Intelligent compensation method for binaural cupping

Abstract

The present application relates to the technical field of binaural ear can deformation compensation, and discloses a binaural ear can deformation intelligent compensation method, which first collects binaural ear can structure parameters and working environment data, analyzes and then constructs a deformation feature set to determine an initial compensation feature library; determines compensation model update information set according to dynamic optimization operation feedback of the initial compensation feature library, obtains target compensation parameters through hierarchical disassembly classification statistics; then acquires compensation resource configuration information set, determines and outputs an adaptive compensation scheme based on the target compensation parameters. This method realizes intelligent and accurate compensation of binaural ear can deformation through multi-dimensional data collection and analysis, dynamic update of the compensation model, and comprehensive consideration of historical data optimization of compensation resource configuration, improves compensation accuracy and efficiency, reduces cost, enhances system stability and adaptability, and is suitable for binaural ear can deformation compensation scenes in multiple industries such as chemical industry and metallurgy.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of bung jar shape deformation compensation, in particular to a bung jar shape deformation intelligent compensation method. BACKGROUND

[0002] In the field of industrial production and manufacturing, bung jars, as a common container or structural component, are widely used in chemical industry, metallurgy, energy and other industries. However, in actual use, bung jars will be affected by various factors and will deform, which not only affects their normal function, but also may cause safety hazards.

[0003] From the aspect of structural parameters, the different structural parameters of the bung jar itself, such as geometric shape, size, material properties, etc., will cause differences in its deformation under the same external conditions. For example, different structural geometric feature groups will make the position correlation weight and structural influence score of each feature point different, thereby affecting the overall deformation mode.

[0004] The complexity of the working environment also has an important influence on the deformation of the bung jar. Changes in the load cycle will cause the bung jar to bear different loads at different stages, thereby causing different key observation stages, and the rate of change of the temperature change environment feature will also affect the mechanical properties of the material, making the deformation of the bung jar more complex.

[0005] The existing bung jar deformation compensation method has many shortcomings when faced with these complex situations. Traditional methods often have difficulty in comprehensively and accurately collecting and analyzing the structural parameters and working environment data of the bung jar, and cannot construct a complete deformation feature set, resulting in an imperfect initial compensation feature library, making it difficult to accurately predict and compensate for deformation. In terms of updating and optimizing the compensation model, traditional methods lack the ability to dynamically adjust, and cannot update the model in a timely manner according to real-time monitoring data, resulting in a decline in compensation effect over time.

[0006] In terms of compensation resource allocation and scheme determination, traditional methods often do not fully consider historical data and actual situations, making it difficult to achieve optimal resource allocation and adaptability of the compensation scheme, resulting in high compensation costs and poor results.

[0007] With the development of industrial automation and intelligentization, higher requirements are placed on the precision and efficiency of bung jar deformation compensation. Therefore, there is an urgent need for a method that can intelligently and efficiently compensate for bung jar deformation to solve the problems existing in the prior art and improve the safety and reliability of bung jars. SUMMARY

[0008] The purpose of the present application is to provide a bung jar shape deformation intelligent compensation method to solve the problems raised in the background art.

[0009] To achieve the above object, the present application provides the following technical solutions: a binaural ear-shaped deformation intelligent compensation method, the method comprising:

[0010] Collecting binaural ear structure parameters and working environment data, analyzing the binaural ear structure parameters and working environment data, constructing a deformation feature set, determining and outputting an initial compensation feature library;

[0011] According to the dynamic optimization operation feedback of the initial compensation feature library, determine the compensation model update information set;

[0012] Analyzing the compensation model update information set, performing hierarchical disassembly classification statistics on the compensation model update information set, and determining the target compensation parameter;

[0013] Obtaining a compensation resource configuration information set, determining and outputting an adaptive compensation scheme based on the target compensation parameter based on the compensation resource configuration information set.

[0014] Preferably, the analyzing the binaural ear structure parameters and working environment data, constructing a deformation feature set, determining and outputting an initial compensation feature library comprises:

[0015] Analyzing the binaural ear structure parameters and working environment data, extracting a structure geometric feature group, a mechanical distribution feature and a temperature change environment feature;

[0016] Analyzing the structure geometric feature group, determining the position correlation weight and the structure influence score corresponding to each feature point;

[0017] According to the load cycle period in the working environment data, determining a binaural ear key observation stage, and according to the load cycle period and the binaural ear key observation stage, determining a deformation attenuation coefficient;

[0018] According to the position correlation weight, the structure influence score, the deformation attenuation coefficient and the mechanical distribution feature, constructing a multi-dimensional feature vector;

[0019] According to the multi-dimensional feature vector, performing matching retrieval on a preset historical deformation case database, determining and outputting the initial compensation feature library.

[0020] Preferably, the constructing a multi-dimensional feature vector according to the position correlation weight, the structure influence score, the deformation attenuation coefficient and the mechanical distribution feature comprises:

[0021] According to the size deviation degree of the structure geometric feature, determining a structure weight parameter, according to the stress concentration degree of the mechanical distribution feature, determining a mechanical weight parameter, and according to the change rate of the temperature change environment feature, determining an environmental weight parameter;

[0022] The structure weight parameter, the mechanical weight parameter, the environment weight parameter and the deformation attenuation coefficient are weighted and fused to form a composite feature vector including a structure dimension, a mechanical dimension and an environment dimension.

[0023] Preferably, the dynamic optimization operation according to the initial compensation feature library feedback determines a compensation model update information set, including:

[0024] According to the initial compensation feature library, a real-time deformation visualization interface is provided for the deformation monitoring system, and the feature space mapping effect is updated synchronously according to the monitoring data trajectory fed back by the system;

[0025] According to the monitoring data trajectory, potential deformation characteristics of the double ear pot are identified through mode clustering analysis, the sorting priority of the initial features is dynamically adjusted, and a feature update scheme is generated;

[0026] The feature update scheme is analyzed, the stability of the compensation model is evaluated, the model stability feedback information is determined, the corresponding feature optimization suggestion is fed back to the control end, and an adjustment confirmation instruction is received;

[0027] According to the adjustment confirmation instruction, the compensation model update information set is determined.

