Hydraulic oil deterioration detection method and device, storage medium and computer equipment

Abstract

The hydraulic oil degradation detection method, device, storage medium and computer equipment provided by the application, the method comprises the following steps: acquiring detection data and inputting the detection data into a preset target model, the target model comprises a plurality of sub-models, then determining a plurality of evidence bodies corresponding to the detection data, and combining a preset optimization function to determine a target optimization function. Then, based on the target optimization function, the correction weight of each evidence body is determined to correct the corresponding evidence body, and a plurality of target evidence bodies corresponding to the detection data are obtained, then the probability distribution function of each sub-model on the detection data is determined, and a fusion detection result is generated. By correcting a plurality of evidence bodies, the weights occupied by the output results of each sub-model can be considered comprehensively. Moreover, by fusing each probability distribution function to obtain a fusion detection result, the state of the hydraulic oil can be analyzed comprehensively from multiple dimensional levels, thereby improving the accuracy of the hydraulic oil degradation detection.

Description

Technical Field

[0001] This application relates to the field of performance testing technology, and in particular to a method, apparatus, storage medium, and computer equipment for detecting hydraulic oil deterioration. Background Technology

[0002] High-voltage circuit breakers are commonly used in high-voltage circuits to interrupt and close no-load and load currents. Within a circuit breaker, the operating mechanism is a crucial component for its reliable function. For hydraulic operating mechanisms, deterioration and performance degradation of the hydraulic oil can lead to problems such as oil leakage in the high-voltage circuit breaker operating mechanism, frequent pump starts, and prolonged inability of the pump to build pressure, posing a serious threat to the safe and reliable operation of the power grid and equipment. Therefore, hydraulic oil testing is essential for maintaining the normal operation of hydraulic circuit breakers.

[0003] Currently, hydraulic oil is generally tested using models trained on sample data. However, this testing method considers only a single and fixed perspective, making it difficult to comprehensively analyze the state of hydraulic oil from multiple dimensions, which leads to low accuracy in detecting hydraulic oil degradation. Summary of the Invention

[0004] The purpose of this application is to at least solve one of the aforementioned technical defects, particularly the technical defect that the existing technology considers only one and fixed angles for the detection of hydraulic oil, making it difficult to comprehensively analyze the state of hydraulic oil from multiple dimensions, thus resulting in low accuracy in detecting hydraulic oil degradation.

[0005] In a first aspect, this application provides a method for detecting hydraulic oil deterioration, the method comprising:

[0006] The process involves acquiring data to be detected and inputting it into a pre-trained target model to determine multiple evidence bodies corresponding to the data to be detected; wherein the target model includes multiple sub-models.

[0007] Based on the multiple evidence bodies corresponding to the data to be detected and the preset optimization function, the target optimization function is determined;

[0008] Based on the objective optimization function, determine the corrected weights for each piece of evidence corresponding to the data to be detected;

[0009] The corresponding evidence body is corrected by using the correction weight of each evidence body corresponding to the data to be detected, so as to obtain the target evidence body for each evidence body corresponding to the data to be detected.

[0010] Based on the multiple target evidence bodies corresponding to the data to be detected, determine the probability allocation function of each sub-model in the target model on the data to be detected, so as to generate a fusion detection result corresponding to the data to be detected.

[0011] In one embodiment, the training process of the target model uses oil data collected by various oil sensors under various operating conditions as training data.

[0012] In one embodiment, the step of determining the corrected weight of each piece of evidence corresponding to the data to be detected based on the objective optimization function includes:

[0013] The objective optimization function is solved according to a pre-agreed method to obtain the evidence weight of each evidence body corresponding to the data to be detected;

[0014] Determine the target weight; wherein, the target weight is the largest evidence weight among the evidence weights of each evidence body corresponding to the data to be detected;

[0015] The relative weight of each piece of evidence corresponding to the data to be detected is calculated based on the evidence weight of each piece of evidence corresponding to the data to be detected and the target weight, respectively.

[0016] The relative weight of each piece of evidence corresponding to the data to be detected is determined as its corresponding corrected weight.

[0017] In one embodiment, the step of determining the plurality of evidence bodies corresponding to the data to be detected includes:

[0018] Determine the model type of each sub-model in the target model;

[0019] Based on the model type of each sub-model in the target model, determine the preset loss function corresponding to each sub-model;

[0020] Obtain the output results of the data to be detected on each sub-model in the target model;

[0021] Based on the preset loss function corresponding to each sub-model and the output results of the data to be detected on each sub-model of the target model, the target loss function of the data to be detected on each sub-model of the target model is determined.

[0022] The target loss function of the data to be detected on each sub-model of the target model is determined as the evidence body corresponding to the data to be detected.

