CN-121350484-B - Comprehensive environment adaptability test system for key performance of identification

Abstract

The invention relates to the technical field of environment adaptability test, and discloses a comprehensive environment adaptability test system for identifying key performance. The system comprises a multi-environment factor acquisition processing module, an environment scene dynamic modeling module, a performance index evaluation module, a failure mode prediction module, an environment tolerance analysis module, an integrated data modeling virtual model and a test scheme optimization module, wherein the multi-environment factor acquisition processing module acquires data and normalizes preprocessing, the environment scene dynamic modeling module constructs a multi-dimensional association map and updates a topological structure, the performance index evaluation module evaluates performance degradation indexes under a composite environment by means of a space-time feature fusion algorithm, the failure mode prediction module generates a potential failure mode under multi-environment coupling according to performance degradation trends and stress accumulation effects, the integrated data builds a virtual model and outputs a specific environment profile tolerance residual error, the test scheme optimization module generates a self-adaptive test sequence according to the residual error and the map, and the environment scene simulation control module combines multiple information to regulate test cabin parameters. The system can comprehensively and accurately evaluate the performance and tolerance of the equipment in a composite environment and has the characteristic of dynamic adjustment.

Inventors

• ZOU CHUANYU
• CHEN YONGQUAN

Assignees

• 中国标准化研究院

Dates

Publication Date

20260508

Application Date

20251024

Claims (7)