[0028] Preferably, the feature update scheme is analyzed, the stability of the compensation model is evaluated, the model stability feedback information is determined, including:

[0029] According to the feature update scheme, the feature dimension number, the compensation confidence interval, the historical compensation result distribution and the feature importance sorting of the model are extracted;

[0030] According to the structure attribute of each feature, the compensation error tolerance threshold, the historical compensation accuracy ratio and the continuous compensation consistency rate of the model are determined;

[0031] Based on the feature dimension number, the compensation confidence interval and the historical compensation result distribution, according to the feature importance sorting and the compensation error tolerance threshold, a statistical stability judgment condition is constructed;

[0032] According to the load cycle period and the key observation stage of the double ear pot, the effective time range of model compensation is determined;

[0033] Based on the historical compensation accuracy ratio and the effective time range, according to the continuous compensation consistency rate and the feature importance sorting, a time sequence stability judgment condition is constructed;

[0034] According to the statistical stability judgment condition and the time sequence stability judgment condition, it is judged whether the statistical stability or the time sequence stability corresponding to the current feature update scheme meets the requirements, and a judgment conclusion is determined as the model stability feedback information.

[0035] Preferably, the feature dimension number, the compensation confidence interval and the historical compensation result distribution are used to construct a statistical stability judgment condition according to the feature importance ranking and the compensation error tolerance threshold, specifically including:

[0036] The correlation between the feature dimension number and the compensation confidence interval is calculated, the matching degree between the historical compensation result distribution and the feature importance ranking is analyzed, and the critical range of statistical stability is set in combination with the compensation error tolerance threshold;

[0037] The historical compensation accuracy proportion and the effective time range are used to construct a time sequence stability judgment condition according to the continuous compensation consistency rate and the feature importance ranking, specifically including:

[0038] The fluctuation amplitude of the historical compensation accuracy proportion within the effective time range is calculated, the time sequence correlation between the continuous compensation consistency rate and the feature importance ranking is analyzed, and the critical range of time sequence stability is set.

[0039] Preferably, the compensation model update information set is analyzed, the compensation model update information set is hierarchically disassembled, classified and counted, and the target compensation parameter is determined, including:

[0040] According to the compensation model update information set, the key feature optimization set and the auxiliary parameter adjustment set are counted;

[0041] According to the model update period set by the institution and the load cycle period corresponding to different double-ear cans in the compensation model update information set, a plurality of update batches are divided, and an update batch information set is determined;

[0042] Based on the update batch information set, the key feature optimization set is analyzed, and according to the parameter drift proportion of each feature in the historical update record corresponding to each feature in the key feature optimization set, the parameter redundancy of each feature corresponding to each update batch is determined;

[0043] According to the auxiliary parameter adjustment set, the update batch information set and the parameter redundancy, the target compensation parameter is constructed.

[0044] Preferably, the compensation resource configuration information set is used to determine and output an adaptive compensation scheme according to the target compensation parameter, including:

[0045] Based on the compensation resource configuration information set, the target compensation parameter is analyzed, and a plurality of target compensation resources corresponding to each feature optimization item in the target compensation parameter are determined;

[0046] According to the compensation resource configuration information set, the historical cooperation record corresponding to each target compensation resource is extracted to determine the historical application times, the historical optimization cost, the historical compensation accuracy ratio and the historical execution risk of the current mechanism under the corresponding target compensation resource for each feature optimization item;

[0047] According to the historical application times, the historical optimization cost, the historical compensation accuracy ratio and the historical execution risk of each feature optimization item under the corresponding target compensation resource, a comprehensive cost optimization process is performed on each feature optimization item to determine the optimal compensation resource corresponding to each feature optimization item, so as to construct and output the adaptive compensation scheme.

[0048] Preferably, the comprehensive cost optimization process on each feature optimization item according to the historical application times, the historical optimization cost, the historical compensation accuracy ratio and the historical execution risk of each feature optimization item under the corresponding target compensation resource specifically includes:

[0049] The weighted sum of the historical application times and the historical optimization cost is calculated as a resource use cost indicator, and the weighted sum of the historical compensation accuracy ratio and the historical execution risk is calculated as an effect risk cost indicator;

[0050] The resource use cost indicator and the effect risk cost indicator are comprehensively compared, and the compensation resource with the lowest comprehensive cost is selected as the optimal resource.

[0051] Preferably, the potential deformation feature of the double-ear tank is identified by mode clustering analysis, which includes:

[0052] The monitoring data trajectory is divided into time windows, and the deformation fluctuation amplitude and frequency in each time window are extracted;

[0053] The initial classification standard of the clustering algorithm is set as the historical high compensation demand case feature library, and the clustering iteration is performed according to the matching degree of the deformation fluctuation amplitude and frequency with the initial classification standard;

[0054] The reliability of the clustering result is evaluated by classification purity, and the clustering group with a matching degree exceeding a preset critical value with the historical high compensation feature is selected, and the corresponding feature combination is determined as the potential deformation feature of the double-ear tank.

[0055] Compared with the prior art, the beneficial effects of the present application are:

[0056] The method collects the double-ear tank structure parameters and the working environment data, analyzes and constructs the deformation feature set and determines the initial compensation feature library, which lays a solid foundation for subsequent compensation work. This comprehensive data collection and analysis method can fully consider the structural characteristics of the double-ear tank and the working environment factors, so that the initial compensation feature library is more accurate and perfect, thereby improving the prediction and compensation accuracy of the deformation.

[0057] In terms of updating the compensation model, the dynamic optimization operation according to the initial compensation feature library feedback is used to determine the compensation model update information set, which is hierarchically disassembled, classified and counted to determine the target compensation parameter. This process realizes the dynamic updating and optimization of the compensation model, can timely adjust the model according to real-time monitoring data, adapt to the change of the binaural pot shape, and ensure the stability and reliability of the compensation effect. Through pattern clustering analysis to identify potential deformation characteristics and dynamically adjust the feature ranking priority, the adaptability of the model to complex deformation conditions is further improved.

[0058] In terms of determining the compensation scheme, the compensation resource configuration information set is obtained, and the adaptive compensation scheme is determined and output according to the target compensation parameter. Through analysis of historical collaboration records, including historical application times, optimization cost, compensation accuracy ratio and execution risk, etc., each feature optimization item is comprehensively optimized in terms of cost, and the optimal compensation resource is selected, realizing the optimal allocation of resources, reducing the compensation cost, and improving the effectiveness and reliability of the compensation scheme.

[0059] When constructing the multi-dimensional feature vector, the method considers the weighted fusion of structure weight parameters, mechanical weight parameters, environmental weight parameters and deformation decay coefficients, comprehensively describes the deformation characteristics of the binaural pot from multiple dimensions, and further improves the accuracy and reliability of the model. When evaluating the stability of the compensation model, the statistical stability judgment condition and the time sequence stability judgment condition are constructed to evaluate the stability of the model from the statistical and time sequence aspects, so that the model can maintain stable compensation effect under different conditions.

