precision setting are performed to achieve the optimization of the vector, thereby achieving the simulation optimization of the running thread.

[0075] The beneficial effects of the above technical solution are: by constructing several component models to simulate and obtain individual optimization functions and overall optimization functions of different operating indicators, the model is optimized for the first time, and by optimizing the model again according to the vector difference and the required accuracy, the simulation optimization effect is improved, and the smaller error of the heat treatment furnace during the actual operation is ensured.

[0076] The present invention provides a model optimization method for a heat treatment furnace, which constructs a component model of each disassembled component, including:

[0077] The components of each disassembled component are obtained according to the structural disassembly results;

[0078] The components are sequentially assembled in a simulated manner according to the internal structure of the heat treatment furnace to obtain component models corresponding to the disassembled components.

[0079] In this embodiment, the device structure involved in the component structure of each disassembled component is the component part. For example, the disassembled component involves part 01 and part 02. In this case, part 01 and part 02 are the component parts.

[0080] In this embodiment, the internal structure is a structural diagram generated by the structural design of the heat treatment furnace, and is spliced ​​according to the chronological construction sequence of the structural diagram. That is, the splicing result obtained after mechanically or circuit splicing part 01 and part 02 according to the corresponding structural diagram is the component model.

[0081] The beneficial effect of the above technical solution is: by obtaining the components of each disassembled component and performing simulated sequential splicing, it is easy to obtain a component model, which provides a basis for subsequent determination of the running thread and simulation optimization.

[0082] The present invention provides a model optimization method for a heat treatment furnace, which performs simulation operation on all component models according to the operation process of the heat treatment furnace, including:

[0083] Arranging all component models according to the working principle of the heat treatment furnace to obtain a model assembly line;

[0084] The operation process of the heat treatment furnace is matched with the model pipeline according to the control program of the execution stage, and all component models are controlled to perform simulation operation.

[0085] In this embodiment, the working principle of the heat treatment furnace is determined by experts, and it is known what operations different models perform, so the models can be arranged to obtain a model pipeline.

[0086] In this embodiment, the structural components and models involved in the operation process and the model pipeline are in principle one-to-one correspondence, so that the execution stage 1 in the operation process corresponds to model 01, the execution stage 2 corresponds to model 02 and model 03, and the execution stage 3 corresponds to model 01, model 02 and model 03. At this time, simulation can be achieved by controlling through known programs.

[0087] The beneficial effect of the above technical solution is: the pipeline is obtained by arranging the model according to the working principle, and the model is further simulated and operated by matching it with the operation process.

[0088] The present invention provides a model optimization method for a heat treatment furnace, which obtains an operating thread that meets preset operating conditions, including:

[0089] Capture all simulation operation parameters involved in each component model separately;

[0090] The captured parameters are compared and analyzed with the preset operating conditions, and the simulation operating parameters are adjusted to meet the preset operating conditions to obtain the running thread.

[0091] In this embodiment, the simulation process is implemented based on the simulation platform. The parameter capture tool set on the platform will realize automatic parameter capture during the simulation of the model on the platform. They are all set parameter descriptions that need to be captured, and the main thing is to capture the parameter values.

[0092] In this embodiment, the preset running conditions are to ensure that the captured parameters can meet the set conditions to obtain a reasonable thread. For example, the captured parameter is u1, but u1 is not within the range corresponding to the preset running conditions. At this time, it is necessary to adjust u1 to the minimum set parameter of the corresponding range to obtain the running thread, that is, the simulated running threads in the thread all meet the preset running conditions.

[0093] The beneficial effect of the above technical solution is: by capturing the simulation operation parameters and comparing and analyzing them with the preset operation conditions, an operation thread that meets the preset operation conditions is realized, which facilitates subsequent optimization simulation, indirectly improves the optimization effect, and reduces the error of actual operation.

