Meteorological equipment health degree assessment method based on regression and fuzzy comprehensive analysis

Abstract

The invention particularly relates to a meteorological equipment health degree assessment method based on regression and fuzzy comprehensive analysis. The method comprises the following steps: constructing a meteorological equipment health degree assessment index system; the method comprises the steps of collecting multi-source data of meteorological equipment, and completing data preprocessing through missing value filling and abnormal value correction; performing feature extraction on the multi-level indexes by adopting regression analysis, screening out core parameters strongly related to the health degree, and quantifying priorities; combining an analytic hierarchy process, integrating expert experience and feature importance results, constructing a hierarchical judgment matrix, and calculating the dynamic weight of each hierarchy index; and based on a fuzzy comprehensive evaluation method, establishing a multi-dimensional comment set and a membership function for the quantitative index and the qualitative index, fusing the weight matrix and the membership data, and outputting a quantitative score of the health state of the meteorological equipment. According to the method, the accuracy of health assessment of the meteorological equipment in severe environments such as extremely cold environments can be realized, and reliable technical support is provided for intelligent operation and maintenance of the meteorological equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of meteorological equipment health assessment, and in particular to a meteorological equipment health assessment method based on regression and fuzzy comprehensive analysis. Background Art

[0002] Current meteorological equipment health assessments face challenges such as complex evaluation indicators, weak inter-component correlations, long data collection cycles, and difficulty measuring qualitative and quantitative indicators. Therefore, specificity should be considered when developing research plans. Common evaluation methods used in engineering practice include qualitative evaluation, quantitative evaluation, multi-metric evaluation, and data-driven evaluation. Given that meteorological equipment health data contains both qualitative and quantitative indicators, multi-metric evaluation methods can be used as the core approach for real-time health assessment of meteorological equipment.

[0003] One study proposed a "1-4-25" green evaluation system for evaluating the effectiveness of flotation collectors in the coal washing process, in which secondary indicators were divided into four categories and tertiary indicators into 25 categories. This study used the Analytic Hierarchy Process (AHP) to determine the weights of the secondary and tertiary indicators in the evaluation system, but the weight determination was too subjective. Another study divided unit equipment into hierarchical systems, constructed a three-layer structure, and used the AHP to determine the weights of each component in the system. Based on the normal distribution characteristics of the unit's historical health data and the Gaussian threshold principle, the upper and lower limits of the indicators were determined. The degradation degree of the real-time monitoring indicators was calculated and applied to a fuzzy evaluation model to obtain a fuzzy judgment matrix to complete the status assessment of the hydropower unit. This study used the same membership function for all indicators, which did not conform to the mechanistic characteristics of each indicator. In addition, the use of the AHP alone to determine the weights was too subjective. One study used a multi-input weighted priority diagram and the analytic hierarchy process to calculate subjective weights, and entropy weighting and the excess multiple method to calculate objective weights. It then introduced moment estimation theory to optimally combine these weights, and finally combined them with a fuzzy comprehensive evaluation method to assess the indoor air quality of subway stations. This study employed four methods to determine the optimal weights, including the entropy weighting method, which calculates the information entropy of each indicator to determine the objective weight. The greater the fluctuation in the indicator value, the more likely it is to be identified as an important indicator. However, this method is inconsistent with the characteristics of meteorological equipment in extremely cold climates. For example, environmental noise is random and volatile, but random instantaneous noise has little impact on the health of the equipment.

[0004] At present, there are still problems in the health assessment of meteorological equipment, such as the lack of an indicator system, weak data correlation, strong subjectivity in weight determination, and difficulty in integrating qualitative and quantitative indicators.

[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present invention, and therefore may include information that does not constitute prior art known to ordinary technicians in this field. Summary of the Invention

[0006] The present invention provides a meteorological equipment health assessment method based on regression and fuzzy comprehensive analysis, which can achieve the accuracy of meteorological equipment health assessment in harsh environments such as extreme cold, and thus can overcome the defects existing in the existing technology to a certain extent.