[0060] By dividing the update batches and determining the parameter redundancy, the updating of the compensation model is more orderly and efficient, which can adapt to the load cycle period and model update period of different binaural pots, and improve the adaptability and flexibility of the system. In summary, the method of the present application can comprehensively and accurately compensate the deformation of the binaural pot, improve the compensation precision and efficiency, reduce the compensation cost, and enhance the stability and adaptability of the system. BRIEF DESCRIPTION OF DRAWINGS

[0061] Figure 1 The working principle diagram of the binaural pot deformation intelligent compensation method described in the present application;

[0062] Figure 2 The design diagram for constructing the initial compensation feature library;

[0063] Figure 3 The design diagram for dynamic updating of the compensation model;

[0064] Figure 4 The design diagram for model stability evaluation;

[0065] Figure 5 design diagram generated for the adaptive compensation scheme. DETAILED DESCRIPTION

[0066] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0067] Please refer to Figures 1-5 The intelligent compensation method for the bellow-shaped ear can is implemented as follows:

[0068] The structure parameters of the bellow-shaped ear can and the operation environment data are collected, the structure parameters of the bellow-shaped ear can and the operation environment data are analyzed, a deformation feature set is constructed, and an initial compensation feature library is determined and output. The structure parameters include the geometric dimensions of the bellow-shaped ear can, material properties, etc., and the operation environment data covers temperature changes, load conditions, etc. Through analysis of these data, relevant features are extracted.

[0069] According to the dynamic optimization operation feedback of the initial compensation feature library, a compensation model update information set is determined. The initial compensation feature library is optimized and adjusted according to the feedback in actual application, thereby generating the information set required for model updating.

[0070] The compensation model update information set is analyzed, the compensation model update information set is classified and counted by hierarchical disassembly, and target compensation parameters are determined. Through in-depth analysis and classification of the update information set, the target compensation parameters that need to be adjusted are determined.

[0071] An adaptive compensation scheme is determined and output based on the target compensation parameters and the compensation resource configuration information set. In combination with the configuration of the compensation resources and the target compensation parameters, a suitable compensation scheme is formulated.

[0072] In the embodiment 1, the structure parameters of the bellow-shaped ear can and the operation environment data are comprehensively collected and analyzed in the implementation of constructing the deformation feature set and determining the initial compensation feature library. The structure parameters of the bellow-shaped ear can cover geometric dimension parameters, such as the height and diameter of the can body, the position coordinates of the ears, the connection angle between the ears and the can body, and other specific structural dimension information, and also include material attribute parameters, such as the elastic modulus, yield strength, and Poisson's ratio of the material, and other physical property data. The operation environment data includes temperature change data, such as the real-time monitoring value of the environmental temperature, the rate and period of temperature change, etc., and load condition data, such as the size of the load applied on the bellow-shaped ear can, the type of the load (static load or dynamic load), the load cycle period, etc.

[0073] After analyzing these data, the extraction of the structural geometric feature group, the mechanical distribution feature, and the temperature variation environment feature begins. For the extraction of the structural geometric feature group, a detailed analysis of each structural part of the double-ear can is required. Taking the can body as an example, the curvature radius of the cylindrical surface, the uniformity of the wall thickness, and other features are extracted; for the double-ear structure, the cross-sectional shape of the ear, the transition fillet size at the connection between the ear and the can body, and other features are extracted. When extracting the mechanical distribution feature, through finite element analysis and other methods, the stress distribution cloud diagram of the double-ear can under different load conditions is obtained, and then the position of the stress concentration area, the stress value, and other features are extracted, as well as the strain distribution features such as the strain size and strain gradient of each part. The extraction of the temperature variation environment feature is to process the temperature monitoring data to obtain the temperature variation curve with time, and then extract the peak value, valley value, and change rate of the temperature variation.

[0074] The structural geometric feature group is analyzed in depth to determine the position correlation weight and the structure influence score corresponding to each feature point. When determining the position correlation weight, the importance of the feature point position in the overall structure of the double-ear can is considered. For example, the connection between the double-ear and the can body is a key stress part, and the position correlation weight of the feature point at this part is relatively high; while the center area at the bottom of the can body is relatively small under general load conditions, and the position correlation weight of the feature point at this part is relatively low. When determining the structure influence score, the influence of the geometric size change of each feature point on the overall deformation of the double-ear can is analyzed. For example, the cross-sectional size change of the double-ear directly affects its carrying capacity, and thus has a greater impact on the deformation of the can body, so the structure influence score of this feature point is higher; while a slight size deviation at a non-critical part of the can body has a smaller impact on the overall deformation, so its structure influence score is lower.

[0075] According to the load cycle period in the operating environment data, the key observation stage of the double-ear can is determined. The load cycle period refers to the time taken for the load to return to the initial state after undergoing a series of changes. Within the load cycle period, the stress state of the double-ear can changes periodically, and the deformation in different stages is also different. By analyzing the load cycle period, it is determined which stages the deformation of the double-ear can is likely to be more significant, and these stages are the key observation stages. For example, the stage when the load reaches the maximum value and the stage when the load starts to unload from the maximum value may be the key observation stages. Then, according to the load cycle period and the key observation stage of the double-ear can, the deformation attenuation coefficient is determined. The deformation attenuation coefficient is used to describe the attenuation law of the deformation of the double-ear can over time after experiencing the load cycle. Through the analysis of the historical monitoring data, a relationship model between the load cycle period, the key observation stage, and the deformation attenuation coefficient is established, and thus the corresponding deformation attenuation coefficient is determined.

[0076] After obtaining the position correlation weight, structure influence score, deformation attenuation coefficient and mechanical distribution characteristics, the multi-dimensional feature vector is constructed. The multi-dimensional feature vector contains feature information of multiple dimensions and can comprehensively describe the deformation characteristics of the double-ear tank. The position correlation weight, structure influence score, deformation attenuation coefficient and stress, strain and other characteristics in the mechanical distribution characteristics are taken as the components of the vector to form a multi-dimensional vector space. The position and weight of each feature in the vector are determined according to the influence degree of the feature on the deformation of the double-ear tank, so that the multi-dimensional feature vector can accurately reflect the actual deformation characteristics of the double-ear tank.

[0077] According to the constructed multi-dimensional feature vector, the preset historical deformation case database is matched and searched. The preset historical deformation case database stores a large number of double-ear tank deformation cases processed in the past, each of which contains a corresponding multi-dimensional feature vector and a corresponding compensation scheme. By calculating the similarity between the multi-dimensional feature vector constructed at present and the case feature vector in the database, a historical case with high similarity is found. According to the compensation scheme corresponding to the similar case, an initial compensation feature library is determined and output. The initial compensation feature library contains preliminary compensation strategies and related parameters for the current double-ear tank deformation condition.