[0094] The present invention provides a model optimization method for a heat treatment furnace, which constructs a separate optimization function for each operating index and an overall optimization function for the operating thread, including:

[0095] Acquire processing data during the actual heat treatment process, and at the same time, obtain the operating index of each actual component in the heat treatment furnace from the component-index mapping table;

[0096] Extracting the processed data according to the operation index to obtain first data matching the operation index, and decomposing the index structure of the operation index to determine a discriminant item;

[0097] Mapping the discriminant item with the first data and comparing them with the standard item to construct a first comparison function for each operation index;

[0098] Determining a separate optimization function according to the comparison difference of each function parameter in the first comparison function and the parameter weight;

[0099] Based on all the individual optimization functions and the operational relationships between different actual components, an overall optimization function of the running thread is constructed.

[0100] In this embodiment, the processed data is the data captured during the heat treatment process.

[0101] In this embodiment, the component-index mapping table includes different components and operating indicators matching the components, and is set in advance.

[0102] In this embodiment, for example, the operation index includes index 1 and index 2, and the first data matching the operation index 1 is data 1, and the first data matching the operation index 2 is data 2.

[0103] In this embodiment, the operation index has a certain index structure, for example, index head description-indicator middle description-indicator tail description, then, after disassembling the index structure, the corresponding discriminant item can be obtained.

[0104] The first data is mapped to the discrimination item to obtain the actual data under each discrimination item, and then by comparing with the standard data under the standard item, the first comparison function B01 (bi2, ci2) can be obtained, where bi2 is for the i2th discrimination item and ci2 is for the i2th standard item.

[0105] In this embodiment, the comparison difference is the difference between the actual and the standard under the same item, and the parameter weight is pre-set based on the operation index.

[0106] In this embodiment, a separate optimization function D01 (B01, qz) is used, where qz represents a weight.

[0107] In this embodiment, the operating relationship refers to the actual correlation between components during the operation of the heat treatment furnace. For example, component 1 drives component 2 to work, or the result generated by component 1 is transmitted back to component 2, that is, the working conditions between different components are determined.

[0108] In this embodiment, the overall optimization function ZT(D01, Yz), Yz represents the distribution structure related to the operation relationship.

[0109] The beneficial effects of the above technical solution are: by obtaining operating indicators and matching data to indicators, and then decomposing the indicator structure to obtain discriminant items and mapping data, comparing with standard items, constructing a comparison function, and combining parameter weights to determine a separate optimization function, and combining the operating relationship to construct an overall optimization function, that is, starting from both the individual and the overall aspects, ensuring the rationality of subsequent optimization, and indirectly improving the optimization accuracy.

[0110] The present invention provides a model optimization method for a heat treatment furnace, which constructs an overall optimization function of the running thread based on all individual optimization functions and the operating relationship between different actual components, including:

[0111] Determine the operation stage of each actual component in the heat treatment furnace according to the thread mode of the operation thread, and construct a first heat treatment efficiency of each actual component involved in the operation stage at different actual heat treatment moments to construct a heat treatment diagram of the corresponding actual component, and then combine the operation relationship between different actual components in the corresponding operation stage to calculate the overall efficiency at each actual heat treatment moment in the corresponding operation stage;

[0112]

[0113] Among them, XZ t represents the overall efficiency at the tth actual heat treatment moment in the corresponding operation stage; n01 represents the total number of actual components involved in the corresponding operation stage; y1 i1 represents the first heat treatment efficiency of the i1th actual component at the tth actual heat treatment moment in the corresponding operation stage; n01 i1 y2 represents the number of components that have an operating relationship with the i1th actual component, determined based on the operating relationship between different actual components in the corresponding operating stage; i2 represents the first heat treatment efficiency of the i2th component that has an operating relationship with the i1th actual component at the tth actual heat treatment moment in the corresponding operation stage; δ1 i2 It represents the operation factor between the i1th actual component and the i2th component that have an operation relationship with the ith actual component at the tth actual heat treatment moment in the corresponding operation stage; δ1 i1 It represents the factor average value of the operating factor that satisfies the normal distribution probability with the ith actual component at the tth actual heat treatment time in the corresponding operation stage; σi1 2 represents all δ1 i2 Variance of 2 represents all y1 i1 The variance of