[0007] Other features and advantages of the present invention will become apparent from the following detailed description, or may be learned in part by practice of the present invention.

[0008] According to a first aspect of the present invention, a method for evaluating the health of meteorological equipment based on regression and fuzzy comprehensive analysis is provided, the method comprising:

[0009] Step 1: Construct a meteorological equipment health assessment index system. The index system includes parameter layer, component layer, subsystem layer and system layer, covering multi-level indicators of hardware performance, environmental adaptability and working stability;

[0010] Step 2: Collect multi-source data from meteorological equipment and complete data preprocessing by filling missing values and correcting outliers;

[0011] Step 3: Based on multi-source data, regression analysis is used to extract features from multi-level indicators, screen out core parameters that are strongly correlated with health, and quantify their priorities;

[0012] Step 4: Combine the analytic hierarchy process (AHP) to integrate expert experience and feature importance results, build a hierarchical judgment matrix, and calculate the dynamic weights of indicators at each level;

[0013] Step 5: Based on the fuzzy comprehensive evaluation method, a multi-dimensional comment set and membership function are established for quantitative indicators and qualitative indicators respectively, and the weight matrix and membership data are integrated to output a quantitative score of the health status of meteorological equipment.

[0014] In some exemplary embodiments, in step 1, the meteorological device is a gravity wave meter, and its health status evaluation set is divided into four levels: healthy, sub-healthy, faulty, and ineffective, and each level is further divided into three levels of status.

[0015] In some exemplary embodiments, in step 2, a hot card filling method is used to address the data missing problem, specifically including:

[0016] Find the object most similar to the object containing the null value;

[0017] Fill missing values with the corresponding values of the most similar objects;

[0018] The outliers are corrected by judging the LOF outlier factor points.

[0019] In some exemplary embodiments, in step three, Lasso regression is used for feature extraction, the value of the regularization parameter λ is determined by cross-validation, and the coordinate descent method is used to solve the optimal parameter β.

[0020] In some exemplary embodiments, the specific step of determining the value of the regularization parameter λ by cross-validation includes:

[0021] Step 1: Prepare data: Divide multi-source data into training set, validation set and test set; cross-validation is only performed on the training set;

[0022] Step 2: Define parameter range: Determine a reasonable range of alpha values;

[0023] Step 3: Split the training data: Use k-fold cross validation to divide the training data into k subsets of equal size;

[0024] Step 4: Perform k rounds of training and validation: For each candidate value of λ, perform the following steps:

[0025] 1) For each of the k subsets, use the subset as the validation set, and combine the remaining k-1 subsets as the current training set;

[0026] 2) Use the current training set to train the Lasso parameter model and use the selected λ value;

[0027] 3) Evaluate the performance of the Lasso parameter model on the validation set;

[0028] 4) Record the verification performance of each round;

[0029] Step 5: Calculate the average performance: For each value of λ, calculate its average performance on the k validation sets;

[0030] Step 6: Select the best λ: Select the λ value that makes the Lasso parameter model have the best average performance in cross validation;

[0031] Step 7: Use the best λ to train the Lasso parameter model: Use all the training data and the best λ value to retrain the Lasso parameter model;

[0032] Step 8 Final evaluation: Use the test set to perform a final evaluation of the Lasso parameter model to estimate the performance of the Lasso parameter model on unseen data.

[0033] In some exemplary embodiments, the specific steps of the coordinate descent method include:

[0034] Initialize the parameter vector β;

[0035] Loop through each parameter β j , fix other parameters and solve the univariate optimization problem;

[0036] Update parameter β j , until the parameter vector β converges.

[0037] In some exemplary embodiments, in step 4, the specific steps of the analytic hierarchy process (AHP) include:

[0038] Construct a weighted judgment matrix based on experts’ ratings of the relative importance of different indicators;

[0039] The weight coefficient is calculated using the sum-product method;

[0040] Conduct consistency checks to ensure the rationality of weight distribution;

[0041] The evaluation results of multiple experts are integrated to generate a multi-expert weight vector.

