Device predictive maintenance method and device based on artificial intelligence

Abstract

The embodiment of the invention provides an equipment predictive maintenance method and device based on artificial intelligence, relates to the technical field of data processing, and aims to adjust a model focus by identifying statistical characteristics of data and paying attention to characteristics with most significant changes. Moreover, the training sample set of the feature extraction model is adjusted based on the abnormal evaluation result, abnormal data can be suppressed, the model is protected from being interfered by the data, and the feature extraction effect can be optimized. Quantum state modeling can better capture a nonlinear relation in data, information loss of a traditional linear dimension reduction algorithm is avoided, and therefore the expressive power of features after dimension reduction is improved. According to the embodiment of the invention, through data compression and optimization feature extraction based on quantum state amplitude, the prediction precision and reliability of the model are improved, the quality of data input into the classifier model is ensured, and the risk of misjudgment is reduced.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an artificial intelligence-based equipment predictive maintenance method and device. Background Art

[0002] With the development of industrial automation and information technology, prebaked anode production equipment plays an important role in the aluminum electrolysis industry. The operating status of these equipment directly affects production efficiency and product quality. Traditional equipment maintenance methods rely on regular inspections and empirical judgments, which are not only time-consuming, but also difficult to achieve real-time and accurate assessment of equipment status. In this context, intelligent equipment health management technology is particularly important. Using advanced data analysis technologies such as machine learning and artificial intelligence, equipment status can be monitored in real time and potential failures can be predicted, thereby reducing downtime, reducing maintenance costs, and improving production efficiency.

[0003] Traditional methods usually rely on batch processing or periodic sampling, which leads to time delays in data processing and feature extraction. Moreover, when sudden abnormal conditions occur during equipment operation, traditional methods fail to effectively filter these abnormal data, resulting in model misjudgment. In addition, traditional dimensionality reduction algorithms are mostly based on linear transformations and cannot effectively process data with nonlinear structures, resulting in information loss. Due to information loss and linear limitations, the features after dimensionality reduction are difficult to fully reflect the complexity and diversity of the original data, affecting the performance of the classifier. Summary of the invention

[0004] In view of this, the object of the present invention is to provide an equipment predictive maintenance method and device based on artificial intelligence, which can improve the prediction accuracy of equipment failure status.

[0005] In a first aspect, an embodiment of the present invention provides an artificial intelligence-based equipment predictive maintenance method, wherein the method includes: obtaining operating status monitoring data of a target device; analyzing the statistical characteristics of the status monitoring data, and identifying key features in the operating status monitoring data based on the statistical characteristics; wherein, key features are identified using a pre-built feature extraction model, and a training sample set for training the feature extraction model is adjusted based on the results of an abnormality assessment; the quantum state of the key features is simulated, and data compression is performed on the key features based on the quantum state amplitude of the quantum state to generate parameters to be measured; the pre-built classifier model is used to classify and predict the parameters to be measured, and the equipment status prediction results of the parameters to be measured are output; and the predicted fault of the target device is determined based on the equipment status prediction results.

[0006] In combination with the first aspect, an embodiment of the present invention provides a first implementation of the first aspect, wherein the step of determining the predicted failure of the target device based on the device state prediction result includes: obtaining the device predicted state indicated by the device state prediction result; and determining the device predicted state as the predicted failure of the target device.

[0007] In combination with the first aspect, an embodiment of the present invention provides a second implementation of the first aspect, wherein the step of analyzing the statistical characteristics of the status monitoring data and identifying the key features in the operating status monitoring data based on the statistical characteristics includes: calculating the information gain of each feature of the status monitoring data; determining the statistical characteristics of the status monitoring data based on the information gain; and determining the features whose information gain exceeds a preset gain threshold as the key features in the operating status monitoring data.

[0008] In combination with the first aspect, an embodiment of the present invention provides a third implementation of the first aspect, wherein the step of adjusting the training sample set of the training feature extraction model based on the results of the abnormality evaluation includes: training a preset neural network using a preset training sample set to determine the historical mean and standard deviation of the training sample set; determining the abnormality evaluation results of the training sample set based on the historical mean and standard deviation; and adjusting the feature values ​​of the features of the training sample set based on the results of the abnormality evaluation.

[0009] In combination with the first aspect, an embodiment of the present invention provides a fourth implementation of the first aspect, wherein the steps of simulating the quantum state of the key feature, compressing the data of the key feature based on the quantum state amplitude of the quantum state, and generating the parameter to be measured include: converting the key feature into a quantum bit, calculating the quantum noise level of the key feature; determining the probability amplitude corresponding to the quantum bit; and compressing the key feature based on the probability amplitude and the quantum noise level.

[0010] In combination with the first aspect, an embodiment of the present invention provides a fifth implementation of the first aspect, wherein the step of compressing key features based on probability amplitude and quantum noise level includes: converting the key features into low-dimensional feature representations through a preset autoencoder, and compressing the key features; wherein the training method of the autoencoder includes: training the preset autoencoder using a preset training sample set; evaluating the data complexity of the training sample set based on the information entropy of the training sample set; adjusting the coding layer dimension of the autoencoder according to the data complexity and a preset coding dimension adjustment coefficient; and adjusting the output layer of the autoencoder based on the adjusted coding layer dimension.