[0078] In the implementation of constructing the multi-dimensional feature vector, the structure weight parameter needs to be determined according to the size deviation degree of the structure geometric feature. The size deviation of the structure geometric feature refers to the difference between the actual structure size of the double-ear tank and the design size. For different structure parts, the influence degree of the size deviation on the overall deformation is different. For example, the wall thickness deviation of the tank body will directly affect its carrying capacity. If the wall thickness is less than the design value, the tank body may be more prone to deformation under the action of load, so the greater the size deviation degree of the structure feature, the higher the corresponding structure weight parameter; the size deviation of a non-key part at the bottom of the double-ear tank has relatively small influence on the overall deformation, and the increase of the structure weight parameter is also relatively low. When determining the structure weight parameter, a quantitative relationship between the size deviation degree and the weight parameter needs to be established, and through analysis and summary of a large amount of historical data, a corresponding mapping rule is formed, so that the structure weight parameter can accurately reflect the influence degree of the size deviation on the deformation.

[0079] The stress concentration degree according to the mechanical distribution characteristic is used to determine the mechanical weight parameter. The stress concentration refers to the phenomenon that the stress value at some positions of the double-ear tank structure is much higher than the average stress due to factors such as geometric shape and load action. The higher the stress concentration degree of a region, the easier the plastic deformation or even the fracture occurs, and the greater the influence on the deformation of the double-ear tank. When determining the mechanical weight parameter, the stress distribution of the double-ear tank under different working conditions needs to be obtained through finite element analysis and other methods, the stress concentrated regions are identified, and the stress concentration coefficients of the regions are calculated. The stress concentration coefficient is an important index for measuring the stress concentration degree, and the greater the value, the higher the stress concentration degree. Then, the corresponding mechanical weight parameter is determined according to the size of the stress concentration coefficient. For example, the connection between the double ears and the tank body is often a region with obvious stress concentration, if the stress concentration coefficient of the region is large, the mechanical weight parameter of the region is also large accordingly; while the stress distribution of the main part of the tank body is relatively uniform, the stress concentration coefficient is small, and the mechanical weight parameter is also low. In this way, the mechanical weight parameter can accurately reflect the influence of the mechanical distribution characteristic on the deformation of the double-ear tank.

[0080] The change rate of the temperature change environment characteristic is used to determine the environmental weight parameter. The change rate of the temperature change environment characteristic refers to the degree of change of the environmental temperature with time. The rapid change of the temperature will cause the thermal expansion and cold shrinkage effect of the material of the double-ear tank, when the change rate of the temperature is too fast, a large thermal stress will be generated in the material, thereby causing the deformation of the tank body. Different change rates of the temperature have different influences on the deformation of the double-ear tank, the faster the change rate, the greater the influence on the deformation, and the higher the corresponding environmental weight parameter. When determining the environmental weight parameter, the temperature monitoring data needs to be processed to calculate the change amount of the temperature in a unit time, i.e. the change rate of the temperature. Then, the corresponding environmental weight parameter is determined according to the size of the change rate of the temperature through a preset rule or model. For example, when the change rate of the temperature exceeds a certain threshold, the environmental weight parameter will have a large increase; while when the change rate of the temperature is small, the increase of the environmental weight parameter is also small.

[0081] After obtaining the structure weight parameter, the mechanical weight parameter, the environmental weight parameter, and the previously determined deformation attenuation coefficient, these parameters need to be weighted and fused to form a composite feature vector containing structure, mechanics, and environment dimensions. The process of weighted fusion needs to consider the influence weight of each parameter on the deformation of the binaural pot. According to different application scenarios and actual needs, appropriate weight coefficients are assigned to each parameter. The structure dimension mainly reflects the influence of structure geometric characteristics on deformation, the mechanical dimension reflects the role of mechanical distribution characteristics, the environmental dimension considers the influence of temperature change environment characteristics, and the deformation attenuation coefficient describes the attenuation law of deformation over time. During fusion, the structure weight parameter, the mechanical weight parameter, and the environmental weight parameter are multiplied by their respective weight coefficients, and then combined with the deformation attenuation coefficient to form a multi-dimensional vector. For example, the weight coefficient of the structure weight parameter can be set to 0.3, the weight coefficient of the mechanical weight parameter is 0.4, the weight coefficient of the environmental weight parameter is 0.2, and the weight coefficient of the deformation attenuation coefficient is 0.1. Through such weight distribution, the composite feature vector can comprehensively consider the factors of each dimension and fully describe the deformation characteristics of the binaural pot.

[0082] During the weighted fusion process, attention needs to be paid to the rationality and scientificity of the weight coefficients. The determination of the weight coefficients can be achieved through expert experience, historical data statistical analysis, etc. For example, by analyzing a large number of binaural pot deformation cases, the influence proportion of structure, mechanics, environment, etc. on deformation under different working conditions is determined, thereby providing a basis for setting the weight coefficients. At the same time, the correlation between each parameter needs to be considered to avoid repeated calculation or omission of important information.

[0083] After forming the composite feature vector, it needs to be verified and optimized. The composite feature vector can be applied to known binaural pot deformation cases to observe its description effect on deformation characteristics. If the composite feature vector can accurately reflect the deformation characteristics in the case, it means that the setting of the weight coefficients and the fusion process are reasonable; if there is deviation, the weight coefficients need to be adjusted and optimized until the composite feature vector can meet the requirements.

[0084] In the embodiment of determining the compensation model update information set in the dynamic optimization operation according to the initial compensation feature library feedback, a real-time deformation visualization interface needs to be built for the deformation monitoring system based on the initial compensation feature library. This interface needs to integrate the bellow structure parameters and real-time monitoring data, and present the bellow deformation state in the form of three-dimensional model or dynamic curve graph, etc. For example, the deformation level of different regions is marked by color coding, the red region represents the part with larger deformation level, and the green region represents the part with smaller deformation level, while the displacement data of key feature points, stress change trend and other information are displayed in real time on the interface. When the system obtains a new monitoring data track, the feature space mapping effect needs to be updated synchronously, that is, the distribution of feature vectors in the multi-dimensional space is adjusted according to the new data, so that the visualization interface can reflect the latest deformation features of the bellow in real time, and provide intuitive data support for subsequent dynamic optimization.