[0042] In some exemplary embodiments, in step 5, the specific steps of the fuzzy comprehensive evaluation method include:

[0043] Establishing the index system factor set and expert comment set;

[0044] Construct a fuzzy relationship matrix to determine the membership of each indicator to the comment set;

[0045] Select fuzzy operators to calculate the result set of fuzzy comprehensive evaluation;

[0046] Membership functions are designed for quantitative indicators and qualitative indicators respectively to quantify their health status.

[0047] In some exemplary embodiments, the membership function includes at least one of a triangular membership function, a trapezoidal membership function, and a Gaussian membership function, and different function forms are designed for different indicators.

[0048] According to a second aspect of the present invention, a storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method for evaluating the health of meteorological equipment based on regression and fuzzy comprehensive analysis as described in the first aspect is implemented.

[0049] The embodiment of the present invention provides a method for evaluating the health of meteorological equipment based on regression and fuzzy comprehensive analysis. First, a set of multi-level meteorological equipment health evaluation indicators with engineering practice value is systematically proposed. The evaluation indicators start from the composition details and construction mechanism of each component of the meteorological equipment, and accurately and in detail capture the working characteristics of meteorological equipment operating in real extreme cold climates. Secondly, regression analysis is used to extract features from multi-level indicators to obtain the priority of each indicator in the health evaluation of meteorological equipment, and assist in the subsequent calculation of indicator weights. The hierarchical analysis method is used to combine the results of expert scoring and feature extraction to construct an importance matrix and calculate the weights of hierarchical indicators. Finally, the fuzzy comprehensive evaluation method is used to construct a comment set and its corresponding membership function from the operating mechanism level of the meteorological equipment for qualitative and quantitative indicators, and the weights and membership matrix are combined to calculate the health status of the meteorological equipment corresponding to each data.

[0050] Compared with existing meteorological equipment health assessment technologies, this invention significantly improves the objectivity, accuracy, and engineering applicability of the assessment by building a multi-level evaluation system and integrating innovative methods. This is specifically demonstrated in the following aspects:

[0051] 1. Traditional methods mostly rely on equipment mechanism models or single indicator threshold judgments, which are difficult to reflect the overall health status of equipment in complex environments. To this end, the present invention innovatively constructs a multi-level indicator framework covering hardware performance, environmental adaptability and operational stability, comprehensively capturing equipment health characteristics from component mechanisms to system behaviors. Combined with the fuzzy comprehensive evaluation method, the fuzzy relationship between qualitative indicators (such as mechanical fatigue) and quantitative indicators (such as data acquisition error rate) is quantified through the membership function, solving the problem of insufficient processing of uncertain information by traditional methods. Compared with existing technologies, this method can more systematically evaluate the comprehensive performance degradation trend of equipment in harsh environments such as extreme cold.

[0052] 2. The analytic hierarchy process (AHP) widely used in the prior art is highly dependent on expert experience scoring, and is prone to bias in weight allocation due to subjective cognitive differences. To this end, the present invention proposes a hierarchical feature extraction strategy of "regression analysis + supervised learning", which automatically screens key indicators and quantifies priorities based on historical equipment operation data, and combines AHP to build a dynamic weight calculation model. Through the collaborative optimization of data-driven and expert knowledge, the subjectivity of weight allocation is effectively reduced. Existing studies mostly use a single evaluation model, which makes it difficult to take into account the sparsity of data and the correlation of indicators in complex scenarios. The present invention organically combines fuzzy comprehensive evaluation, feature extraction and hierarchical analysis for the first time: using regression analysis to solve the redundancy problem of high-dimensional indicators of meteorological equipment, strengthening domain knowledge embedding through AHP, and finally using fuzzy evaluation to achieve unified quantification of qualitative and quantitative indicators.

[0053] In summary, the present invention solves the core problems of the existing technology, such as strong subjectivity, single dimension, and insufficient generalization ability, through methodological innovation and technological integration. It has significant advancement and practical value in the field of meteorological equipment health management, and provides reliable technical support for the intelligent operation and maintenance of meteorological equipment.