[0023] In one embodiment, the step of determining the probability assignment function of each sub-model in the target model on the data to be detected includes:

[0024] Determine the preset probability function for each sub-model;

[0025] Obtain the output results of the data to be detected on each sub-model in the target model;

[0026] For each sub-model in the target model, the output result of the data to be detected on the sub-model and the target evidence body corresponding to the sub-model among the multiple target evidence bodies corresponding to the data to be detected are substituted into the preset probability function corresponding to the sub-model to obtain the probability allocation function of the sub-model on the data to be detected.

[0027] In one embodiment, the step of correcting the corresponding evidence body using the correction weights of each evidence body corresponding to the data to be detected includes:

[0028] For each piece of evidence corresponding to the data to be detected, the product of the correction weight corresponding to the evidence and the evidence is determined as the target evidence of the sub-model corresponding to the evidence on the data to be detected, so as to complete the correction of the evidence of the sub-model corresponding to the evidence on the data to be detected.

[0029] In one embodiment, the expression for the optimization function is:

[0030]

[0031] In the formula, The function represents the optimization function, where n is the number of sub-models in the target model, and w is the value of w. i E represents the weight of evidence i. i Let i represent evidence body i, and ME be the centroid vector of the evidence matrix composed of all evidence bodies. <w i E i ME> represents w i E i The dot product of ME, d i Let be the distance between evidence body i and the ME vector.

[0032] Secondly, this application provides a hydraulic oil deterioration detection device, the device comprising:

[0033] A data acquisition module is used to acquire data to be detected and input the data to be detected into a pre-trained target model to determine multiple evidence bodies corresponding to the data to be detected; wherein, the target model includes multiple sub-models;

[0034] The function determination module is used to determine the target optimization function based on multiple evidence bodies corresponding to the data to be detected and a preset optimization function;

[0035] The weight determination module is used to determine the corrected weight of each piece of evidence corresponding to the data to be detected based on the target optimization function.

[0036] The evidence body correction module is used to correct the corresponding evidence body by using the correction weight of each evidence body corresponding to the data to be detected, so as to obtain the target evidence body for each evidence body corresponding to the data to be detected.

[0037] The probability detection module is used to determine the probability allocation function of each sub-model in the target model on the data to be detected based on multiple target evidence bodies corresponding to the data to be detected, so as to generate a fusion detection result corresponding to the data to be detected.

[0038] Thirdly, this application provides a storage medium storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the steps of the hydraulic oil deterioration detection method as described in any of the above embodiments.

[0039] Fourthly, this application provides a computer device, including: one or more processors, and a memory;

[0040] The memory stores computer-readable instructions, and when the one or more processors execute the computer-readable instructions, they perform the steps of the hydraulic oil deterioration detection method as described in any of the above embodiments.

[0041] As can be seen from the above technical solutions, the embodiments of this application have the following advantages:

[0042] This application provides a hydraulic oil degradation detection method, apparatus, storage medium, and computer equipment. The method includes: acquiring data to be tested during hydraulic oil degradation detection and inputting it into a preset target model, which includes multiple sub-models; determining multiple evidence bodies corresponding to the data to be tested; and determining a target optimization function based on a preset optimization function. Then, based on the target optimization function, determining the correction weight of each evidence body to correct the corresponding evidence body, obtaining multiple target evidence bodies corresponding to the data to be tested, and determining the probability allocation function of each sub-model on the data to be tested, ultimately generating a detection result corresponding to the data to be tested. By correcting the multiple evidence bodies determined based on the multiple sub-models in the target model, the weight of the output results of each sub-model can be comprehensively considered, improving the reliability of each evidence body corresponding to the data to be tested. Furthermore, the probability allocation function of each sub-model on the data to be tested is determined to obtain a fused detection result. Thus, by fusing information obtained from multiple sub-models, the state of hydraulic oil can be comprehensively analyzed from multiple dimensions, thereby improving the accuracy of hydraulic oil degradation detection. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 A schematic flowchart of a hydraulic oil deterioration detection method provided in an embodiment of this application;

[0045] Figure 2 A flowchart illustrating the steps of determining the corrected weights of each piece of evidence corresponding to the data to be detected based on the target optimization function, as provided in this application embodiment;

[0046] Figure 3 A flowchart illustrating the steps for determining multiple evidence bodies corresponding to the data to be detected, as provided in an embodiment of this application;

[0047] Figure 4 A flowchart illustrating the steps for determining the probability assignment function of each sub-model in the target model on the data to be detected, as provided in an embodiment of this application;

[0048] Figure 5 This is a schematic diagram of the structure of a hydraulic oil deterioration detection device provided in an embodiment of this application;

[0049] Figure 6 This is an internal structural diagram of a computer device provided in an embodiment of this application. Detailed Implementation

[0050] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0051] In one embodiment, this application provides a method for detecting hydraulic oil deterioration. The following embodiments illustrate the application of this method to a server. It is understood that the hydraulic oil deterioration detection method can be performed on a single server or a server cluster consisting of multiple servers, and this application does not impose any specific limitations on this.