[0011] In combination with the first aspect, an embodiment of the present invention provides a sixth implementation of the first aspect, wherein a method for constructing a classifier model includes: training a preset neural decision tree using a preset training sample set, and calculating a feature influence parameter matrix of the preset training sample set; calculating the feature influence weight of the training sample set based on the feature influence parameter matrix and the learning status of the neural decision tree; determining the classification error of each decision path of the neural decision tree based on the feature influence weight; optimizing the decision rules of the neural decision tree based on the classification error, and determining the matrix update amount of the feature influence parameter matrix; until the neural decision tree meets the preset training accuracy, freezing the parameters of the neural decision tree according to the preset parameter ratio, and constructing a classifier model.

[0012] In combination with the first aspect, an embodiment of the present invention provides a seventh implementation of the first aspect, wherein the above method also includes: obtaining operating status monitoring samples of a preset device, labeling the operating status monitoring samples according to the fault status corresponding to the preset device, and constructing an initial sample set; using a preset generator of a generative adversarial network to perform sample expansion on the initial sample set to generate initial expanded samples; using a discriminator of the generative adversarial network to discriminate the initial sample set and the initial expanded samples respectively; updating the generator parameters of the generator based on the discrimination results of the initial sample set and the discrimination results of the initial expanded samples to update the initial expanded samples; until the initial expanded samples meet the preset expansion requirements, constructing a training sample set based on the initial sample set and the initial expanded samples.

[0013] In combination with the first aspect, an embodiment of the present invention provides an eighth implementation of the first aspect, wherein the step of updating the generator parameters of the generator based on the discrimination results of the initial sample set and the discrimination results of the initial expanded samples includes: determining the loss function of the generator based on the discrimination results of the initial sample set and the discrimination results of the initial expanded samples; determining the fitness function of the generative adversarial network based on the dynamic changes corresponding to the generative adversarial network; and updating the generator parameters of the generator based on the loss function and the fitness function.

[0014] In a second aspect, an embodiment of the present invention provides an artificial intelligence-based equipment predictive maintenance device, wherein the device includes: a data acquisition module, used to acquire the operating status monitoring data of the target equipment; a feature extraction module, used to analyze the statistical characteristics of the status monitoring data, and identify key features in the operating status monitoring data; wherein a pre-built feature extraction model is used to identify key features, and a training sample set for training the feature extraction model is adjusted based on the results of abnormality assessment; a data processing module, used to simulate the quantum state of the key features, and based on the quantum state amplitude of the quantum state, the key features are compressed to generate parameters to be measured; an execution module, used to classify and predict the parameters to be measured using a pre-built classifier model, and output the equipment status prediction results of the parameters to be measured; an output module, used to determine the predicted fault of the target equipment based on the equipment status prediction results.

[0015] The embodiments of the present invention bring about the following beneficial effects: The embodiments of the present invention provide an artificial intelligence-based equipment predictive maintenance method and device, which adjusts the model focus by identifying the statistical characteristics of the data, focusing on the features with the most significant changes. Moreover, the training sample set of the feature extraction model is adjusted based on the abnormal evaluation results, which can suppress abnormal data, protect the model from interference from these data, and optimize the effect of feature extraction. Quantum state modeling can better capture nonlinear relationships in the data, avoid information loss in traditional linear dimensionality reduction algorithms, and thus improve the expressiveness of features after dimensionality reduction. The embodiments of the present invention improve the prediction accuracy and reliability of the model by compressing data based on quantum state amplitude and optimizing feature extraction, thereby ensuring the quality of data input to the classifier model and reducing the risk of misjudgment.

[0016] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description and the drawings.

[0017] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1A flowchart of an artificial intelligence-based equipment predictive maintenance method provided by an embodiment of the present invention; Figure 2 A flowchart of another artificial intelligence-based equipment predictive maintenance method provided by an embodiment of the present invention; Figure 3 A flowchart of a method for constructing a training sample set provided by an embodiment of the present invention; Figure 4 A schematic diagram of the structure of an artificial intelligence-based equipment predictive maintenance device provided in an embodiment of the present invention; Figure 5 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the following embodiments of the present disclosure are described by specific specific examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all of the embodiments. The present disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present disclosure.

[0021] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described in the present invention may be embodied in a wide variety of forms, and any specific structure and / or function described in the present invention is merely illustrative. Based on the present disclosure, it should be understood by those skilled in the art that an aspect described in the present invention may be implemented independently of any other aspect, and two or more of these aspects may be combined in various ways. For example, any number of aspects described in the present invention may be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described in the present invention may be used to implement this device and / or practice this method.

[0022] It should also be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present disclosure. The drawings only show components related to the present disclosure rather than being drawn according to the number, shape and size of components in actual implementation. The type, quantity and proportion of each component in actual implementation may be changed at will, and the component layout type may also be more complicated. In addition, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the aspects described can be practiced without these specific details.

[0023] The present invention provides an artificial intelligence-based equipment predictive maintenance method and device, which can improve the prediction accuracy of equipment failure status.