[0054] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The accompanying drawings are incorporated into and constitute a part of this specification, illustrate embodiments consistent with the present invention, and together with the description, serve to explain the principles of the present invention. Obviously, the drawings described below are only some embodiments of the present invention, and it is clear that those skilled in the art can derive other drawings based on these drawings without inventive effort.

[0056] Figure 1 This is a framework diagram of the meteorological equipment health assessment method based on regression and fuzzy comprehensive analysis;

[0057] Figure 2 This is a flow chart of the meteorological equipment health assessment method based on regression and fuzzy comprehensive analysis;

[0058] Figure 3 This is a schematic diagram of the structure of the gravity wave instrument;

[0059] Figure 4 Schematic diagram of the data completion preprocessing process for hot card filling;

[0060] Figure 5 Schematic diagram of the feature extraction process based on Lasso regression. DETAILED DESCRIPTION

[0061] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be embodied in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete and will fully convey the concepts of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0062] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the blocks shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0063] In view of the shortcomings and deficiencies of the existing technology, the present invention proposes a multi-level intelligent assessment method for the health of meteorological equipment, and constructs a systematic evaluation system for the operating characteristics of equipment in extremely cold climates. Figure 1 As shown in the figure, first, based on the composition and operating mechanism of equipment components, a multi-level health evaluation indicator covering hardware performance, environmental adaptability, and operational stability is designed. Second, multi-source data is collected according to the equipment operation cycle, and data preprocessing is completed through missing value filling, outlier correction, and standardization. Subsequently, regression analysis is used to extract features from the indicators, screen out core parameters strongly correlated with health, quantify their priorities, and assist in weight allocation. Furthermore, combined with the analytic hierarchy process (AHP), expert experience and feature importance results are integrated to construct a hierarchical judgment matrix, and the dynamic weights of indicators at each level are calculated. Finally, based on the fuzzy comprehensive evaluation method, multi-dimensional comment sets and membership functions are established for quantitative indicators (such as sensor accuracy offset rate) and qualitative indicators (such as mechanical structure fatigue). The weight matrix and membership data are integrated to output a quantitative score of the equipment health status. This method innovatively integrates mechanism analysis, data-driven development, and expert knowledge, significantly improving the accuracy and engineering applicability of meteorological equipment health assessment in harsh environments such as extreme cold.

[0064] refer to Figure 2 As shown, the following steps may be specifically included:

[0065] Step 1: Construct a meteorological equipment health assessment index system. The index system includes parameter layer, component layer, subsystem layer and system layer, covering multi-level indicators of hardware performance, environmental adaptability and working stability;

[0066] Step 2: Collect multi-source data from meteorological equipment and complete data preprocessing by filling missing values and correcting outliers;

[0067] Step 3: Use regression analysis to extract features from multi-level indicators, screen out core parameters that are strongly correlated with health, and quantify their priorities;

[0068] Step 4: Combine the analytic hierarchy process (AHP) to integrate expert experience and feature importance results, build a hierarchical judgment matrix, and calculate the dynamic weights of indicators at each level;

[0069] Step 5: Based on the fuzzy comprehensive evaluation method, a multi-dimensional comment set and membership function are established for quantitative indicators and qualitative indicators respectively, and the weight matrix and membership data are integrated to output a quantitative score of the health status of meteorological equipment.

[0070] Below, each step in this exemplary implementation will be described in more detail with reference to the accompanying drawings and embodiments.

[0071] Step 1: Construction of meteorological equipment health assessment index system

[0072] The meteorological equipment studied in this paper is a gravity wave instrument. The equipment consists of three independent subsystems. Each subsystem has its corresponding working components. Therefore, when evaluating the health of the equipment, the indicators should be divided according to different levels. The levels can be divided into parameter layer, component layer, subsystem layer and system layer from low to high. The architecture is as follows: Figure 3 shown.