[0052] like Figure 1 As shown, this application provides a method for detecting hydraulic oil deterioration, the method comprising:

[0053] Step S101: Obtain the data to be detected and input it into the pre-trained target model to determine the multiple evidence bodies corresponding to the data to be detected.

[0054] The target model comprises multiple sub-models, each of which is a different neural network model. The target model receives hydraulic parameters and, based on these parameters, outputs the corresponding hydraulic oil degradation detection results.

[0055] It is understood that the data to be tested refers to the hydraulic fluid parameters of the hydraulic oil to be tested for degradation. The data to be tested includes, but is not limited to, temperature, density, kinematic viscosity, dielectric constant, moisture content, pH value, etc. The evidence body describes the output of one of the sub-models in the target model. Specifically, since the target model includes multiple sub-models, when the data to be tested is input into the target model, it is input into each sub-model within the target model. Therefore, the multiple evidence bodies corresponding to the data to be tested correspond one-to-one with each sub-model in the target model.

[0056] In this step, when hydraulic oil deterioration detection is required, corresponding oil parameters can be obtained from various oil sensors. These parameters are used as the data to be detected, and an instruction containing this data is generated. This instruction is then sent to the server. Upon receiving the instruction, the server parses it to obtain the data to be detected. The parsed data is then input into the target model, and multiple evidence bodies corresponding to the data to be detected are determined based on the output of each sub-model in the target model.

[0057] In one example, the sub-model in the target model can be a BPNN (Back Propagation Neural Network), an SVM (Support Vector Machine), or an ELM (Extreme Learning Machine).

[0058] Furthermore, the training process for the target model can include:

[0059] Oil feature data is acquired and normalized to obtain feature parameter data. A diagnostic model is then established based on the feature parameter data. The feature parameter data is divided into training and testing sets according to a preset ratio. The diagnostic model is iteratively trained using the training and testing sets. When the latest diagnostic model meets preset conditions, it is determined as the target model. The diagnostic model includes multiple sub-models, and the number of sub-models in the diagnostic model is consistent with the target model. The preset conditions can be that the number of iterations corresponding to the diagnostic model reaches a preset number, or that the absolute value of the difference between the parameter values ​​of the diagnostic model and the parameter values ​​of the previous iteration is less than a preset threshold. This application does not impose specific limitations on this.

[0060] In one example, the preset ratio can be 7:3, where the former represents the proportion of the training set and the latter represents the proportion of the test set. Oil characteristic data can include temperature, density, kinematic viscosity, dielectric constant, moisture content, solid particles, flash point, pour point, acid value, base value, and calorific value, etc.

[0061] It is understandable that iteratively training the diagnostic model using the training set and the test set means iteratively training each sub-model of the diagnostic model using the training set and the test set respectively.

[0062] Step S102: Determine the target optimization function based on the multiple evidence bodies corresponding to the data to be detected and the preset optimization function.

[0063] The optimization objective of the optimization function is to minimize the sum of squared distances between each piece of evidence and the centroid of the evidence matrix formed by all the pieces of evidence. This allows for a more accurate match between each piece of evidence and its centroid, enhancing the compactness of the clustering and ultimately improving the accuracy of the detection results obtained after fusing the various pieces of evidence.

[0064] In this step, multiple pieces of evidence corresponding to the data to be detected are substituted into the optimization function to obtain the target optimization function. In one example, the construction process of the optimization function can be represented as follows:

[0065] Assume the weight of evidence is W = (w1, w2, ..., w n ), where 0≤w i ≤1 and The evidence matrix M consists of n sets of evidence, and can be represented as M = [E1; E2; ...; E...]. i …;E n ]. Where n represents the number of sub-models in the target model.

[0066] (1) The centroid vector ME can be calculated from the evidence matrix M. At this time, the evidence body E i Distance d between the centroid vector and the centroid vector iIt can be represented as follows:

[0067]

[0068] In the formula, n is the number of sub-models in the target model, and w i E represents the weight of evidence i. i Let i represent evidence body i, and ME be the centroid vector of the evidence matrix composed of all evidence bodies. <w i E i ME> represents w i E i The dot product of ME and ME.

[0069] (2) Based on the formula obtained in step (1), an optimization function can be established with the optimization objective being to minimize the sum of squares of the distances between each piece of evidence and the centroid of the evidence matrix formed by all the pieces of evidence.

[0070] Step S103: Determine the corrected weights for each piece of evidence corresponding to the data to be detected based on the objective optimization function.

[0071] Among them, the correction weight is used to correct the corresponding evidence body.

[0072] In this step, when the objective optimization function is determined, the corrected weights for each piece of evidence corresponding to the data to be detected can be determined by solving the objective optimization function. It is understood that various optimization algorithms, including particle swarm optimization, can be used to solve the objective optimization function. This application does not impose specific limitations in this regard.