[0114] Based on the overall efficiency at different actual heat treatment moments, an overall efficiency diagram of the corresponding operation stage is constructed;

[0115] According to all the thermal treatment diagrams and the overall efficiency diagram in the same operation stage, a processing array in the corresponding operation stage is constructed;

[0116] According to the phase pipeline order of different running phases in the running thread and in combination with the processing array, a first efficiency graph and a second efficiency graph are constructed;

[0117] Comparing and analyzing the first efficiency graph and the second efficiency graph to obtain a first difference value in each operation stage;

[0118] At the same time, the first efficiency values ​​of adjacent operating stages in the first efficiency diagram are compared and analyzed to obtain second difference values ​​in adjacent stages;

[0119] Comparing and analyzing the second efficiency values ​​of adjacent operating stages in the second efficiency diagram, and obtaining third difference values ​​in adjacent stages;

[0120] Compare and analyze the second difference value and the third difference value in the same adjacent stage to obtain the fifth difference value in the corresponding adjacent stage;

[0121] Inserting the fifth difference value in the adjacent stage after the first difference value in the operation stage consistent with the previous stage according to the time of the previous stage, to obtain an analysis diagram;

[0122] Determine the distribution probability of the difference values ​​contained in all moments of each two operation stages in the analysis diagram at different difference levels to obtain a distribution structure;

[0123] All distribution structures are mapped to all individual optimization functions as a whole, and the overall optimization function of the running thread is constructed.

[0124] In this embodiment, different thread modes involve different operating stages. For example, the operating stages of the heat treatment furnace include: stage 1-stage 2-stage 3, stage 1 is related to actual component 01, stage 2 is related to actual component 02 and actual component 03, and stage 3 is related to actual component 01, actual component 02 and actual component 03. The heat treatment efficiency of each component can be determined, and the heat treatment efficiency = actual heat treatment condition / standard heat treatment condition.

[0125] In this embodiment, there may be multiple moments in each operation stage, which may constitute an overall efficiency graph for the stage.

[0126] In this embodiment, the heat treatment diagram is directed to the heat treatment efficiency of the same component at different heat treatment times.

[0127] In this embodiment, the processing array for stage 2 is: the average heat treatment efficiency of the first heat treatment efficiency of component 01 and the first heat treatment efficiency of component 03, that is, the overall efficiency.

[0128] In this embodiment, the stage flow sequence is the order in which different stages are run in sequence, and the two efficiencies obtained in each stage are plotted separately to obtain a first efficiency graph and a second efficiency graph.

[0129] In this embodiment, the first efficiency diagram: stage 1: s01, stage 2: s02, stage 3: s03;

[0130] Second efficiency diagram: stage 1: s11, stage 2: s12, stage 3: s13;

[0131] First difference: stage 1: s01-s11, stage 2: s02-s12, stage 3: s03-s13;

[0132] The second difference: stage 1 and stage 2: s01-s02, stage 2 and stage 3: s02-s03;

[0133] The third difference: stage 1 and stage 2: s11-s12, stage 2 and stage 3: s12-s13;

[0134] Fifth difference: stage 1 and stage 2: s01-s02-(s11-s12), stage 2 and stage 3: s02-s03-(s12-s13);

[0135] In this embodiment, s01-s02-(s11-s12) is inserted before s01-s11, and s02-s03-(s12-s13) is inserted before s02-s12, and an analysis diagram can be obtained.

[0136] In this embodiment, the difference level is obtained by comparing the difference values ​​in the analysis diagram with a level mapping table. The mapping table is constructed by including different difference values ​​and expert level calibration of the difference values, and can be directly obtained.

[0137] In this embodiment, the distribution probability refers to the number of differences existing at the difference level / the number of all differences, and the distribution result is a structure that performs statistics on different difference levels and the distribution probabilities at the difference levels.

[0138] In this embodiment, after determining the corresponding difference levels under two different operating stages, the functions of the corresponding indicators under the corresponding stages can also be determined, and the overall optimization function can be obtained through one-to-one correspondence. That is, all individual optimization functions are adjusted according to the difference levels existing in the distribution structure, and the corresponding functions are adjusted according to the difference levels to obtain the overall optimization function, so as to minimize the differences.