[0073] Based on the operating environment data and performance of gravity wave instruments over a period of time, combined with the experience and recommendations of meteorological equipment component manufacturers and meteorological researchers, four health statuses can be categorized: healthy, subhealthy, faulty, and ineffective. Due to the relatively complex components and hierarchical structure of gravity wave instruments, a single evaluation set cannot accurately summarize their health status. Therefore, each health level is further categorized into three levels based on the detailed operating conditions and indicator deviations of different components. The detailed evaluation set divisions and the corresponding status descriptions for each level are shown in Table 1.

[0074] Table 1 Classification of meteorological equipment health status evaluation sets

[0075]

[0076] Evaluation indicators of different dimensions and levels can be set according to the software, hardware, data, and criteria used during the operation of meteorological equipment. The hardware devices contained in meteorological equipment include CCD cameras, TEC air conditioners, filter wheels, temperature sensors, humidity sensors, air pressure sensors, noise sensors, voltage sensors, and servers. The entire meteorological equipment system also covers software of different modules, including server programs that provide data persistence and real-time analysis, client programs that temporarily store and send data, data cleaning programs, image verification programs, meteorological equipment health assessment programs, meteorological equipment fault detection programs, and meteorological equipment life degradation analysis programs. At the same time, a large amount of status data will be generated during the operation of meteorological equipment. These data include temperature, humidity, air pressure, noise, images, and other data. The accurate and efficient reception of these data is also the key to the normal operation of the meteorological equipment system. In addition, the quality management of meteorological equipment needs to comply with the quality assurance standards in extremely cold climates. Combined with the technical status and development characteristics of military equipment, according to the GJB 9001C quality management system, when formulating the health evaluation indicators of meteorological equipment, consideration should be given to introducing qualitative indicators such as safety, environmental adaptability, reliability, maintainability, testability, and supportability. These indicators ensure the comprehensive quality of the equipment. The specific evaluation indicators are shown in Table 2.

[0077] Table 2 Meteorological equipment health evaluation index system

[0078]

[0079]

[0080]

[0081]

[0082] Step 2: Construction of meteorological equipment dataset

[0083] The data processed in this paper comes from the Tuli River National Climate Reference Station. During the actual data collection process, due to the inconsistent duty cycles of various sensors, data gaps or outlier interference may occur. Therefore, before building the model, the data should be balanced to ensure that the data used for health assessment at a given moment contains the results of all indicators. This prevents data loss caused by differences in sensor transmission cycles from affecting the health assessment results of meteorological equipment. Therefore, this paper uses a hot card filling method to fill in the missing data.

[0084] The preprocessing steps for completing meteorological equipment data for thermal card filling are as follows:

[0085] Step 1: Find the most similar objects: Find the objects in the dataset that are most similar to the object containing the null value. The determination of similarity varies depending on the needs of the specific problem and can be based on the characteristics and attributes of the object.

[0086] Step 2: Fill missing values: Use the corresponding value of the most similar object found to fill the missing values. For example, if a feature value of object A is missing, and object B is the most similar object to A, then the value of the feature of object B is used to fill the missing value of A.

[0087] Step 3: LOF outlier factor point judgment.

[0088] Step 3: Extracting health features of meteorological equipment based on regression analysis

[0089] When evaluating the health of meteorological equipment, a total of 54 indicators are considered across four different dimensions: hardware, software, data, and quality criteria. Scoring each indicator individually without considering its importance to the overall meteorological equipment can result in significant deviations from actual engineering applications. To address this issue, a feature extraction method can be used to calculate the weights for each indicator. This weighted assessment yields an assessment that better reflects the health trends of meteorological equipment in real-world environments.

[0090] When using regression analysis for feature extraction, the cross-validation method is used to obtain the value of the regularization parameter λ. The steps are as follows:

[0091] Step 1: Prepare the data: Split the dataset into a training set and a test set. Cross-validation is usually performed only on the training set.

[0092] Step 2: Define parameter range: Determine a reasonable range of alpha values. This range can be linear, logarithmic, or any other distribution based on previous experience.

[0093] Step 3: Split the training data: Use k-fold cross validation to split the training data into k subsets of equal size. Common choices are k = 5 or k = 10.