[0073] Step S104: Correct the corresponding evidence body by using the correction weight of each evidence body corresponding to the data to be detected, and obtain the target evidence body for each evidence body corresponding to the data to be detected.

[0074] Among them, the target evidence body is the modified evidence body, and the target evidence body corresponding to the data to be detected corresponds one-to-one with the evidence body.

[0075] In this step, the correction weights of each piece of evidence corresponding to the data to be detected are used to correct the corresponding pieces of evidence, thereby obtaining the corrected pieces of evidence, that is, the multiple target pieces of evidence corresponding to the data to be detected.

[0076] In one example, suppose the data to be detected contains three pieces of evidence: evidence A, evidence B, and evidence C. Evidence A has a corrected weight of weight a, evidence B has a corrected weight of weight b, and evidence C has a corrected weight of weight c. Then, by correcting evidence A using weight a, we obtain target evidence A1 corresponding to evidence A; by correcting evidence B using weight b, we obtain target evidence B1 corresponding to evidence B; and by correcting evidence C using weight c, we obtain target evidence C1 corresponding to evidence C. In this case, the multiple target pieces of evidence corresponding to the data to be detected are target evidence A1, target evidence B1, and target evidence C1.

[0077] Step S105: Based on the multiple target evidence bodies corresponding to the data to be detected, determine the probability allocation function of each sub-model in the target model on the data to be detected, so as to generate the fusion detection result corresponding to the data to be detected.

[0078] The probability allocation function is a function composed of the probabilities of various degradation scenarios occurring. The independent variable of the probability allocation function is the degradation scenario, and the dependent variable is the probability.

[0079] In this step, based on multiple target evidence bodies corresponding to the data to be detected, the probability allocation function of each sub-model on the data to be detected is determined. Then, evidence is synthesized on the probability allocation functions of each sub-model on the data to be detected to obtain the target probability allocation function. Finally, a fusion detection result corresponding to the data to be detected is generated based on the target probability allocation function.

[0080] In one example, the target probability assignment function A can be expressed as:

[0081] A(H1) = 0.6, where H1 represents the case of high remaining lifetime and A(H1) represents the probability of high remaining lifetime.

[0082] A(H2) = 0.3, where H2 represents the case of moderate remaining lifespan and A(H2) represents the probability of moderate remaining lifespan.

[0083] A(H3) = 0.1, where H3 represents the case of low remaining lifetime and A(H3) represents the probability of low remaining lifetime.

[0084] By combining the target probability assignment function A, the fusion detection result can be obtained. In this case, if A(H1) is relatively large, a result indicating that the hydraulic oil has a high remaining life can be generated, and this result is used as the fusion detection result. Assuming that A(H3) is relatively large, a result indicating that the hydraulic oil needs to be replaced or maintained can be generated, and this result is used as the fusion detection result.

[0085] This application provides a hydraulic oil degradation detection method, apparatus, storage medium, and computer equipment. The method includes: acquiring data to be tested during hydraulic oil degradation detection and inputting it into a preset target model, which includes multiple sub-models; determining multiple evidence bodies corresponding to the data to be tested; and determining a target optimization function based on a preset optimization function. Then, based on the target optimization function, determining the correction weight of each evidence body to correct the corresponding evidence body, obtaining multiple target evidence bodies corresponding to the data to be tested, and determining the probability allocation function of each sub-model on the data to be tested, ultimately generating a detection result corresponding to the data to be tested. By correcting the multiple evidence bodies determined based on the multiple sub-models in the target model, the weight of the output results of each sub-model can be comprehensively considered, improving the reliability of each evidence body corresponding to the data to be tested. Furthermore, the probability allocation function of each sub-model on the data to be tested is determined to obtain a fused detection result. Thus, by fusing information obtained from multiple sub-models, the state of hydraulic oil can be comprehensively analyzed from multiple dimensions, thereby improving the accuracy of hydraulic oil degradation detection.

[0086] In one embodiment, the training process of the target model uses oil data collected by various oil sensors under various operating conditions as training data.

[0087] Among them, oil sensors refer to sensors used to detect and monitor the state of liquids (such as oil, lubricating oil, etc.) in mechanical equipment. Various oil sensors can collect oil parameters such as temperature, density, kinematic viscosity, dielectric constant, moisture content, solid particles, flash point, pour point, acid value, alkalinity, and calorific value of hydraulic oil.

[0088] In this embodiment, multiple hydraulic parameters can be collected using various hydraulic sensors, and the collected data can be used as training data. This avoids relying on a single indicator to detect hydraulic oil degradation, enriches the training data of the target model, and improves the generalization ability of the target model. Moreover, due to the diversity of training data, the target model can comprehensively evaluate the degradation state of hydraulic oil, thereby improving the accuracy of hydraulic oil degradation detection.