1. 1. An integrated environmental suitability test system for identifying key performance, comprising: the multi-environmental factor acquisition processing module is used for acquiring environmental data monitored by various sensors deployed in the test environmental cabin and executing normalization preprocessing; The environment scene dynamic modeling module is used for constructing a multidimensional environment association graph based on the preprocessed environment data and dynamically updating the topological structure of the multidimensional environment association graph according to the environment data flow; the performance index evaluation module is used for receiving the multidimensional environment association map and the preprocessed environment data and evaluating performance degradation indexes of the tested equipment under the composite environment stress through a space-time feature fusion algorithm; the failure mode prediction module generates a potential failure mode set of the tested equipment in the multi-environment coupling scene according to the evolution trend of the performance degradation index and the environmental stress accumulation effect; The environment tolerance analysis module integrates the performance degradation indexes and the potential failure mode set, builds an environment stress tolerance virtual model, and outputs tolerance residual errors of the tested equipment under a specific environment profile; the test scheme optimizing module generates a self-adaptive environment test scene sequence according to the tolerance residual error and the multidimensional environment association map; The environment scene simulation control module is used for dynamically regulating and controlling the parameter combination of the test environment cabin by combining the potential failure mode set, the self-adaptive environment test scene sequence and the tolerance residual error; the specific process of the performance index evaluation module for evaluating the performance degradation index comprises the following steps: Inputting a dynamic environmental stress propagation network and the preprocessed environmental data into a space-time feature fusion model, and extracting environmental stress spatial distribution features through a graph neural network layer; Capturing a long-term dependence characteristic of the evolution of the environmental stress along with time by utilizing an expansion convolution layer, and fusing the spatial distribution characteristic and the long-term dependence characteristic to form an environmental performance degradation characteristic vector; mapping the environmental performance degradation characteristic vector into a performance degradation index through a full connection layer, and outputting a key performance degradation curve of the tested equipment under the composite environmental stress; The specific process of constructing the environment stress tolerance virtual model by the environment tolerance analysis module comprises the following steps: Collecting physical properties and structural design parameters of materials of equipment to be tested, and establishing a multi-physical field coupling simulation model based on a finite element method; Inputting the performance degradation index and the potential failure mode set into the multi-physical field coupling simulation model, solving an internal stress distribution equation of the equipment, and outputting an environmental stress tolerance reference value; comparing the deviation of the tolerance reference value and the actually measured performance degradation index, and triggering a model parameter correction mechanism if the deviation exceeds a tolerance threshold value; Adjusting the material aging coefficient of the multi-physical field coupling simulation model through a gradient back propagation algorithm until the tolerance deviation converges to a preset tolerance interval; The specific flow of generating the self-adaptive environment test scene sequence by the test scheme optimization module comprises the following steps: taking the environment stress tolerance residual error, the test resource constraint and the failure risk level as optimization targets to construct a test scene decision space; Initializing a root node of a test scene decision tree, traversing an environmental factor combined action space to generate a multi-layer child node branch; selecting a branch node with the highest confidence upper limit from a root node through a Monte Carlo tree search algorithm, and expanding unexplored environment factor combination actions; simulating a device performance degradation path after the selected environmental factor combination action is executed, calculating a test cost benefit function and backtracking and updating decision tree node weights; When the change rate of the test cost benefit function is lower than the convergence threshold, outputting an optimal environment test scene sequence and a corresponding load distribution strategy; the Monte Carlo tree search algorithm performs selective expansion on decision trees, the search process starts from a root node, and a confidence upper limit value of a child node is calculated at each decision point: ; Wherein, the The upper limit value of the confidence is indicated, Representing the child node currently being evaluated, Representing the parent node of the node in question, It is the node that accumulates the test profit value, The parent node accesses the count and, The number of node accesses is recorded and, In order to explore the weight of the coefficient for balancing development and exploration, the algorithm preferably selects the branch with the highest confidence upper limit value to extend downwards, when encountering the insufficiently explored node, a new child node is created, the initial access frequency of the new node is set as one, and the test profit value carries out interpolation estimation based on the father node historical data.
2. 2. The system for identifying key performance synthetic environmental suitability testing of claim 1, wherein the specific flow for performing normalization preprocessing comprises: receiving the original environmental data collected by the temperature sensor, the vibration sensor, the electromagnetic sensor and the humidity sensor, grouping the data according to the sensor types, and aligning sampling time stamps of the data of each group through a time sliding window; Calculating covariance matrixes of the environmental data of each group, decomposing and determining the principal component directions of the data of each group based on covariance eigenvalues, and executing standardized scaling processing along the principal component directions so as to eliminate dimension differences; traversing each group of standardized data, calculating the mahalanobis distance distribution of the standardized data, and marking data points exceeding the range of three times of standard deviation as environment abnormal values; Filling up missing data points through local weighted regression, adopting a median substitution method to process environment abnormal values, and finally mapping each group of data to a zero-to-one interval through range scaling.
3. 3. The system for testing the adaptability of the comprehensive environment for identifying key performances according to claim 2, wherein the specific flow of constructing the multidimensional environment association graph by the dynamic modeling module of the environment scene comprises the following steps: extracting the time sequence characteristics of the preprocessed temperature data, vibration data, electromagnetic interference data and humidity data, and constructing an environmental factor relation matrix; taking historical environmental stress data, equipment performance degradation records and environmental factor coupling relations as map initial nodes, and calculating environmental association strength among the nodes through a map attention mechanism; And updating the weight coefficient of the environmental factor relation matrix based on the environmental data flow, and reconstructing the adjacent matrix of the environmental association strength through the graph rolling network to generate a dynamic environmental stress propagation network.
4. 4. The system for testing the comprehensive environmental adaptability for identifying key performances according to claim 3, wherein the specific flow of the environmental scene simulation control module for dynamically regulating and controlling the parameter combination of the testing environmental cabin comprises the following steps: Analyzing temperature gradient, vibration spectrum, electromagnetic strength and humidity change curves in the optimal environment test scene sequence; Adjusting the heater power, the vibration table frequency, the electromagnetic field generator strength and the humidifier output of the test environmental chamber according to the environmental factor coupling relation matrix; collecting performance response data of the tested equipment under an environmental stress sequence in real time, and feeding back the performance response data to the environmental scene dynamic modeling module to update a multidimensional environmental correlation map; And if the deviation degree of the performance response data and the expected degradation curve exceeds the regulation threshold value, the test scheme optimization module is triggered again to generate a substitute environment test scene sequence.
5. 5. The system for identifying critical performance synthetic environmental suitability test of claim 4, wherein the particular flow of the failure mode prediction module to generate the set of potential failure modes comprises: based on the node connection strength of the environmental stress propagation network, identifying an environmental factor combination chain with high association degree; predicting stress accumulation tracks of each environmental factor combination chain through a time convolution network, and calculating failure probability density at track intersection points; and aggregating the environmental factor combination chains with the failure probability density exceeding a critical threshold value to generate a potential failure mode coding table ordered according to the failure risk.
6. 6. The system for identifying key performance integrated environmental suitability testing of claim 1, further comprising an environmental scenario classification module for managing input data of the environmental scenario simulation control module: receiving real-time environment test scene request data, and extracting environment factor combination feature vectors of the real-time environment test scene request data; Calculating cosine similarity between the feature vector and a historical environment test scene feature library, and classifying the feature vector as a compliance environment test scene if the similarity is higher than a scene matching threshold; And if the similarity is lower than a scene matching threshold, calling the environment scene dynamic modeling module to construct a temporary environment association map, and outputting the newly-added environment test scene characteristics to a history characteristic library.
7. 7. The system for testing the comprehensive environmental suitability for identifying key performance according to claim 1, wherein the specific process of outputting the tolerance residual by the environmental tolerance analysis module comprises: carrying out point-by-point differential operation on the tolerance reference value predicted by the environment stress tolerance virtual model and the actually measured performance degradation index; Carrying out Gaussian filtering smoothing treatment on the differential result to generate a tolerance residual curve and transmitting the tolerance residual curve to a test scheme optimization module; When the slope change of the tolerance residual curve exceeds the early warning threshold, triggering the environmental scene simulation control module to pause the current test sequence and start the safety protocol.