[0094] Step 4: Perform k rounds of training and validation: For each candidate value of λ, perform the following steps:

[0095] 1) For each of the k subsets, use the subset as the validation set, and combine the remaining k-1 subsets as the current training set.

[0096] 2) Train the model using the current training set and the selected lambda value.

[0097] 3) Evaluate the performance of the model on the validation set, usually using mean squared error (MSE) or other relevant performance metrics.

[0098] 4) Record the verification performance of each round.

[0099] Step 5: Calculate the average performance: For each value of λ, calculate its average performance on the k validation sets.

[0100] Step 6: Select the optimal λ: Select the λ value that gives the best average performance of the model in cross-validation. The average performance can be the minimum average error, or the best performance while taking into account the complexity of the model.

[0101] Step 7: Use the best λ to train the model: Retrain the model using all the training data and the best λ value.

[0102] Step 8 Final Evaluation: Optionally, perform a final evaluation of the model using the held-out test set to estimate the model's performance on unseen data.

[0103] Given the large number of samples, distinct features, and weak correlation in the meteorological equipment dataset, the coordinate descent method can be used to find the optimal parameter B. Coordinate descent is an algorithm for solving optimization problems that iteratively optimizes the parameters of the objective function to gradually approach the optimal solution. In the context of regression, coordinate descent is particularly effective because it can efficiently handle L1 regularization terms. The basic steps of coordinate descent are as follows:

[0104] Step 1 Initialization: Select an initial parameter vector β. This vector can be a zero vector, a random vector, or a heuristic choice.

[0105] Step 2 loops through the coordinates: each parameter β j (where j = 1, 2, 3, ..., p), perform the following steps:

[0106] 1) Fix other parameters: except β j All parameters except are considered constants.

[0107] 2) Solve the univariate optimization problem: find β that minimizes the objective function j In Lasso regression, this usually means finding the β that minimizes the following function for a given λ and other fixed parameters: j :

[0108]

[0109] 3) Update parameters: Update β with the value obtained in the previous step j

[0110] Step 3 Check convergence: If the parameter vector β does not change significantly, or the reduction of the objective function is lower than a certain threshold, stop the iteration; otherwise, return to step 2 to continue the iteration.

[0111] Step 4 outputs the result: When the algorithm converges, it outputs the final parameter vector β, which contains the coefficients of the model.

[0112] Step 4: Determination of hierarchical weights based on analytic hierarchy process

[0113] The determination of hierarchical weights based on the hierarchical analysis method mainly includes the following four steps.

[0114] Step 1: Construction of weight judgment matrix

[0115] The weighted judgment matrix is composed of the experts' evaluation of the relative importance of different evaluation indicators. The scoring criteria are shown in Table 3.

[0116] Table 3 Description of AHP scoring matrix

[0117]

[0118] In the evaluation analysis, m experts can be used to score n indicators, and m weight judgment matrices can be obtained.

[0119] Step 2 Calculation of weight coefficient

[0120] In specific evaluations, the sum-product method can be used to calculate the weight coefficients and obtain the weight judgments of m experts. The specific process is as follows:

[0121] Weight judgment matrix A m Normalize by columns, the elements in the matrix satisfy:

[0122]

[0123] Where, Finally, the normalized matrix can be obtained By taking the average value of each row, we can get the weight coefficient:

[0124]

[0125] The weight coefficient vector W of the score of the mth method is calculated from this m , and so on, the weight coefficient vectors of other methods can be obtained.

[0126] Step 3: Consistency check

[0127] For the weight coefficient vector determined by AHP, a consistency check is required. The specific analysis steps are as follows:

[0128] 1) Calculate the normalized matrix The characteristic root λ mmax :

[0129]

[0130] Where n is the order of the judgment matrix.

[0131] 2) Calculate the consistency index CI:

[0132]

[0133] 3) Calculate the consistency ratio CR

[0134]

[0135] According to the consistency index table, the smaller the CI, the better the consistency of the evaluation results. When CR < 0.1, it has a relatively satisfactory consistency. If CR ≥ 0.1, the judgment matrix needs to be modified. The average random consistency index value is shown in Table 4.