[0089] like Figure 2 As shown, in one embodiment, the step of determining the corrected weight of each piece of evidence corresponding to the data to be detected according to the objective optimization function includes:

[0090] Step S201: Solve the objective optimization function according to the pre-agreed method to obtain the evidence weight of each evidence body corresponding to the data to be detected.

[0091] The evidence weight is the weight value that makes the objective optimization function achieve its optimization objective.

[0092] In this step, the objective optimization function can be solved according to a pre-agreed method to obtain the evidence weight of each piece of evidence corresponding to the data to be detected. For example, the Lagrange function and the Newton-Raphson method can be used to solve it.

[0093] Step S202: Determine the target weight.

[0094] The target weight is the largest evidence weight among the evidence weights of each piece of evidence corresponding to the data to be detected.

[0095] Step S203: Calculate the relative weight of each piece of evidence corresponding to the data to be detected based on the evidence weight and target weight of each piece of evidence corresponding to the data to be detected.

[0096] Specifically, the relative weights can be calculated using the following formula:

[0097]

[0098] In the formula, w i Let the weight of evidence i be max{w1,w2,…,w n} represents the target weights, n is the number of sub-models in the target model, and w i ' represents the relative weight of evidence i.

[0099] Step S204: Determine the relative weight of each piece of evidence corresponding to the data to be detected as its corresponding corrected weight.

[0100] Understandably, by calculating corrected weights to modify the multiple evidence bodies determined by multiple sub-models in the target model, the weights of the output results of each sub-model can be comprehensively considered, thereby improving the reliability of each evidence body corresponding to the data to be detected. This, in turn, improves the accuracy of hydraulic oil deterioration detection.

[0101] like Figure 3 As shown, in one embodiment, the step of determining multiple evidence bodies corresponding to the data to be detected includes:

[0102] Step S301: Determine the model type of each sub-model in the target model.

[0103] The model type can include regression and classification. Regression models are used to predict the values ​​of one or more continuous variables. Classification models are used to predict the values ​​of one or more discrete variables.

[0104] In this step, the model type of each sub-model in the target model can be read from the configuration information of the target model, thereby completing the determination of the model type of each sub-model in the target model.

[0105] Step S302: Determine the preset loss function for each sub-model based on its model type in the target model.

[0106] Understandably, there exists a pre-defined mapping table to record the mapping relationship between model types and their corresponding pre-defined loss functions. In this case, based on the model type of each sub-model in the target model, the pre-defined loss function corresponding to that model type can be looked up in the mapping table, thus determining the pre-defined loss function for each sub-model.

[0107] Step S303: Obtain the output results of the data to be detected on each sub-model in the target model.

[0108] Step S304: Based on the preset loss function corresponding to each sub-model and the output results of the data to be detected on each sub-model in the target model, determine the target loss function of the data to be detected on each sub-model of the target model.

[0109] In this step, the output of the data to be detected on each sub-model is substituted into the preset loss function corresponding to that sub-model to obtain the target loss function of the data to be detected on that sub-model. This determines the target loss function of the data to be detected on each sub-model of the target model.

[0110] Step S305: Determine the target loss function of the data to be detected on each sub-model of the target model as the evidence body corresponding to the data to be detected.

[0111] In one example, suppose submodel A is a backpropagation neural model, belonging to the regression type. Its evidence body output can be described by the approximation error of the neural network, as expressed below:

[0112]

[0113] In the formula, E represents the evidence body, l represents the number of neurons in sub-model A, and y i k represents the output of the i-th neuron in sub-model A. i This represents the expected output of the i-th neuron in sub-model A. It is understandable that, since the obtained evidence body is needed as the basic data for judging the state of hydraulic oil, accurately describing the output of the sub-model is crucial. Therefore, using the corresponding loss function to describe the output of the corresponding sub-model, based on different model types, can improve the accuracy of the evidence body, thereby improving the accuracy of hydraulic oil degradation detection.

[0114] like Figure 4As shown, in one embodiment, the step of determining the probability assignment function of each sub-model in the target model on the data to be detected includes:

[0115] Step S401: Determine the preset probability function corresponding to each sub-model.

[0116] Step S402: Obtain the output results of the data to be detected on each sub-model in the target model.

[0117] Step S403: For each sub-model in the target model, substitute the output result of the data to be detected on the sub-model and the target evidence body corresponding to the sub-model among the multiple target evidence bodies corresponding to the data to be detected into the preset probability function corresponding to the sub-model to obtain the probability allocation function of the sub-model on the data to be detected.

[0118] In this embodiment, step S403 is performed on each sub-model in the target model to obtain the probability allocation function of each sub-model on the data to be detected.