[0085] The mode clustering analysis is performed on the monitoring data track to identify potential deformation features. In specific implementation, the monitoring data track is first divided into time windows, and the length of the time window needs to be determined according to the load cycle period and the key observation stage. For example, if the load cycle period is 10 minutes, the time window can be set to 1 minute, so as to capture the deformation features in different stages within each cycle. In each time window, the deformation fluctuation amplitude and frequency features are extracted, the deformation fluctuation amplitude refers to the difference between the maximum and minimum values of the deformation level in the time period, and the frequency refers to the number of times of significant changes in the deformation level.

[0086] The initial classification standard of the clustering algorithm is set as the historical high compensation demand case feature library, which stores the feature vectors of the cases that need high compensation in the past. The matching degree of the deformation fluctuation amplitude and frequency extracted in the current time window with the initial classification standard is calculated for clustering iteration. The matching degree can be calculated by using the Euclidean distance, cosine similarity and other methods, for example, for two feature vectors, the square root of the sum of squares of the difference values of corresponding dimensions is calculated, and the smaller the value is, the higher the matching degree is. In the iteration process, the clustering center is adjusted constantly until the clustering result tends to be stable or the preset iteration number is reached.

[0087] The reliability of the clustering result is evaluated by the classification purity, which refers to the proportion of samples belonging to the same category in each cluster. For example, if more than 80% of the samples in a cluster belong to the historical high compensation demand case, it is considered that the purity of the cluster is high. The cluster group with a matching degree higher than a preset threshold with the historical high compensation feature is selected, and the preset threshold can be set according to actual application requirements, such as 0.7, and the corresponding feature combination is determined as the potential deformation feature of the bellow. These potential features may be caused by new load working conditions, environmental changes and other factors, which are not covered by the initial compensation feature library.

[0088] After identifying the potential deformation features, the initial feature ranking priority needs to be dynamically adjusted and a feature update scheme is generated. The initial feature ranking priority is originally determined based on factors such as structural geometric features, mechanical distribution features, etc. When potential deformation features are found, the influence of these new features on the deformation of the binaural ear cup needs to be evaluated. For example, if a potential feature has a larger deformation variable and a higher frequency of occurrence, its priority should be raised to a higher position in the initial features. By adjusting the ranking priority of the features, the compensation model can pay more attention to features that have a greater impact on deformation, thereby improving the accuracy of compensation. The feature update scheme needs to clearly list the potential features to be added, the adjusted feature priority, and the corresponding parameter settings, etc.

[0089] The feature update scheme is analyzed and the stability of the compensation model is evaluated. First, the feature dimension number, compensation confidence interval, historical compensation result distribution, and feature importance ranking of the model are extracted. The feature dimension number refers to the number of features considered by the model, the compensation confidence interval represents the reliable range of the compensation result, the historical compensation result distribution reflects the effectiveness of past compensation, and the feature importance ranking reflects the degree of influence of each feature on compensation.

[0090] According to the structural attributes of each feature, the compensation error tolerance threshold, historical compensation accuracy ratio, and continuous compensation consistency rate of the model are determined. The compensation error tolerance threshold refers to the maximum allowable compensation error range, for example, for the deformation variable of the key part, the error tolerance threshold can be set to ±0.1 mm; the historical compensation accuracy ratio refers to the proportion of cases within the tolerance threshold in historical compensation; the continuous compensation consistency rate refers to the proportion of consistent results in consecutive compensation.

[0091] Based on the feature dimension number, compensation confidence interval, and historical compensation result distribution, combined with the feature importance ranking and compensation error tolerance threshold, the statistical stability judgment condition is constructed. For example, the correlation between the feature dimension number and the compensation confidence interval is calculated, if the compensation confidence interval decreases when the feature dimension number increases, it means that the stability of the model may improve; the matching degree between the historical compensation result distribution and the feature importance ranking is analyzed, if the compensation result deviation of important features is large, it may affect the stability of the model. By setting the critical range of statistical stability, such as the feature dimension number being between 5-10, the compensation confidence interval not exceeding ±5%, etc., the statistical stability of the model is judged.

[0092] According to the load cycle period and the key observation stage of the double-ear tank, the effective time range of model compensation is determined. For example, in the key observation stage within the load cycle period, the effective time range of model compensation is the starting time to the ending time of the stage. Based on the historical compensation accuracy ratio and the effective time range, combined with the continuous compensation consistency rate and the feature importance ranking, the time sequence stability judgment condition is constructed. The fluctuation amplitude of the historical compensation accuracy ratio in the effective time range is calculated. If the fluctuation amplitude is small, it indicates that the stability of the model in the time sequence is good. The time sequence correlation of the continuous compensation consistency rate and the feature importance ranking is analyzed. If the continuous compensation consistency rate of the important features is high, the time sequence stability of the model is high. The critical range of time sequence stability is set, such as the fluctuation amplitude of the historical compensation accuracy ratio is not more than 10%, the continuous compensation consistency rate is not less than 80%, etc., to judge the time sequence stability of the model.

[0093] According to the statistical stability judgment condition and the time sequence stability judgment condition, it is judged whether the statistical stability or the time sequence stability corresponding to the current feature update scheme meets the requirements. If it meets the requirements, it is considered that the model after updating has good stability; if it does not meet the requirements, the feature update scheme needs to be adjusted. The judgment conclusion is used as the model stability feedback information, and the corresponding feature optimization suggestion is fed back to the control end, such as suggesting to increase the weight of a certain feature or adjust the setting of a certain parameter, etc.

[0094] After the control end receives the adjustment confirmation instruction, the compensation model update information set is determined according to the instruction. The compensation model update information set needs to include the feature update scheme, the model stability feedback information, and the adjusted parameter setting, etc.

[0095] In the embodiment of evaluating the stability of the compensation model, the feature dimension number, the compensation confidence interval, the historical compensation result distribution, and the feature importance ranking of the model need to be extracted according to the feature update scheme. The feature dimension number refers to the total number of feature parameters included in the current compensation model, such as the specific number of different dimensions of features including structural geometric features, mechanical distribution features, and temperature change environment features, etc. The compensation confidence interval is a range index for measuring the reliability of the compensation result, such as the compensation value of a certain deformation variable has a confidence interval of ±0.5mm, indicating that the actual deformation variable has a high probability of falling within the interval. The historical compensation result distribution is a statistical analysis of all past compensation cases, showing the distribution state such as the case number ratio under different compensation accuracy. The feature importance ranking is a priority ranking according to the influence of each feature on the compensation result, for example, the stress concentration feature of the double-ear connection may be ranked higher, and the size feature of the non-key part of the tank is ranked relatively later.