[0136] Table 4 Average random consistency index values

[0137]

[0138] Step 4 Weight coefficient weighting

[0139] W m The evaluation results of a single expert for the evaluation index are subjective and will have a certain impact on the final result. The evaluation results of multiple experts are integrated to obtain the multi-expert weight vector W, which is calculated as

[0140]

[0141] Where B m is the evaluation weight of different experts in the expert group, B m The specific selection can be made according to the actual situation. In the present invention, it is assumed that the weights of the experts are consistent.

[0142] Step 5: Calculation of meteorological equipment health based on fuzzy comprehensive analysis

[0143] The calculation of meteorological equipment health based on fuzzy comprehensive analysis mainly includes the following six steps.

[0144] Step 1: Establish an indicator system factor set

[0145] Z=[z1,z2,...,z n ]

[0146] Where n is the number of comments in the comment set

[0147] Step 2: Establish an expert comment set. The comment set is the conclusion of the quantitative evaluation of the indicator factors. The comment set is as follows:

[0148] V=[v1,v2,...,v n ]

[0149] Where p is the number of evaluation levels set.

[0150] Step 3: Establish the fuzzy relationship matrix R

[0151] First, we need to construct a single indicator membership subset R i ,Right now

[0152] R i =[r i1 ,r i2 ,r i3 ,...,r ip ]

[0153]

[0154] Where R i is the membership degree of a certain indicator to the fuzzy subset of the comment set V in the evaluation process, where i = (1, 2, ..., n), k = (1, 2, ..., p), and Z ik Represents the score (degree of membership) of the kth evaluation of the i-th indicator. Gradually construct the fuzzy relationship matrix R of all indicators:

[0155]

[0156] Step 4: Create a result set of fuzzy comprehensive evaluation

[0157] The evaluation result set can be expressed as Where: B is the membership of the evaluation target to V, W is the weight vector, is a fuzzy operator.

[0158] Table 5 Calculation methods and characteristics of four fuzzy operators

[0159]

[0160]

[0161] Step 5: Construction of membership function

[0162] According to the fuzzy comprehensive evaluation method, the membership function converts qualitative indicators into the membership of each comment, so the selection of the membership function is particularly important. Common membership functions include triangular membership function, trapezoidal membership function, Gaussian membership function, S-shaped membership function, bell-shaped membership function, etc.

[0163] During use, the operating status of meteorological equipment is often correlated with multiple components. Therefore, when determining the membership function, it is important to consider the possibility that individual indicators may deviate from their standard values while the equipment continues to function normally. Taking the membership function for temperature as an example, since the equipment operates in extreme climates for a long time, the instrument is relatively cold-resistant. Therefore, the downward trend of the membership function for the evaluation "healthy" should be lower in low-temperature environments than in high-temperature conditions. For qualitative indicators, the determination of the membership function requires the experience of experts or researchers with extensive engineering experience. The membership functions for quantitative indicators in meteorological equipment health assessment are shown in Table 6.

[0164] Table 6 Membership function of meteorological equipment indicators

[0165]

[0166]

[0167]

[0168]

[0169]

[0170] Furthermore, the above-described figures are merely illustrative of the processes included in the method according to exemplary embodiments of the present invention and are not intended to be limiting. It is readily understood that the processes illustrated in the above-described figures do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0171] Other embodiments of the present invention will readily occur to those skilled in the art after considering the specification and practicing the invention herein. This application is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, with the true scope and spirit of the invention being indicated by the claims.

[0172] It should be understood that the present invention is not limited to the exact construction described above and shown in the drawings and that various modifications and variations can be made without departing from the scope thereof, which is limited only by the appended claims.