[0119] In one example, assuming sub-model A is a backpropagation neural model or an extreme learning machine model, the probability assignment function of sub-model A on the data to be tested can be determined by the following formula:

[0120]

[0121] In the formula, S represents the population, i.e., the sample space, and y(A) j ) represents the output of submodel A for the j-th type of degradation, m(A j Let S denote the probability assignment function for the j-th type of degradation. The expression for S is as follows:

[0122]

[0123] In the formula, E represents the evidence body, and l represents the number of types of deterioration.

[0124] Assuming sub-model B is an SVM model, its corresponding probability assignment function can be expressed by the following formula:

[0125] m(A i ) = rp i

[0126] Wherein, m(A) i ) indicates that the output result is A i The basic probability assignment, r is the accuracy of the output result of submodule B corresponding to the training set, p i The i-th class result is output by sub-model B for the test sample.

[0127] Understandably, by determining the probability allocation function of each sub-model on the data to be tested, the probability distribution of hydraulic oil degradation can be considered from multiple dimensions. This aims to comprehensively assess the condition of the hydraulic oil and ultimately improve the accuracy of hydraulic oil degradation detection.

[0128] In one embodiment, the step of correcting the corresponding evidence body using the correction weights of each evidence body corresponding to the data to be detected includes:

[0129] For each piece of evidence corresponding to the data to be detected, the product of the correction weight corresponding to the evidence and the evidence itself is determined as the target evidence of the sub-model corresponding to the evidence on the data to be detected, so as to complete the correction of the evidence of the sub-model corresponding to the evidence on the data to be detected.

[0130] In this embodiment, for each piece of evidence corresponding to the data to be detected, the product of the correction weight corresponding to the evidence and the evidence itself is determined as the target evidence for the sub-model corresponding to the evidence on the data to be detected, i.e., it is determined as the target evidence. This completes the correction for each piece of evidence corresponding to the data to be detected.

[0131] In one example, the process of modifying the body of evidence can be represented by the following expression:

[0132] M' E =(E1',E'2,…,E') n )=(w1E1,w2E2,…,w n E n )

[0133] In the formula, w1E1, w2E2, ..., w n E n E1', E'2, ..., E' represents the product of the modified weights corresponding to each piece of evidence and the corresponding piece of evidence. n M' represents each target piece of evidence. E This is the corrected evidence matrix.

[0134] It is understandable that by correcting the evidence body corresponding to the test data, the accuracy between different sub-models can be taken into account, and different correction weights can be assigned. Then, by using the correction weights to correct the corresponding evidence body, the accuracy of the global contribution of each evidence body can be improved, thereby improving the accuracy of hydraulic oil deterioration detection.

[0135] In one embodiment, the expression for the optimization function is:

[0136]

[0137] In the formula, The function represents the optimization function, where n is the number of sub-models in the target model, and w is the value of w. i E represents the weight of evidence i. i Let i represent evidence body i, and ME be the centroid vector of the evidence matrix composed of all evidence bodies. <w i E i ME> represents w i E i The dot product of ME, d i Let i be the distance between the evidence body i and the ME vector.

[0138] Where, 0≤w i ≤1 and

[0139] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0140] The hydraulic oil deterioration detection device provided in the embodiments of this application is described below. The hydraulic oil deterioration detection device described below can be referred to in correspondence with the hydraulic oil deterioration detection method described above.

[0141] like Figure 5 As shown, this application provides a hydraulic oil deterioration detection device 500, the device comprising:

[0142] The data acquisition module 501 is used to acquire the data to be detected and input the data to be detected into a pre-trained target model to determine multiple evidence bodies corresponding to the data to be detected; wherein, the target model includes multiple sub-models;

[0143] The function determination module 502 is used to determine the target optimization function based on multiple evidence bodies corresponding to the data to be detected and a preset optimization function;

[0144] The weight determination module 503 is used to determine the corrected weight of each piece of evidence corresponding to the data to be detected based on the objective optimization function.

[0145] The evidence body correction module 504 is used to correct the corresponding evidence body by using the correction weight of each evidence body corresponding to the data to be detected, so as to obtain the target evidence body for each evidence body corresponding to the data to be detected.

[0146] The probability detection module 505 is used to determine the probability allocation function of each sub-model in the target model on the data to be detected based on multiple target evidence bodies corresponding to the data to be detected, so as to generate a fusion detection result corresponding to the data to be detected.

[0147] In one embodiment, the weight determination module includes:

[0148] The evidence weight determination submodule is used to solve the target optimization function according to a pre-agreed method to obtain the evidence weight of each evidence body corresponding to the data to be detected;

[0149] The target weight determination submodule is used to determine the target weight; whereby the target weight is the largest evidence weight among the evidence weights of each piece of evidence corresponding to the data to be detected.

[0150] The relative weight determination submodule is used to calculate the relative weight of each piece of evidence corresponding to the data to be detected based on the evidence weight and the target weight of each piece of evidence corresponding to the data to be detected, respectively.