[0096] The compensation error tolerance threshold, historical compensation accuracy ratio and continuous compensation consistency rate of the model are determined according to the structural properties of each feature. The compensation error tolerance threshold needs to be set in combination with the functional importance and structural safety of the part where the feature is located. For example, for the double-ear part directly bearing the main load, the compensation error tolerance threshold of the deformation variable may be set to ±0.1 mm, and the threshold of the non-stress key area of the side wall of the tank can be relaxed to ±0.5 mm. The historical compensation accuracy ratio is obtained by counting the ratio of the number of cases whose compensation error is within the corresponding tolerance threshold to the total number of cases in the historical compensation cases. For example, in 100 historical cases, 85 cases have compensation errors within the threshold range, and the accuracy ratio is 85%. The continuous compensation consistency rate is the ratio of the cases whose results are consistent with each other in the same feature in continuous compensation. For example, a feature has consistent results in 9 out of 10 consecutive compensations, and the consistency rate is 90%.

[0097] Based on the number of feature dimensions, compensation confidence interval and historical compensation result distribution, combined with feature importance ranking and compensation error tolerance threshold, a statistical stability judgment condition is constructed. In specific implementation, first, the correlation between the number of feature dimensions and the compensation confidence interval is calculated. For example, by analyzing historical data, it is found that when the number of feature dimensions is 8-12, the fluctuation range of the compensation confidence interval is the smallest, and thus it is determined that this number interval is the optimal range. At the same time, the matching degree of the historical compensation result distribution and the feature importance ranking is analyzed. If the proportion of the features with high importance ranking showing large deviation in the compensation result distribution is high, it indicates that the statistical stability of the model may have problems. Combined with the compensation error tolerance threshold, the critical range of statistical stability is set, for example, the number of feature dimensions needs to be kept in the range of 6-15, the upper limit of the compensation confidence interval should not exceed ±8%, and the deviation rate of important features in the historical compensation result should be less than 15%. When these conditions are met at the same time, it can be determined that the statistical stability of the model meets the requirements.

[0098] At the same time, according to the load cycle period and the key observation stage of the double-ear tank, the effective time range of the model compensation is determined. The load cycle period refers to the time span from the beginning to the end of a complete cycle of the load borne by the double-ear tank. For example, the load cycle period of a certain double-ear tank is 24 hours, and the peak load stage (such as 8:00-10:00 every day) is the key observation stage. The effective time range of the model compensation is the specific time period of the peak stage.

[0099] Based on the historical compensation accuracy ratio and the effective time range, combined with the continuous compensation consistency rate and the feature importance ranking, the time sequence stability judgment condition is constructed. First, the fluctuation range of the historical compensation accuracy ratio in the effective time range is calculated, for example, in the key observation stage of the past 10 load cycle periods, the historical compensation accuracy ratio fluctuates between 75%-85%, and the fluctuation range is 10%. Then analyze the time sequence correlation of continuous compensation consistency rate and feature importance ranking. If the consistency rate of important features in continuous compensation shows a significant downward trend, it may indicate that the stability of the model in time sequence is poor. By setting the critical range of time sequence stability, such as the fluctuation range of the historical compensation accuracy ratio in the effective time range does not exceed 15%, the continuous compensation consistency rate on important features is not less than 70%, and the overall consistency rate is not less than 80%, etc., which is used as the judgment basis of time sequence stability.

[0100] After constructing the statistical stability judgment condition and the time sequence stability judgment condition, the stability of the current feature update scheme needs to be comprehensively judged. Check whether each critical index of statistical stability meets the requirements, such as whether the feature dimension number is within the preset interval, whether the compensation confidence interval exceeds the upper limit, whether the bias rate of important features in the historical compensation result meets the requirements, etc. At the same time, check the conditions of time sequence stability, such as whether the fluctuation of accuracy ratio in the effective time range is within the allowed range, whether the continuous compensation consistency rate reaches the critical value, etc.

[0101] If the statistical stability judgment condition and the time sequence stability judgment condition meet the requirements, it is concluded that the current feature update scheme will not have a significant negative impact on the model stability; if one or more conditions do not meet the requirements, it is determined that the model stability is at risk, and the feature update scheme needs to be further adjusted. Finally, the judgment conclusion is used as the model stability feedback information, which needs to clearly indicate the stability state of the model in the statistical level and the time sequence level, as well as the specific factors that may affect the stability, such as too many feature dimensions or the continuous compensation consistency rate of an important feature is too low, etc.

[0102] During the entire evaluation process, the accuracy of data extraction needs to be ensured, such as the distribution of historical compensation results needs to be based on a sufficient number of valid cases, and the importance of features needs to be scientifically evaluated in combination with actual stress analysis and structural functions. At the same time, the setting of critical range needs to refer to industry standards, historical experience and specific use scenarios of the binaural earphone, to ensure the scientificity and rationality of stability evaluation. In addition, when it is found that the stability does not meet the requirements, the specific problem points in the feature update scheme need to be traced back in time, such as whether too many non-key features are added, causing the number of feature dimensions to exceed the reasonable range, or whether the parameter setting of an important feature is adjusted, causing the continuous compensation consistency rate to decrease, etc., to provide accurate basis for subsequent feature optimization suggestions, so as to ensure that the compensation model can maintain good stability after updating and continuously provide reliable support for binaural earphone deformation compensation.

[0103] In the implementation of determining the adaptive compensation scheme based on the compensation resource configuration information set and the target compensation parameter, the target compensation parameter corresponding to each feature optimization item needs to be determined based on the compensation resource configuration information set. The compensation resource configuration information set includes information such as the type, specification, quantity and technical parameters of the compensation resources available in the organization, such as compensation pads, stress release devices and temperature control equipment of different models. The target compensation parameter is obtained by hierarchical disassembly and classification statistics of the compensation model update information set, involving structural geometric feature optimization parameters, mechanical distribution adjustment parameters and temperature change environment adaptation parameters. Taking a feature optimization item "stress concentration compensation at the binaural connection" as an example, the corresponding target compensation resources after analysis may include wedge-shaped compensation pads A, flexible stress release sheets B and local heating control devices C, etc., which all have potential compensation ability for this feature optimization item.

[0104] From the compensation resource configuration information set, the historical cooperation records corresponding to each target compensation resource are extracted to determine the historical application times, historical optimization cost, historical compensation accuracy ratio and historical execution risk of the current organization for each feature optimization item under the corresponding target compensation resource. The historical cooperation record is a summary of past cases using the resource for compensation, for example, the target compensation resource "wedge-shaped compensation pad A" in the feature optimization item "stress concentration compensation at the binaural connection", the historical application times is 50 times, the material cost, labor cost, etc. of each application constitutes the historical optimization cost, among which 40 times of compensation results meet the error tolerance threshold, the historical compensation accuracy ratio is 80%, and 3 times of installation position deviation has occurred in the past application. Historical execution risk events. These data need to be systematically sorted and counted through the historical project archives, resource use account book, etc. of the organization, to ensure the integrity and accuracy of the data.