[0024] For ease of understanding, an artificial intelligence-based equipment predictive maintenance method provided by an embodiment of the present invention is first described. Figure 1 A flowchart of an artificial intelligence-based equipment predictive maintenance method provided by an embodiment of the present invention is shown, referring to Figure 1 , the method comprises the following steps: Step S102, obtaining the operating status monitoring data of the target device.

[0025] The data of the present invention is collected from data acquired by various sensors during the operation of the equipment, including but not limited to temperature sensors, vibration sensors, pressure sensors and current sensors. All collected data are stored in a standardized JSON format. In one embodiment, the attributes of the data include: Ta (temperature), indicating the operating temperature of the equipment components; Va (vibration frequency), indicating the vibration frequency generated when the equipment is running; Pa (pressure), indicating the working pressure of the liquid or gas in the equipment; Ia (current), indicating the current size when the equipment is running; Ra (rotation speed), indicating the rotation speed of the equipment components; Sa (sound level), indicating the sound level generated when the equipment is running; Ha (humidity), indicating the humidity of the environment around the equipment; Oa (oil quality), indicating the quality status of the lubricating oil; Ua (use time), indicating the operation time of the equipment since the last maintenance; Ca (configuration parameters), indicating the configuration parameters of the equipment operation. It should be noted that this embodiment only illustrates a data format and type of the present invention. In actual applications, the attributes of the data are usually more than 10 attributes, and the number of attributes of the data may reach dozens or even hundreds.

[0026] Step S104: Analyze the statistical characteristics of the status monitoring data, and identify key features in the operating status monitoring data based on the statistical characteristics.

[0027] The embodiment of the present invention performs detailed statistical analysis on the state monitoring data and identifies important statistics. In the embodiment of the present invention, by analyzing the statistical characteristics of the input data, the most significant features of the changes are identified, and those features that have a significant impact on the changes in the model output are captured. According to the importance of the captured features, the focus of the network is adjusted to prioritize those key features. Based on the statistical characteristics, the key features that can reflect the changes in the operating status of the equipment are selected.

[0028] Among them, the embodiment of the present invention uses a pre-built feature extraction model to identify key features, and the training sample set of the feature extraction model is adjusted based on the results of abnormality assessment, the corresponding features are suppressed, the model is protected from interference from these data, and the effect of feature extraction is guaranteed.

[0029] Step S106, simulating the quantum state of the key feature, and based on the quantum state amplitude of the quantum state, performing data compression on the key feature to generate a parameter to be measured.

[0030] Quantum state amplitude can significantly reduce data dimensions and improve computing efficiency while ensuring information integrity. The embodiments of the present invention combine quantum states to perform data compression to better capture nonlinear relationships in data and avoid information loss in traditional linear dimensionality reduction algorithms.

[0031] Step S108, using the pre-built classifier model to perform classification prediction on the parameters to be measured, and outputting the device state prediction results of the parameters to be measured.

[0032] Step S110: determining the predicted fault of the target device according to the device status prediction result.

[0033] Among them, the predicted state of the device indicated by the device state prediction result can be obtained, and the predicted state of the device can be determined as the predicted fault of the target device. In one embodiment, the collected raw data is input into the trained feature extraction and feature dimension reduction model for feature processing, and further, the processed features are input into the classifier model for classification, and then the classification results are obtained. In this embodiment, the classification categories include "early faults of bearing wear", "early faults of motor overheating", "early faults of insufficient lubrication" and other early states of various faults, and correspondingly, they are output in the form of corresponding labels, such as the specific numbers (0-9) indicated by the labels.

[0034] In summary, an embodiment of the present invention provides an artificial intelligence-based equipment predictive maintenance method, which adjusts the model focus by identifying the statistical characteristics of the data, focusing on the features with the most significant changes. Moreover, the training sample set of the feature extraction model is adjusted based on the abnormal evaluation results, which can suppress abnormal data, protect the model from interference from these data, and optimize the effect of feature extraction. Quantum state modeling can better capture nonlinear relationships in the data, avoid information loss in traditional linear dimensionality reduction algorithms, and thus improve the expressiveness of features after dimensionality reduction. The embodiment of the present invention improves the prediction accuracy and reliability of the model by compressing data based on quantum state amplitude and optimizing feature extraction, ensuring the quality of data input to the classifier model and reducing the risk of misjudgment.

[0035] Furthermore, in combination with the above embodiments, the embodiments of the present invention also provide another equipment predictive maintenance method based on artificial intelligence. Figure 2 A flowchart of another method for predictive maintenance of equipment based on artificial intelligence provided by an embodiment of the present invention is shown, referring to Figure 2 , the method comprises the following steps: Step S202, obtaining the operating status monitoring data of the target device.

[0036] Step S204, calculating the information gain of each feature of the condition monitoring data.

[0037] Step S206: determining the statistical characteristics of the condition monitoring data based on the information gain.

[0038] Step S208: determining the features whose information gain exceeds a preset gain threshold as key features in the operating status monitoring data.

[0039] In specific implementation, the embodiment of the present invention extracts features from the running status monitoring data through a preset feature extraction model to determine the key features in the running status monitoring data. Specifically, feature selection is achieved by analyzing the statistical characteristics in the data stream in real time, calculating the information gain of each feature, and selecting features whose information gain exceeds the threshold, which is expressed as:

[0040] In the formula, s i is the data feature, is the data feature s i The corresponding eigenvalue; Y is the target variable, specifically the equipment fault status label, is the preset information gain threshold; is the information gain of the feature. Preferably, Set to 0.1.