[0151] The weight determination submodule is used to determine the relative weight of each piece of evidence corresponding to the data to be detected as its corresponding corrected weight.

[0152] In one embodiment, the data acquisition module includes:

[0153] The model type determination submodule is used to determine the model type of each sub-model in the target model;

[0154] The loss function determination submodule is used to determine the preset loss function for each sub-model based on the model type of each sub-model in the target model.

[0155] The output acquisition submodule is used to acquire the output results of the data to be detected on each sub-model in the target model;

[0156] The loss function determination submodule is used to determine the target loss function of the data to be detected on each sub-model of the target model based on the preset loss function corresponding to each sub-model and the output results of the data to be detected on each sub-model in the target model.

[0157] The evidence body determination submodule is used to determine the target loss function of the data to be detected on each sub-model of the target model as the evidence body corresponding to the data to be detected.

[0158] In one embodiment, the probability detection module includes:

[0159] The function determines the submodule, which is used to determine the preset probability function corresponding to each sub-model;

[0160] The result acquisition submodule is used to acquire the output results of the data to be detected on each sub-model in the target model;

[0161] The probability detection submodule is used to, for each submodel in the target model, substitute the output result of the data to be detected on the submodel and the target evidence body corresponding to the submodel among the multiple target evidence bodies of the data to be detected into the preset probability function corresponding to the submodel, so as to obtain the probability allocation function of the submodel on the data to be detected.

[0162] In one embodiment, the evidence body correction module includes:

[0163] The evidence body correction sub-model is used to determine the target evidence body of the sub-model corresponding to the evidence body on the data to be detected by multiplying the correction weight of the evidence body with the evidence body for each evidence body corresponding to the data to be detected, so as to complete the correction of the evidence body of the sub-model corresponding to the evidence body on the data to be detected.

[0164] The division of modules in the above-described hydraulic oil deterioration detection device is merely illustrative. In other embodiments, the hydraulic oil deterioration detection device can be divided into different modules as needed to complete all or part of its functions. Each module in the above-described hydraulic oil deterioration detection device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0165] In one embodiment, this application also provides a storage medium storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the hydraulic oil deterioration detection method as described in any of the above embodiments.

[0166] In one embodiment, this application also provides a computer device storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the hydraulic oil deterioration detection method as described in any of the above embodiments.

[0167] Indicatively, such as Figure 6 As shown, Figure 6This is a schematic diagram of the internal structure of a computer device 600 provided in an embodiment of this application. The computer device 600 can be provided as a server. (Refer to...) Figure 6 The computer device 600 includes a processing component 602, which further includes one or more processors, and memory resources represented by memory 601 for storing instructions, such as application programs, that can be executed by the processing component 602. The application programs stored in memory 601 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 602 is configured to execute instructions to perform the hydraulic oil deterioration detection method of any of the above embodiments.

[0168] The computer device 600 may also include a power supply component 603 configured to perform power management of the computer device 600, a wired or wireless network interface 604 configured to connect the computer device 600 to a network, and an input / output (I / O) interface 605. The computer device 600 may operate on an operating system stored in memory 601, such as Windows Server™, Mac OS X™, Unix™, Linux™, Free BSD™, or similar.

[0169] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0170] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising a…" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. In this document, the singular forms "a," "an," and "the" may also include the plural forms unless the context clearly indicates otherwise. It should also be understood that the terms “comprising / including” or “having” specify the presence of the stated features, wholes, steps, operations, components, parts or combinations thereof, but do not exclude the possibility of the presence or addition of one or more other features, wholes, steps, operations, components, parts or combinations thereof. Meanwhile, the term “and / or” as used in this specification includes any and all combinations of the associated listed items.

[0171] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referred to each other.