[0105] After obtaining the historical application times, historical optimization cost, historical compensation accuracy ratio and historical execution risk, a comprehensive cost optimization processing is performed on each feature optimization item. In specific implementation, a weighted sum of the historical application times and the historical optimization cost is calculated as a resource use cost indicator. The historical application times reflect the popularity and maturity of the resource in actual application, and a resource with more application times usually has advantages in procurement cost and operation proficiency; the historical optimization cost directly reflects the economic investment of using the resource. For example, if the weight of the historical application times is set to 0.3, the weight of the historical optimization cost is set to 0.7, the historical application times of a target compensation resource is 50 times (converted into a standardized value of 0.8), and the historical optimization cost is 1000 yuan (converted into a standardized value of 0.6), then the resource use cost indicator is 0.3*0.8+0.7*0.6=0.66.

[0106] A weighted sum of the historical compensation accuracy ratio and the historical execution risk is calculated as an effect risk cost indicator. The historical compensation accuracy ratio reflects the reliability of the compensation effect of the resource, and the higher the ratio, the more stable the effect; the historical execution risk reflects the operation risk, safety risk and the like that may be faced when using the resource, and the more the number of risk events, the higher the risk level. For example, if the weight of the historical compensation accuracy ratio is set to 0.6, the weight of the historical execution risk is set to 0.4, the historical compensation accuracy ratio of a resource is 80% (converted into a standardized value of 0.8), and the historical execution risk events are 3 times (converted into a standardized value of 0.3, and the lower the value, the higher the risk), then the effect risk cost indicator is 0.6*0.8+0.4*0.3=0.6.

[0107] The resource use cost indicator and the effect risk cost indicator are compared comprehensively, and the compensation resource with the lowest comprehensive cost is selected as the optimal resource. The comprehensive cost is calculated by weighting and aggregating the two indicators, for example, if the weight of the resource use cost indicator is set to 0.5, and the weight of the effect risk cost indicator is set to 0.5, then the comprehensive cost of a resource is 0.5*0.66+0.5*0.6=0.63. The comprehensive cost of all target compensation resources corresponding to each feature optimization item is calculated in this way, and the resource with the lowest comprehensive cost is selected as the optimal compensation resource of the feature optimization item.

[0108] In the conversion of the standardized values, a unified quantification standard needs to be established. For example, the historical application times can be standardized according to the ratio of the actual times to the maximum application times in the organization; the historical optimization cost can be standardized according to the ratio of the actual cost to the minimum cost (or cost interval classification); the historical compensation accuracy ratio can be directly converted into a standardized value of 0-1 according to the percentage value; and the historical execution risk can be converted into a standardized value according to the ratio of the number of risk events to the total application times (or risk level classification), and the more the risk events, the higher the standardized value (or the lower, which needs to be unified according to the definition of the indicator).

[0109] In the weight setting process, the application scene of the double-ear tank, the priority of compensation demand and other factors need to be combined. For example, for the industrial scene with extremely high safety requirements, the weights of historical compensation accuracy ratio and historical execution risk can be appropriately increased to ensure the reliability of compensation effect and operation safety; for the cost-sensitive scene, the weight of historical optimization cost can be increased to achieve economic optimization. The weight setting can be determined by expert evaluation, Delphi method and other methods to ensure the scientificity and rationality of weight distribution.

[0110] After determining the optimal compensation resources for each feature optimization item, these optimal resources are combined to construct an adaptive compensation scheme. The scheme needs to clearly indicate the optimal compensation resources corresponding to each feature optimization item, the use method of the resources, installation parameters, operation process and other contents. For example, for the "stress concentration compensation at the double-ear connection" feature optimization item, the optimal resource is "flexible stress relief piece B", and the scheme needs to specify the installation position, fixing method, pre-tightening force requirement and other specific parameters of the relief piece, as well as the matters needing attention and quality detection standards in the installation process.

[0111] The constructed adaptive compensation scheme needs to be verified for feasibility. The combination of resources in the scheme can be verified for whether it can achieve the expected compensation effect, whether there are resource conflicts or compatibility problems, through simulation, small-scale test and other ways. For example, through finite element simulation analysis, the stress concentration relief effect of using "flexible stress relief piece B" at the double-ear connection is verified to ensure that its deformation and stress distribution meet the design requirements.

[0112] In the whole implementation process, the accuracy and integrity of historical data need to be ensured to avoid calculation errors of comprehensive cost caused by data deviation. At the same time, the weight setting and standardization conversion method need to be consistent and interpretable to facilitate fair comparison between different feature optimization items and compensation resources. In addition, the adaptive compensation scheme needs to be operable and adjustable, which can be dynamically optimized according to the sudden situation in actual application or new monitoring data to ensure that the compensation effect on the deformation of the double-ear tank is always in the optimal state.

[0113] It should be noted that in this text, relationship terms such as first and second are only used to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the term "includes" "contains" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0114] While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary of the principles and application of the present application. Numerous modifications and adaptions can be effected without departing from the spirit and scope of the present application, which is not limited to the exact construction and arrangement described. It is intended, therefore, to cover all modifications and adaptions that fall within the scope of the claims and their equivalents.

Claims

1. A method for intelligent compensation of deformation in a double-eared can, characterized in that, The method comprises the following steps: Collecting double-ear tank structure parameters and working environment data, analyzing the double-ear tank structure parameters and working environment data, constructing a deformation feature set, determining and outputting an initial compensation feature library; According to the dynamic optimization operation feedback by the initial compensation feature library, determine the compensation model update information set; Analyzing the compensation model update information set, classifying and counting the compensation model update information set by hierarchical decomposition, and determining the target compensation parameter; Obtain the compensation resource configuration information set, based on the compensation resource configuration information set, according to the target compensation parameter, determine and output the adaptive compensation scheme; The analysis of the double-ear tank structure parameters and the working environment data, the construction of the deformation feature set, the determination and output of the initial compensation feature library, comprising: Analyzing the double-ear tank structure parameters and working environment data, extracting structure geometric feature group, mechanical distribution feature and temperature change environment feature; Analyzing the structure geometric feature group, determining the position correlation weight and structure influence score corresponding to each feature point; According to the load cycle period in the working environment data, determine the double-ear tank key observation stage, and according to the load cycle period and the double-ear tank key observation stage, determine the deformation attenuation coefficient; According to the position correlation weight, the structure influence score, the deformation attenuation coefficient and the mechanical distribution feature, a multi-dimensional feature vector is constructed; According to the multi-dimensional feature vector, the preset historical deformation case database is matched and searched, and the initial compensation feature library is determined and output.