[0041] Furthermore, the present invention adopts feature selection based on information gain, and information gain is realized by mutual information, and the calculation method is expressed as:

[0042] In the formula, is the target variable The entropy of is a given feature hour The conditional entropy of .

[0043] Existing feature extraction methods are difficult to capture changes in key features in real time, and lack strategies to suppress abnormal data. They are easily affected by abnormal data, resulting in a decline in model performance, which has an adverse impact on the accuracy of fault prediction. In the feature extraction process, the embodiments of the present invention not only focus on the statistical characteristics of the features, but also use adaptive abnormal feature suppression strategies to dynamically suppress the influence of abnormal data in the feature extraction model, thereby enhancing the robustness of the model to abnormal data.

[0044] Furthermore, the embodiment of the present invention also adjusts the training sample set of the training feature extraction model based on the result of the abnormal evaluation, and dynamically adjusts the influence of abnormal feature values ​​that may cause the performance of the model to deteriorate by identifying and adjusting them, thereby protecting the model from interference from these data. The abnormal evaluation method is expressed as follows:

[0045] In the formula, and The data features are The historical mean and standard deviation of the corresponding eigenvalues; is the abnormality assessment result, which is close to 1 for normal data and close to 0 for abnormal data, and is divided with a threshold of 0.5.

[0046] Feature suppression and adjustment are performed based on the results of abnormal evaluation, expressed as:

[0047] In the formula, It is a parameter that is dynamically adjusted based on the abnormal evaluation results.

[0048] Furthermore, the calculation method of the parameters dynamically adjusted based on the abnormal evaluation results is expressed as:

[0049] In the formula, is the minimum value function; is the maximum value function; is the threshold parameter for abnormal evaluation, which is used to determine the transition boundary of the feature from abnormal to normal. Preferably, Set to 0.5.

[0050] Furthermore, the embodiment of the present invention also evaluates the feature uncertainty of the neural network for the training sample set. Based on the feature uncertainty, the network parameters of the neural network are updated. Until the neural network meets the preset training requirements, a feature extraction model is constructed.

[0051] Among them, uncertainty estimation is achieved through Bayesian posterior variance. Specifically, the variance of the neural network parameters helps to control the sensitivity of the model to features with high uncertainty. The calculation method is expressed as:

[0052] In the formula, is the posterior distribution The mean of is a given data set, which represents all the data that the network has received; In a given data set The parameters under the condition are The probability distribution function of Indicates the parameters The differential of .

[0053] Uncertainty metrics are used to evaluate the confidence of the current model for a specific input. If the uncertainty of certain features is too high, the model will automatically adjust its parameters to learn these features more finely and reduce the uncertainty of the prediction. The way the network parameters are adjusted is expressed as:

[0054] In the formula, is the learning rate of the neural network; is the regularized learning rate of the neural network; is L2 regularization; For the The parameters of the neural network at the iteration; For the The parameters of the neural network at the iteration; is the regularization coefficient of the neural network, Represents the variance of the neural network parameters; The parameter is When the input is When you get The posterior probability distribution function of ; For the The input features of the neural network at this iteration.

[0055] The parameters of the neural network include the weights and biases of the neural network. In one embodiment, the initialization method is expressed as:

[0056] In the formula, represents the parameters of the neural network model used for feature extraction, represents the variance of the neural network initialization; Indicates compliance with a specific distribution; represents normal distribution; Represents the initial parameters of the neural network model used for feature extraction.

[0057] Among them, the regularized learning rate of the neural network is set by adaptive adjustment, and the calculation method is expressed as:

[0058] In the formula, is the regularized learning rate of the neural network at the t+1th iteration; is the regularized learning rate of the neural network at the tth iteration; and is the decay rate of the estimated first and second order moments at iteration t, set to a value close to 1.

[0059] In summary, the above steps are repeated until the preset stop iteration condition is met, which means that the model training is completed. In one embodiment, the preset stop iteration condition is reaching a preset maximum number of iterations. Preferably, the preset maximum number of iterations is set to 1000 times.

[0060] In summary, the embodiment of the present invention adopts a 6-layer fully connected neural network for feature extraction. In the prior art, some solutions use neural networks for feature extraction. In some neural network structures, problems such as gradient vanishing, gradient explosion, or falling into local optimal solutions may be encountered, affecting the stability of training and the performance of the model. The present invention adopts a neural network algorithm based on dynamic feature capture, which can effectively extract key features that may indicate equipment failures and enhance the model's adaptability to features of different fault types.

[0061] Step S210, simulating the quantum state of the key feature, and based on the quantum state amplitude of the quantum state, performing data compression on the key feature to generate a parameter to be measured.