[0139] The beneficial effects of the above technical solution are: by calculating the heat treatment efficiency and the overall efficiency of each component to construct the corresponding efficiency diagram, by constructing a processing array and comparative analysis between different diagrams, various differences in adjacent stages are obtained, providing a basis for subsequent analysis of the differences, and by inserting the corresponding positions of the stages, it is convenient to construct analysis diagrams and distribution structures, providing a basis for integrating separate optimization functions, thereby ensuring the reliability of the subsequent overall optimization function and ensuring the efficiency of model optimization.

[0140] The present invention provides a model optimization method for a heat treatment furnace, which performs initial simulation optimization on the component model in the running thread based on the individual optimization function and the overall optimization function to determine the optimized feature vector, including:

[0141] Performing a first optimization on the corresponding component models in the running thread according to the separate optimization function, and recording a first optimization log of each component model;

[0142] Performing a second optimization on all component models in the running thread according to the overall optimization function, and recording a second optimization log of all component models;

[0143] Arrange all first optimization logs in the optimization order and align them with the second optimization logs to determine optimization conflicts;

[0144] The optimization contradiction is eliminated, and the optimization feature vector is obtained by combining the optimization log without contradiction.

[0145] In this embodiment, the first optimization is to reduce the differences in the component models, and the second optimization is to reduce the differences in all the involved component models.

[0146] In this embodiment, since the individual optimization function is the basis of the overall optimization function, it does not mean that the optimization log of the individual optimization function is completely consistent with the optimization log of the overall optimization function. Therefore, the purpose of the alignment process is to align the optimization of the components to select the best optimization method for optimization. At this time, it is necessary to eliminate the optimization methods with poor optimization effects and conflicts, analyze the retained logs, and obtain the optimized feature vector. The retained logs contain different optimization parameters, which can be directly extracted. That is, the optimized feature vector = {optimization parameter 1, optimization parameter 2, optimization parameter 3...}.

[0147] The beneficial effect of the above technical solution is that by optimizing according to the individual optimization function and the overall optimization function respectively, the optimization contradiction is eliminated, so that the optimized feature vector is effectively obtained.

[0148] The present invention provides a model optimization method for a heat treatment furnace, which further optimizes the optimized feature vector according to the vector difference between the optimized feature vector and the initial feature vector of the preset operating conditions and the required accuracy, including:

[0149] Comparing the optimized feature vector with the initial feature vector to determine a first distance between the same elements;

[0150] extracting second distances that do not meet the consistency criterion and third distances that meet the consistency criterion from all first distances, and determining the distance variation coefficient of each second distance according to the distance criteria of all third distances;

[0151]

[0152] Among them, L2 j1 represents the j1th second distance; XL L3 L3 represents the fitting distance after fitting based on all third distances; j2 represents the j2th third distance; m1 represents the number of third distances; Δ1 represents the absolute difference between the first point and the last point on the fitting line obtained after fitting all third distances; σL 2 represents the variance of all third distances; Δ2 represents the absolute difference between any point on the left straight line and any point on the right straight line on the fitted straight line obtained after fitting all third distances; b2 j1 represents the distance variation coefficient of the j1th second distance;

[0153] According to the following formula, each second distance is corrected separately:

[0154]

[0155] Among them, L0 j1 represents the distance after correction of the j1th second distance; y01 represents the accuracy conversion factor of the coefficient of variation; yJ represents the required accuracy; ∝0 represents the accuracy conversion factor; In represents the logarithmic function symbol;

[0156] According to the corrected distance, a re-optimized vector is constructed.

[0157] In this embodiment, the first distance is calculated according to a distance function, and the result is the absolute value of the distance difference.

[0158] For example, there are distances G1, G2, G3 and G4. At this time, the distance differences of G1, G2 and G4 are all tending to be consistent, and only G3 is more prominent. At this time, G1, G2 and G4 are regarded as the third distance, and G3 is regarded as the second distance.