Claims

1. A meteorological equipment health assessment method based on regression and fuzzy comprehensive analysis, characterized in that: The method comprises the following steps: Step 1: Construct a meteorological equipment health assessment index system. The index system includes parameter layer, component layer, subsystem layer and system layer, covering multi-level indicators of hardware performance, environmental adaptability and working stability; Step 2: Collect multi-source data from meteorological equipment and complete data preprocessing by filling missing values and correcting outliers; Step 3: Based on multi-source data, regression analysis is used to extract features from multi-level indicators, screen out core parameters that are strongly correlated with health, and quantify their priorities; Step 4: Combine the analytic hierarchy process (AHP) to integrate expert experience and feature priority results, build a hierarchical judgment matrix, and calculate the dynamic weights of indicators at each level; Step 5: Based on the fuzzy comprehensive evaluation method, a multi-dimensional comment set and membership function are established for quantitative indicators and qualitative indicators respectively, and the weight matrix and membership data are integrated to output a quantitative score of the health status of meteorological equipment.

2. The method according to claim 1, characterized in that In the step 1, the meteorological equipment is a gravity wave instrument, and its health status evaluation set is divided into four levels: healthy, sub-healthy, faulty, and ineffective, and each level is further divided into three levels of status.

3. The method according to claim 1, characterized in that In step 2, a hot card filling method is used to handle the data missing problem, specifically including: Find the object most similar to the object containing the null value; Fill missing values with the corresponding values of the most similar objects; The outliers are corrected by judging the LOF outlier factor points.

4. The method according to claim 1, wherein In the step 3, Lasso regression is used for feature extraction, the value of the regularization parameter λ is determined by cross-validation, and the coordinate descent method is used to solve the optimal parameter β.

5. The method according to claim 4, characterized in that The specific steps of determining the value of the regularization parameter λ by cross-validation include: Step 1: Prepare data: Divide multi-source data into training set, validation set and test set; cross-validation is only performed on the training set; Step 2: Define parameter range: Determine a reasonable range of alpha values; Step 3: Split the training data: Use k-fold cross validation to divide the training data into k subsets of equal size; Step 4: Perform k rounds of training and validation: For each candidate value of λ, perform the following steps: 1) For each of the k subsets, use the subset as the validation set, and combine the remaining k-1 subsets as the current training set; 2) Use the current training set to train the Lasso parameter model and use the selected λ value; 3) Evaluate the performance of the Lasso parameter model on the validation set; 4) Record the verification performance of each round; Step 5: Calculate the average performance: For each value of λ, calculate its average performance on the k validation sets; Step 6: Select the best λ: Select the λ value that makes the Lasso parameter model have the best average performance in cross validation; Step 7: Use the best λ to train the Lasso parameter model: Use all the training data and the best λ value to retrain the Lasso parameter model; Step 8 Final evaluation: Use the test set to perform a final evaluation of the Lasso parameter model to estimate the performance of the Lasso parameter model on unseen data.

6. The method according to claim 4, characterized in that The specific steps of the coordinate descent method include: Initialize the parameter vector β; Loop through each parameter β j , fix other parameters and solve the univariate optimization problem; Update parameter β j , until the parameter vector β converges.

7. The method according to claim 1, characterized in that In the step 4, the specific steps of the analytic hierarchy process (AHP) include: Construct a weighted judgment matrix based on experts’ ratings of the relative importance of different indicators; The weight coefficient is calculated using the sum-product method; Conduct consistency checks to ensure the rationality of weight distribution; The evaluation results of multiple experts are integrated to generate a multi-expert weight vector.

8. The method according to claim 1, characterized in that In step 5, the specific steps of the fuzzy comprehensive evaluation method include: Establishing the index system factor set and expert comment set; Construct a fuzzy relationship matrix to determine the membership of each indicator to the comment set; Select fuzzy operators to calculate the result set of fuzzy comprehensive evaluation; Membership functions are designed for quantitative indicators and qualitative indicators respectively to quantify their health status.

9. The method according to claim 8, characterized in that The membership function includes at least one of a triangular membership function, a trapezoidal membership function, and a Gaussian membership function, and different function forms are designed for different indicators.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for evaluating the health of meteorological equipment based on regression and fuzzy comprehensive analysis according to any one of claims 1 to 9 is implemented.

Images