[0172] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting hydraulic oil deterioration, characterized in that, The method includes: The process involves acquiring data to be detected and inputting it into a pre-trained target model to determine multiple evidence bodies corresponding to the data to be detected; wherein the target model includes multiple sub-models. Based on the multiple evidence bodies corresponding to the data to be detected and the preset optimization function, the target optimization function is determined; Based on the objective optimization function, determine the corrected weights for each piece of evidence corresponding to the data to be detected; The corresponding evidence body is corrected by using the correction weight of each evidence body corresponding to the data to be detected, so as to obtain the target evidence body for each evidence body corresponding to the data to be detected. Based on the multiple target evidence bodies corresponding to the data to be detected, determine the probability allocation function of each sub-model in the target model on the data to be detected, so as to generate a fusion detection result corresponding to the data to be detected; The step of determining the multiple evidence bodies corresponding to the data to be detected includes: Determine the model type of each sub-model in the target model; Based on the model type of each sub-model in the target model, determine the preset loss function corresponding to each sub-model; Obtain the output results of the data to be detected on each sub-model in the target model; Based on the preset loss function corresponding to each sub-model and the output results of the data to be detected on each sub-model of the target model, the target loss function of the data to be detected on each sub-model of the target model is determined. The target loss function of the data to be detected on each sub-model of the target model is determined as the evidence body corresponding to the data to be detected; The step of determining the probability assignment function of each sub-model in the target model on the data to be detected includes: Determine the preset probability function for each sub-model; Obtain the output results of the data to be detected on each sub-model in the target model; For each sub-model in the target model, the output result of the data to be detected on the sub-model and the target evidence body corresponding to the sub-model among the multiple target evidence bodies corresponding to the data to be detected are substituted into the preset probability function corresponding to the sub-model to obtain the probability allocation function of the sub-model on the data to be detected. The expression for the optimization function is: In the formula, Represents the optimization function. The number of sub-models in the target model. For evidence Weight of evidence Representative evidence body , Let be the centroid vector of the evidence matrix composed of all the evidence pieces. represent and dot product, For evidence and The distance between vectors.

2. The hydraulic oil deterioration detection method according to claim 1, characterized in that, The training process of the target model uses oil data collected by various oil sensors under various working conditions as training data.

3. The hydraulic oil deterioration detection method according to claim 1, characterized in that, The step of determining the corrected weight of each piece of evidence corresponding to the data to be detected based on the objective optimization function includes: The objective optimization function is solved according to a pre-agreed method to obtain the evidence weight of each evidence body corresponding to the data to be detected; Determine the target weight; wherein, the target weight is the largest evidence weight among the evidence weights of each evidence body corresponding to the data to be detected; The relative weight of each piece of evidence corresponding to the data to be detected is calculated based on the evidence weight of each piece of evidence corresponding to the data to be detected and the target weight, respectively. The relative weight of each piece of evidence corresponding to the data to be detected is determined as its corresponding corrected weight.

4. The hydraulic oil deterioration detection method according to any one of claims 1 to 3, characterized in that, The step of correcting the corresponding evidence body by using the correction weights of each evidence body corresponding to the data to be detected includes: For each piece of evidence corresponding to the data to be detected, the product of the correction weight corresponding to the evidence and the evidence is determined as the target evidence of the sub-model corresponding to the evidence on the data to be detected, so as to complete the correction of the evidence of the sub-model corresponding to the evidence on the data to be detected.

5. A hydraulic oil deterioration detection device, characterized in that, The device includes: A data acquisition module is used to acquire data to be detected and input the data to be detected into a pre-trained target model to determine multiple evidence bodies corresponding to the data to be detected; wherein, the target model includes multiple sub-models; The function determination module is used to determine the target optimization function based on multiple evidence bodies corresponding to the data to be detected and a preset optimization function; The weight determination module is used to determine the corrected weight of each piece of evidence corresponding to the data to be detected based on the target optimization function. The evidence body correction module is used to correct the corresponding evidence body by using the correction weight of each evidence body corresponding to the data to be detected, so as to obtain the target evidence body for each evidence body corresponding to the data to be detected. The probability detection module is used to determine the probability allocation function of each sub-model in the target model on the data to be detected based on multiple target evidence bodies corresponding to the data to be detected, so as to generate a fusion detection result corresponding to the data to be detected; The data acquisition module includes: Determine the model type of each sub-model in the target model; Based on the model type of each sub-model in the target model, determine the preset loss function corresponding to each sub-model; Obtain the output results of the data to be detected on each sub-model in the target model; Based on the preset loss function corresponding to each sub-model and the output results of the data to be detected on each sub-model of the target model, the target loss function of the data to be detected on each sub-model of the target model is determined. The target loss function of the data to be detected on each sub-model of the target model is determined as the evidence body corresponding to the data to be detected; The probability detection module includes: Determine the preset probability function for each sub-model; Obtain the output results of the data to be detected on each sub-model in the target model; For each sub-model in the target model, the output result of the data to be detected on the sub-model and the target evidence body corresponding to the sub-model among the multiple target evidence bodies corresponding to the data to be detected are substituted into the preset probability function corresponding to the sub-model to obtain the probability allocation function of the sub-model on the data to be detected. The expression for the optimization function is: In the formula, Represents the optimization function. The number of sub-models in the target model. For evidence Weight of evidence Representative evidence body , Let be the centroid vector of the evidence matrix composed of all the evidence pieces. represent and dot product, For evidence and The distance between vectors.

6. A storage medium, characterized in that: The storage medium stores computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the steps of the hydraulic oil deterioration detection method as described in any one of claims 1 to 4.

7. A computer device, characterized in that, include: One or more processors, and memory; The memory stores computer-readable instructions, which, when executed by the one or more processors, perform the steps of the hydraulic oil deterioration detection method as described in any one of claims 1 to 4.

Images