[0159] In this embodiment, the optimized vector = {parameter 1 corresponding to the corrected distance, parameter 2 corresponding to the corrected distance, ...}, wherein the elements in the optimized vector are directly matched based on the original optimization parameter-corrected distance-corrected optimization parameter mapping table.

[0160] The beneficial effect of the above technical solution is: the optimized feature vector is compared with the initial feature vector to determine the first distance, and the second distance is corrected by calculating the distance variation coefficient to obtain the optimized vector, thereby further ensuring the efficiency of the simulation optimization.

[0161] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A model optimization method for a heat treatment furnace, characterized in that: include: Step 1: Disassemble the heat treatment furnace according to its internal structure, and construct a component model of each disassembled component; Step 2: According to the operation process of the heat treatment furnace, all component models are simulated and run to obtain the running threads that meet the preset running conditions; Step 3: Perform actual heat treatment according to the running thread, obtain the discriminant item under each running index and combine it with the standard item to construct a separate optimization function for each running index and the overall optimization function of the running thread; Step 4: Performing initial simulation optimization on the component model in the running thread based on the individual optimization function and the overall optimization function to determine the optimized feature vector; Step 5: According to the vector difference between the optimized feature vector and the initial feature vector of the preset operating condition and the required accuracy, the optimized feature vector is optimized again to achieve simulation optimization of the running thread.

2. The model optimization method of a heat treatment furnace according to claim 1, characterized in that: Construct component models for each disassembled component, including: The components of each disassembled component are obtained according to the structural disassembly results; The components are sequentially assembled in a simulated manner according to the internal structure of the heat treatment furnace to obtain component models corresponding to the disassembled components.

3. The model optimization method of a heat treatment furnace according to claim 1, characterized in that: According to the operation process of the heat treatment furnace, all component models are simulated and run, including: Arranging all component models according to the working principle of the heat treatment furnace to obtain a model assembly line; The operation process of the heat treatment furnace is matched with the model pipeline according to the control program of the execution stage, and all component models are controlled to perform simulation operation.

4. The model optimization method of a heat treatment furnace according to claim 1, characterized in that: Obtain the running threads that meet the preset running conditions, including: Capture all simulation operation parameters involved in each component model separately; The captured parameters are compared and analyzed with the preset operating conditions, and the simulation operating parameters are adjusted to meet the preset operating conditions to obtain the running thread.

5. The model optimization method of a heat treatment furnace according to claim 1, characterized in that: Constructing a separate optimization function for each running indicator and an overall optimization function for the running thread, including: Acquire processing data during the actual heat treatment process, and at the same time, obtain the operating index of each actual component in the heat treatment furnace from the component-index mapping table; Extracting the processed data according to the operation index to obtain first data matching the operation index, and decomposing the index structure of the operation index to determine a discriminant item; Mapping the discriminant item with the first data and comparing them with the standard item to construct a first comparison function for each operation index; Determining a separate optimization function according to the comparison difference of each function parameter in the first comparison function and the parameter weight; Based on all the individual optimization functions and the operational relationships between different actual components, an overall optimization function of the running thread is constructed.