Description

Comprehensive environment adaptability test system for key performance of identification Technical Field The invention relates to the technical field of environment adaptability test, in particular to a comprehensive environment adaptability test system for key performance identification. Background In the modern industrial field, the operation environment of various devices is increasingly complex, from a tropical region with high temperature and high humidity to a severe cold and dry polar region, from a dusty desert region to a marine scene with high salt mist, the diversification and the coupling of environmental factors provide a severe challenge for the performance of the devices. The traditional environment adaptability testing means are often limited to the simulation of a single environmental factor, such as testing only on one item of temperature, humidity or vibration, so that it is difficult to truly reflect the multi-factor compound action scene faced by the device in practical application. In the prior art, part of test systems attempt to introduce a combination of multiple environmental factors, but have obvious drawbacks in the acquisition and processing of environmental data. Most systems can simply record limited environmental parameters, and lack the ability to normalize and preprocess data, so that environmental data of different types and different magnitudes are difficult to effectively correlate and analyze. Meanwhile, in the modeling process of an environment scene, a static model takes the dominant position, and the topological structure of the model cannot be dynamically adjusted according to environment data acquired in real time, so that the response of the model to complex environment changes is lagged, and the dynamic association relation among environment factors is difficult to accurately capture. In the links of equipment performance evaluation and failure prediction, the traditional method mostly depends on empirical judgment or simple statistical analysis, and lacks deep mining on space-time correlation characteristics between environmental stress and equipment performance degradation. This results in insufficient accuracy of the evaluation of performance degradation indicators, and prediction of potential failure modes is difficult to cover the complications under multiple environmental couplings. In addition, the existing test system is weak in environmental tolerance analysis, and cannot construct an accurate virtual model to quantify the tolerance of equipment under a specific environmental profile, so that the test scheme is difficult to optimize accordingly. The immobilization of the test scheme is another prominent problem, most systems cannot dynamically generate a self-adaptive test scene sequence according to the real-time performance feedback and environmental tolerance analysis result of the equipment, so that redundancy or deficiency exists in the test process, test resources are wasted, and key environmental stress combinations can be omitted, so that the reliability and the comprehensiveness of the test result are affected. Disclosure of Invention The invention aims to provide a comprehensive environment adaptability test system for identifying key performances, so as to solve the problems in the background technology. To achieve the above object, the present invention provides an integrated environment adaptability test system for identifying key performance, the system comprising: the multi-environmental factor acquisition processing module is used for acquiring environmental data monitored by various sensors deployed in the test environmental cabin and executing normalization preprocessing; The environment scene dynamic modeling module is used for constructing a multidimensional environment association graph based on the preprocessed environment data and dynamically updating the topological structure of the multidimensional environment association graph according to the environment data flow; the performance index evaluation module is used for receiving the multidimensional environment association map and the preprocessed environment data and evaluating performance degradation indexes of the tested equipment under the composite environment stress through a space-time feature fusion algorithm; the failure mode prediction module generates a potential failure mode set of the tested equipment in the multi-environment coupling scene according to the evolution trend of the performance degradation index and the environmental stress accumulation effect; The environment tolerance analysis module integrates the performance degradation indexes and the potential failure mode set, builds an environment stress tolerance virtual model, and outputs tolerance residual errors of the tested equipment under a specific environment profile; the test scheme optimizing module generates a self-adaptive environment test scene sequence according to the tolerance residual error and