[0062] Existing dimensionality reduction algorithms are limited by the linear limitations of the coding layer, resulting in large information loss, making it difficult to efficiently compress complex device data, and the feature expression after dimensionality reduction is limited, which affects the subsequent classification effect. The embodiment of the present invention converts key features into quantum bits and calculates the quantum noise level of the key features. Determine the probability amplitude corresponding to the quantum bit, and compress the key features based on the probability amplitude and quantum noise level. In quantum computing, a quantum bit can be in a superposition state between 0 and 1, and its probability amplitude defines the probability of observing 0 or 1 if measured, which contains not only amplitude information but also phase information. Multiple quantum systems interact to form entangled states, and their probability amplitudes jointly describe the quantum state of the entire system, reflecting the non-classical correlation between quantum systems. Quantum noise refers to any factor that interferes with ideal quantum operations and affects the evolution of quantum states. Compressing data based on these parameters can retain the most critical information.

[0063] Specifically, the embodiment of the present invention converts the key features into low-dimensional feature representations through a preset autoencoder, and performs data compression on the key features. In the encoding process of the autoencoder, the embodiment of the present invention further compresses the feature representation of the data by simulating the amplitude of the quantum bit to optimize the representation efficiency of the data. Specifically, the amplitude of the quantum bit is used to further compress the data features, which is expressed as:

[0064] In the formula, is the quantum probability amplitude function, which is used to simulate the quantum state; is a normalization constant that ensures that the sum of probabilities is 1; is the parameter that controls the compression sensitivity of the quantum state; yes No. elements; Indicates that it corresponds to The quantum noise level.

[0065] The embodiment of the present invention uses a quantum probability amplitude function to simulate the quantum state and determine the probability amplitude of the quantum bit. The quantum probability amplitude function assigns an importance weight to each feature in the form of probability distribution. Important features are given higher weights, while unimportant features are compressed or even ignored, ensuring that the data after dimensionality reduction retains the most critical information.

[0066] Furthermore, the quantum noise level allows the model to have different compression levels in different parts of the feature representation, thereby more finely controlling the information loss. The calculation method is expressed as:

[0067] In the formula, is a coefficient that adjusts the sensitivity of quantum noise response. The introduction of randomness through quantum noise can effectively avoid the overfitting of the model to the input data, especially in the case of high feature dimensions. Quantum noise improves the generalization performance of the model by suppressing non-critical features.

[0068] The embodiment of the present invention further trains the autoencoder to ensure that the model can balance the importance of each category of data during the learning process. In the specific implementation, the training method of the autoencoder is as follows: 1) Use the preset training sample set to train the preset autoencoder.

[0069] First, the parameters of the autoencoder algorithm model are initialized. In one embodiment, the parameters of the autoencoder algorithm are initialized using a strategy based on quantum probability distribution, which is expressed as:

[0070]

[0071] In the formula, represents the weight of the autoencoder; Represents the bias of the autoencoder; is the number of features input to the autoencoder; is the first random initialization parameter, is the second random initialization parameter; Represents the imaginary unit, which is used to simulate the complex representation of quantum states.

[0072] During the training process, the data input to the autoencoder can be the data features extracted using the feature extraction model described above. By converting the encoder part into a low-dimensional feature representation, the encoder learns how to map the high-dimensional input data to a low-dimensional latent space, expressed as:

[0073] In the formula, The low-dimensional feature representation after encoding by the autoencoder; is the Sigmoid activation function; is the data vector input to the autoencoder.

[0074] Among them, the embodiment of the present invention uses the reconstruction error and the asymmetric regularization term to calculate the loss, and adjusts the network parameters through the back propagation algorithm to minimize the overall loss. The calculation method of the parameter update amount of the autoencoder is expressed as:

[0075]

[0076] In the formula, and They are the update amounts of the weights and biases of the autoencoder respectively; is the learning rate of the autoencoder; is the regularization coefficient that controls the sparsity of the weights; represents the L2 norm; Represents the L1 norm, which is used to achieve sparse weights; is the reconstructed data feature of the decoder in the autoencoder; is the jth element of the autoencoder weight vector. Preferably, Set to 0.3.

[0077] 2) Based on the information entropy of the training sample set, the data complexity of the training sample set is evaluated.

[0078] The complexity of the input data is evaluated by the information entropy of the input data, and the calculation method is expressed as:

[0079] In the formula, express The probability distribution of the i-th feature in is estimated by the statistical properties of the data distribution.

[0080] 3) Adjust the encoding layer dimension of the autoencoder according to the data complexity and the preset encoding dimension adjustment coefficient.

[0081] The dynamic dimension of the encoding layer is adjusted based on the complexity of the input data, expressed as:

[0082] In the formula, It is the basic coding dimension; is the encoding dimension adjustment coefficient, which is used to control the sensitivity of dimension adjustment; is the complexity of the input data.

[0083] 4) Based on the adjusted encoding layer dimension, the output layer of the autoencoder is adjusted.

[0084] The output layer of the encoder is adjusted using the dynamic dimension, expressed as:

[0085] In the formula, It is the dimension reduction vector after dynamic dimension adjustment, which is input into the decoder of the autoencoder as the feature vector after dimension reduction of the next iteration; is the noise vector of the tth iteration, which is used to enhance the robustness of the autoencoder model during training.

[0086] Furthermore, the adjustment of the noise vector is based on the current reconstruction error of the model, and the calculation method is expressed as:

[0087] In the formula, is the base noise level; is the noise adjustment factor, which is used to control the sensitivity of the error to the noise adjustment; is the model reconstruction error calculated in the tth iteration. Preferably, Set to 0.01.