2. The binaural mannequin compensation method of claim 1, wherein, The multi-dimensional feature vector constructed according to the position correlation weight, the structure influence score, the deformation attenuation coefficient and the mechanical distribution feature, specifically comprises: Determine the structure weight parameter according to the size deviation degree of the structure geometric feature, determine the mechanical weight parameter according to the stress concentration degree of the mechanical distribution feature, and determine the environment weight parameter according to the change rate of the temperature change environment feature; The structure weight parameter, the mechanical weight parameter, the environment weight parameter and the deformation attenuation coefficient are weighted and fused to form a composite feature vector containing structure dimension, mechanical dimension and environment dimension. According to the dynamic optimization operation feedback by the initial compensation feature library, determine the compensation model update information set, comprising:

3. The binaural mannequin compensation method of claim 2, wherein, According to the initial compensation feature library, provide a real-time deformation visualization interface for the deformation monitoring system, and update the feature space mapping effect synchronously according to the monitoring data trajectory feedback by the system; According to the monitoring data trajectory, identify the potential deformation features of the double-ear tank through pattern clustering analysis, dynamically adjust the sorting priority of the initial features, and generate a feature update scheme; Analyzing the feature update scheme, evaluating the stability of the compensation model, determining the model stability feedback information, and feeding back the corresponding feature optimization suggestions to the control end, receiving adjustment confirmation instructions; According to the adjustment confirmation instruction, determine the compensation model update information set. The analysis of the feature update scheme, the evaluation of the stability of the compensation model, and the determination of the model stability feedback information, comprising:

4. The binaural manhole shape variation compensation method of claim 3, wherein, According to the feature update scheme, extract the feature dimension number, compensation confidence interval, historical compensation result distribution and feature importance ranking of the model; ​ According to the structural attributes of each feature, determine the compensation error tolerance threshold, historical compensation accuracy ratio and continuous compensation consistency rate of the model; Based on the feature dimension number, compensation confidence interval and historical compensation result distribution, according to the feature importance ranking and compensation error tolerance threshold, construct a statistical stability judgment condition; According to the load cycle period and the key observation stage of the double-ear tank, determine the effective time range of model compensation; Based on the historical compensation accuracy ratio and the effective time range, according to the continuous compensation consistency rate and the feature importance ranking, construct a time sequence stability judgment condition; According to the statistical stability judgment condition and the time sequence stability judgment condition, judge whether the statistical stability or time sequence stability corresponding to the current feature update scheme meets the requirements, determine the judgment conclusion, and use it as the model stability feedback information.

5. The binaural manhole shape variation compensation method of claim 4, wherein, The statistical stability judgment condition is constructed based on the feature dimension number, compensation confidence interval and historical compensation result distribution, according to the feature importance ranking and compensation error tolerance threshold, and specifically includes: Calculate the correlation degree of feature dimension number and compensation confidence interval, analyze the matching degree of historical compensation result distribution and feature importance ranking, and set the critical range of statistical stability combined with compensation error tolerance threshold; The time sequence stability judgment condition is constructed based on the historical compensation accuracy ratio and the effective time range, according to the continuous compensation consistency rate and the feature importance ranking, and specifically includes: Calculate the fluctuation amplitude of historical compensation accuracy ratio within the effective time range, analyze the time sequence correlation of continuous compensation consistency rate and feature importance ranking, and set the critical range of time sequence stability.

6. The binaural manhole shape variation compensation method of claim 5, wherein, The compensation model update information set is analyzed, the compensation model update information set is hierarchically disassembled, classified and counted, and the target compensation parameter is determined, including: According to the compensation model update information set, count the key feature optimization set and auxiliary parameter adjustment set; According to the model update period set by the institution, according to the load cycle period corresponding to different double-ear tanks in the compensation model update information set, divide several update batches, and determine the update batch information set; Based on the update batch information set, analyze the key feature optimization set, and according to the parameter drift ratio of each feature in the historical update record corresponding to each feature in the key feature optimization set, determine the parameter redundancy of each feature corresponding to each update batch; According to the auxiliary parameter adjustment set, the update batch information set and the parameter redundancy, construct the target compensation parameter.

7. The binaural manhole shape variation compensation method of claim 6, wherein, Based on the compensation resource configuration information set, determine and output the adaptive compensation scheme according to the target compensation parameter, including: Based on the compensation resource configuration information set, analyze the target compensation parameter, and determine a plurality of target compensation resources corresponding to each feature optimization item in the target compensation parameter; According to the compensation resource configuration information set, historical cooperation records corresponding to each target compensation resource are extracted to determine the historical application times, historical optimization costs, historical compensation accuracy ratios and historical execution risks of the current mechanism for each feature optimization item under the corresponding target compensation resource; According to the historical application times, historical optimization costs, historical compensation accuracy ratios and historical execution risks of each feature optimization item under the corresponding target compensation resource, a comprehensive cost optimization process is performed on each feature optimization item to determine the optimal compensation resource corresponding to each feature optimization item, so as to construct and output the adaptive compensation scheme.

8. The binaural manhole shape variation compensation method of claim 7, wherein, The comprehensive cost optimization process on each feature optimization item according to the historical application times, historical optimization costs, historical compensation accuracy ratios and historical execution risks of each feature optimization item under the corresponding target compensation resource specifically includes: The weighted sum of the historical application times and the historical optimization costs is calculated as a resource use cost indicator, and the weighted sum of the historical compensation accuracy ratio and the historical execution risk is calculated as an effect risk cost indicator; The resource use cost indicator and the effect risk cost indicator are comprehensively compared, and the compensation resource with the lowest comprehensive cost is selected as the optimal resource.

9. The binaural mannequin variability compensation method of claim 8, wherein, The potential deformation feature of the double-ear pot is identified through mode clustering analysis, including: The monitoring data trajectory is divided into time windows, and the deformation fluctuation amplitude and frequency in each time window are extracted; The initial classification standard of the clustering algorithm is set as a historical high compensation demand case feature library, and clustering iteration is performed according to the matching degree of the deformation fluctuation amplitude and frequency with the initial classification standard; The reliability of the clustering result is evaluated through classification purity, and the clustering group with a matching degree exceeding a preset critical value with the historical high compensation feature is selected, and the corresponding feature combination is determined as the potential deformation feature of the double-ear pot.

Images