6. A model optimization method for a heat treatment furnace according to claim 5, characterized in that: Based on all the individual optimization functions and the operational relationships between different actual components, an overall optimization function of the running thread is constructed, including: Determine the operation stage of each actual component in the heat treatment furnace according to the thread mode of the operation thread, and construct a first heat treatment efficiency of each actual component involved in the operation stage at different actual heat treatment moments to construct a heat treatment diagram of the corresponding actual component, and then combine the operation relationship between different actual components in the corresponding operation stage to calculate the overall efficiency at each actual heat treatment moment in the corresponding operation stage; Among them, XZ t represents the overall efficiency at the tth actual heat treatment moment in the corresponding operation stage; n01 represents the total number of actual components involved in the corresponding operation stage; y1 i1 represents the first heat treatment efficiency of the i1th actual component at the tth actual heat treatment moment in the corresponding operation stage; n01 i1 y2 represents the number of components that have an operating relationship with the i1th actual component, determined based on the operating relationship between different actual components in the corresponding operating stage; i2 represents the first heat treatment efficiency of the i2th component that has an operating relationship with the i1th actual component at the tth actual heat treatment moment in the corresponding operation stage; δ1 i2 It represents the operation factor between the i1th actual component and the i2th component that have an operation relationship with the ith actual component at the tth actual heat treatment moment in the corresponding operation stage; δ1 i1 It represents the factor average value of the operating factor that satisfies the normal distribution probability with the ith actual component at the tth actual heat treatment time in the corresponding operation stage; σi1 2 represents all δ1 i2 Variance of 2 represents all y1 i1 The variance of Based on the overall efficiency at different actual heat treatment moments, an overall efficiency diagram of the corresponding operation stage is constructed; According to all the thermal treatment diagrams and the overall efficiency diagram in the same operation stage, a processing array in the corresponding operation stage is constructed; According to the phase pipeline order of different running phases in the running thread and in combination with the processing array, a first efficiency graph and a second efficiency graph are constructed; Comparing and analyzing the first efficiency graph and the second efficiency graph to obtain a first difference value in each operation stage; At the same time, the first efficiency values ​​of adjacent operating stages in the first efficiency diagram are compared and analyzed to obtain second difference values ​​in adjacent stages; Comparing and analyzing the second efficiency values ​​of adjacent operating stages in the second efficiency diagram, and obtaining third difference values ​​in adjacent stages; Compare and analyze the second difference value and the third difference value in the same adjacent stage to obtain the fifth difference value in the corresponding adjacent stage; Inserting the fifth difference value in the adjacent stage after the first difference value in the operation stage consistent with the previous stage according to the time of the previous stage, to obtain an analysis diagram; Determine the distribution probability of the difference values ​​contained in all moments of each two operation stages in the analysis diagram at different difference levels to obtain a distribution structure; All distribution structures are mapped to all individual optimization functions as a whole, and the overall optimization function of the running thread is constructed.

7. The model optimization method of a heat treatment furnace according to claim 1, characterized in that: Based on the individual optimization function and the overall optimization function, the component model in the running thread is initially simulated and optimized to determine the optimized feature vector, including: Performing a first optimization on the corresponding component models in the running thread according to the separate optimization function, and recording a first optimization log of each component model; Performing a second optimization on all component models in the running thread according to the overall optimization function, and recording a second optimization log of all component models; Arrange all first optimization logs in the optimization order and align them with the second optimization logs to determine optimization conflicts; The optimization contradiction is eliminated, and the optimization feature vector is obtained by combining the optimization log without contradiction.

8. The model optimization method of a heat treatment furnace according to claim 1, characterized in that: According to the vector difference between the optimized feature vector and the initial feature vector of the preset operating conditions and the required accuracy, the optimized feature vector is optimized again, including: Comparing the optimized feature vector with the initial feature vector to determine a first distance between the same elements; extracting second distances that do not meet the consistency criterion and third distances that meet the consistency criterion from all first distances, and determining the distance variation coefficient of each second distance according to the distance criteria of all third distances; Among them, L2 j1 represents the j1th second distance; XL L3 L3 represents the fitting distance after fitting based on all third distances; j2 represents the j2th third distance; m1 represents the number of third distances; Δ1 represents the absolute difference between the first point and the last point on the fitting line obtained after fitting all third distances; σL 2 represents the variance of all third distances; Δ2 represents the absolute difference between any point on the left straight line and any point on the right straight line on the fitted straight line obtained after fitting all third distances; b2 j1 represents the distance variation coefficient of the j1th second distance; According to the following formula, each second distance is corrected separately: Among them, L0 j1 represents the distance after correction of the j1th second distance; y01 represents the accuracy conversion factor of the coefficient of variation; yJ represents the required accuracy; ∝0 represents the accuracy conversion factor; In represents the logarithmic function symbol; According to the corrected distance, a re-optimized vector is constructed.

Images