[0088] In summary, the embodiment of the present invention utilizes a dynamic coding space adjustment strategy to dynamically adjust the dimension and configuration of the coding layer to continuously adapt to the complexity and variability of the input data during the iteration process, thereby achieving more refined feature representation and compression. Repeat the above steps until the preset stop iteration condition is met, indicating that the model training is completed. In one embodiment, the preset stop iteration condition is to reach a preset maximum number of iterations. Preferably, the preset maximum number of iterations is set to 1000 times.

[0089] In summary, the present invention adopts an autoencoder neural network algorithm based on quantum-inspired sparse coding as a feature dimensionality reduction model. On the basis of the traditional autoencoder algorithm, the probability amplitude of quantum bits is used to simulate the distribution of data features, thereby achieving efficient and meaningful feature dimensionality reduction. The feature dimensionality reduction of traditional autoencoder networks is often limited by the linear limitations and information loss of the coding layer. The present invention adopts a quantum-inspired method and uses nonlinear quantum probability distribution to enhance coding efficiency, optimize the compression and reconstruction process of features, and improve the model's ability to process complex data.

[0090] Step S212: using the pre-built classifier model to perform classification prediction on the parameters to be measured, and outputting the device state prediction results of the parameters to be measured.

[0091] In combination with the above embodiments, the existing classifier model still has the problem of a single parameter update method, which is prone to overfitting during training, lacks a dynamic adjustment mechanism, and is difficult to adapt to the learning needs of different stages and changes in the operating status of the equipment, limiting the accuracy of classification and the generalization ability of the model. The present invention adopts a neural decision tree algorithm as a classifier model. The neural decision tree combines the intuitive classification rules of the traditional decision tree and the nonlinear learning ability of deep learning to optimize the feature selection and decision path in the classification process. On the basis of the traditional neural decision tree, the present invention monitors the learning progress and classification effect of the model in real time, and dynamically adjusts the update frequency and freezing state of the model parameters, thereby adapting to the learning needs of different stages, reducing the risk of overfitting, and improving the generalization ability of the model.

[0092] Specifically, the method for constructing the classifier model of the embodiment of the present invention is as follows: 1) Use the preset training sample set to train the preset neural decision tree and calculate the feature influence parameter matrix of the preset training sample set.

[0093] 2) Based on the feature influence parameter matrix and the learning status of the neural decision tree, calculate the feature influence weights of the training sample set.

[0094] First, the structure and parameters of the neural decision tree are initialized. In one embodiment, the initialization method is expressed as:

[0095]

[0096] In the formula, represents the weight of the neural decision tree; Represents the dimension of the data input to the neural decision tree; Represents the number of decision nodes of the neural decision tree; represents the bias of the neural decision tree; represents a random uniform distribution; Indicates compliance with a specific distribution; represents an all-zero vector.

[0097] The data input to the neural decision tree is forward propagated, and each decision node calculates the decision condition according to its weight and activation function to determine the flow of data. During the forward propagation process, the weight of each input feature in the model decision is dynamically adjusted to achieve real-time optimization of the influence of different features, so that the model can more accurately adapt to the characteristics of different device states, thereby improving classification accuracy and model adaptability.

[0098] Specifically, the influence of a feature is evaluated through a feature influence evaluation function. The influence weight of each feature is calculated based on the current input features and the learning effect of historical data. The calculation method is expressed as follows:

[0099] In the formula, is the influence weight of the feature; is the feature vector corresponding to the data input into the neural decision tree; is the Softmax function; is the feature influence parameter matrix to be learned; is the feature influence bias term; is a Sigmoid activation function to ensure the nonlinearity of the output and the expressiveness of the model. The output of the Softmax function ensures that the sum of all feature weights is 1, so that it can be interpreted as a probability distribution. Among them, the feature influence parameter matrix to be learned is constructed based on the feature influence weights of the training process. The feature influence parameter matrix to be learned and the feature influence bias term are all learning parameters. For example, the parameters can be updated by error back propagation.

[0100] 3) Based on the feature influence weights, the classification error of each decision path of the neural decision tree is determined.

[0101] The features input to the neural decision tree will be weighted by the corresponding influence weights to adjust the contribution of each feature to the decision node. The calculation method is expressed as:

[0102] In the formula, represents the product of the corresponding elements, It is the feature input of the weighted neural decision tree. By adjusting the influence of the input features, the model can respond to important features more flexibly, thereby improving decision accuracy.

[0103] Furthermore, during the forward propagation of the neural decision tree, the feature calculation method is expressed as:

[0104] In the formula, Represents the output of each decision node of the neural decision tree; It is an activation function based on sensitivity enhancement.

[0105] In one embodiment, the activation function based on sensitivity enhancement enhances the nonlinearity and adaptability of the model by controlling the curvature of the activation function. for , then the calculation method of the activation function based on sensitivity enhancement is expressed as:

[0106] In the formula, Control the curvature of the activation function to enhance the model’s sensitivity to input changes; is the threshold parameter that determines the activation offset.

[0107] Furthermore, the effect of each decision path is evaluated using the data input to the neural decision tree. In one embodiment, the loss function of the neural decision tree uses a regularization term to prevent overfitting and enhance the model's ability to adapt to new data. The calculation method is expressed as:

[0108] In the formula, is the number of training samples currently input to the neural decision tree, is the true label of the i-th sample, is the label of the i-th sample predicted by the neural decision tree model, is the regularization coefficient, is the weight matrix of the neural decision tree List; is the L2 norm; It is the information entropy calculation function.

[0109] In the embodiment of the present invention, the information entropy of each classification is calculated in each iteration, and the loss function is adjusted accordingly to enhance the model's ability to handle uncertain and unbalanced data, thereby further improving the classification effect. The calculation method of the information entropy calculation function is expressed as follows:

[0110] In the formula, is the predicted probability of the neural decision tree model for the kth class, is the total number of categories of samples input to the neural decision tree.

[0111] 4) Based on the classification error, the decision rules of the neural decision tree are optimized and the matrix update amount of the feature influence parameter matrix is ​​determined.

[0112] The embodiment of the present invention optimizes the parameters of the neural decision tree according to the classification error. The neural decision tree combines the intuitive classification rules of the traditional decision tree and the nonlinear learning ability of deep learning. It can realize the construction of classification rules based on the decision path implementation method of the traditional decision tree and the nonlinear feature transformation method of deep learning. Specifically, the gradient descent method is used to update the weights and biases of the neural decision tree, and the calculation method of the update amount is expressed as:

[0113]

[0114] In the formula, and They are the update amounts of the weights and biases of the neural decision tree respectively; is the learning rate of the neural decision tree; is the loss function of the neural decision tree.

[0115] As the number of iterations increases, the learning rate of the neural decision tree is dynamically adjusted to improve the convergence speed of the model. The adjustment method is expressed as:

[0116] It is The learning rate of the neural decision tree for each iteration, is the initial learning rate of the neural decision tree, The learning rate of the neural decision tree controls the decay rate. Preferably, Set to 0.01, Set to 0.3.

[0117] The update of the feature influence parameter matrix to be learned depends on the error between the model output and the target output. The update is performed using the back propagation algorithm. The calculation method of the update amount is expressed as:

[0118] In the formula, is the update amount of the feature influence parameter matrix to be learned.

[0119] 5) Until the neural decision tree meets the preset training accuracy, the parameters of the neural decision tree are frozen according to the preset parameter ratio to build a classifier model.

[0120] When the model performs well, some parameters are gradually frozen to reduce the complexity of training and prevent overfitting. Specifically, when the training accuracy of the model reaches the preset threshold, the calculation method of the proportion of frozen parameters is expressed as:

[0121] In the formula, is the parameter ratio of the neural decision tree at the tth iteration; Control the freezing rate and gradually increase the proportion of frozen parameters with the number of iterations to stabilize the training process; is the current iteration number.

[0122] Repeat the above steps until the preset stop iteration condition is met, which means that the model training is completed. In one embodiment, the preset stop iteration condition is reaching a preset maximum number of iterations. Preferably, the preset maximum number of iterations is set to 1000 times.

[0123] Step S214: determining the predicted fault of the target device according to the device status prediction result.

[0124] The embodiment of the present invention uses the trained model to process new samples to achieve predictive maintenance of equipment.

[0125] In summary, another artificial intelligence-based equipment predictive maintenance method provided by an embodiment of the present invention adopts a neural network based on dynamic feature capture for feature extraction, identifies key features by real-time analysis of the statistical characteristics of the input data, and prioritizes features that have a significant impact on the prediction results to improve the model's sensitivity to fault conditions. In addition, through an adaptive abnormal feature suppression strategy, the impact of abnormal data is dynamically suppressed to protect the model's stability, and the model's ability to cope with high-uncertainty data is further enhanced through uncertainty metric evaluation.

[0126] In addition, a quantum-inspired sparse coding autoencoder algorithm is used for feature dimensionality reduction, and efficient compression of data features is achieved by simulating the probability distribution of quantum bits, which solves the problems of large information loss and low coding efficiency in traditional dimensionality reduction methods. In addition, through the dynamic coding space adjustment strategy, the dimension and noise response sensitivity of the coding layer are dynamically adjusted according to the complexity of the input data, which effectively enhances the adaptability of the model to data of different complexities.

[0127] A neural decision tree model is also used as a classifier. By combining the regularity of the decision tree and the nonlinear learning ability of deep learning, the fault classification process is optimized, and the rationality of the feature selection and decision path of the classification process is improved. In addition, through the dynamic adjustment and freezing mechanism of parameters, the model learning progress is monitored in real time, and the parameter update frequency and freezing status are dynamically adjusted, which enhances the generalization ability of the model and reduces overfitting.

[0128] Furthermore, the embodiment of the present invention uses a preset training sample set to train the above model. The embodiment of the present invention also provides another equipment predictive maintenance method based on artificial intelligence, mainly illustrating a method for constructing a training sample set. Figure 3 A flowchart of a method for constructing a training sample set provided by an embodiment of the present invention is shown. Figure 3 , the method comprises the following steps: Step S10, obtaining operation status monitoring samples of preset equipment, marking the operation status monitoring samples according to the fault status corresponding to the preset equipment, and constructing an initial sample set.

[0129] The operation status monitoring samples can refer to the above-mentioned embodiments. Furthermore, the embodiments of the present invention also mark the collected data. The marking method of the present invention is manual marking. In one embodiment, the marked categories include "early faults of bearing wear", "early faults of motor overheating", "early faults of insufficient lubrication" and other early states of various faults.

[0130] Step S11, using a preset generative adversarial network generator to perform sample expansion on the initial sample set to generate initial expanded samples.

[0131] It is understandable that in the task of the present invention, the collection, acquisition, labeling and preprocessing of training data are time-consuming and labor-intensive, and insufficient training samples can easily lead to poor model generalization ability and affect the accuracy of the model. Existing data expansion methods mainly rely on a small amount of sampling technology, which cannot effectively expand the equipment operation data, and the quality and diversity of the generated data are not high, resulting in insufficient model generalization ability and prediction accuracy. The present invention adopts a generative adversarial network to generate samples, and then realizes data expansion. On the basis of the traditional generative adversarial network, the mutation mechanism of the evolutionary algorithm is used to optimize the learning process of the generator and discriminator in the generative adversarial network, thereby improving the quality and diversity of the generated data. In addition, the game learning strategy is used to further refine the confrontation process of the generator and the discriminator, so that each step of the update more accurately reflects the complexity of the data set and its inherent distribution characteristics.

[0132] Among them, the parameters of the generator and the discriminator are randomly initialized. In one embodiment, the initialization method is expressed as:

[0133] In the formula, represents the parameters of the generator, represents the parameters of the discriminator, represents the initialization variance; Indicates compliance with a specific distribution; represents the normal distribution variance of the GAN parameter initialization. Preferably, Set to 0.01.

[0134] Step S12, using the discriminator of the generative adversarial network to discriminate the initial sample set and the initial expanded sample respectively.

[0135] In the standard generative adversarial network training framework, the generator tries to generate new data that is sufficient to deceive the discriminator, and the discriminator strives to distinguish between real data and generated data. The loss function of the traditional generative adversarial network is expressed as:

[0136] In the formula, Represents real data, Represents a random noise distribution Input; represents random noise distribution; express expectations; It is the second division of the real data; represents the discriminator; Represents a generator; represents the discriminator function; represents a generator function; is the discriminator's output on the real data. is the fake data output by the generator.

[0137] Based on the traditional generative adversarial network, the present invention uses game learning strategy for optimization, so that the two networks continuously improve their strategies and counter-strategies during the training process. Specifically, the discriminator's discriminant process fully considers multiple discriminant strategies. The discriminator has different discriminant strategies for real data and data generated by the generator. Based on this, the loss function is calculated. During the training process, the loss function constraint can improve its game confrontation ability and realize the improvement of the generator's ability. The definition of the discriminator's discriminant process for real data is expressed as:

[0138] In the formula, is the scoring function of the discriminator for the real data, is the parameter of the scoring function of the discriminator for the real data.

[0139] Furthermore, the implementation of the scoring function of the discriminator for real data is expressed as:

[0140] In the formula, The number of real data samples input to the discriminator for the current batch; is the hyperparameter of the discriminator’s scoring function for real data; is the hyperbolic tangent function; Train the weight parameters of the discriminator's scoring function for real data; for Transpose of parameters; is the bias training parameter of the discriminator's scoring function for real data.

[0141] And, the discriminator's discriminative process for generated data is defined as:

[0142] In the formula, is the scoring function of the discriminator on the generated data, are the parameters of the discriminator’s scoring function for the generated data.

[0143] Furthermore, the calculation method of the scoring function of the discriminator for the generated data is expressed as:

[0144] In the formula, is the hyperparameter of the scoring function of the discriminator on the generated data; Train the weights of the scoring function for the discriminator on the generated data; for Transpose of parameters; The bias training parameters of the scoring function of the discriminator for the generated data; The number of data samples generated by the generator for the current batch that are fed into the discriminator.

[0145] In summary, the discriminator of the embodiment of the present invention uses different discriminant functions for real data and generated data, so that the discriminator can more accurately model data of different properties. The distribution of real data and generated data may be essentially different, so designing adaptive discriminant functions can better capture these differences and improve the ability to distinguish. At the same time, the authenticity of the generated data is examined from multiple perspectives at the same time, thereby improving the robustness of the entire adversarial process. The authenticity of the generated data can be examined from multiple perspectives at the same time, thereby improving the robustness of the entire adversarial process.

[0146] Step S13, updating the generator parameters of the generator based on the discrimination results of the initial sample set and the discrimination results of the initial expanded samples to update the initial expanded samples.

[0147] The embodiment of the present invention determines the loss function of the generator based on the discrimination results of the initial sample set and the discrimination results of the initial expanded samples; determines the fitness function of the generative adversarial network based on the dynamic changes corresponding to the generative adversarial network; and updates the generator parameters of the generator based on the loss function and the fitness function.

[0148] The parameters of the generator are adjusted using an evolutionary strategy to improve its generation quality. The evolutionary algorithm evaluates and selects the best individuals, expressed as:

[0149] In the formula, is the generator parameter for the tth iteration, is the generator parameter for the t+1th iteration; is the learning rate of the generator at the tth iteration, is the loss function of the generative adversarial network; represents the gradient with respect to the generator parameters; is the fitness function of the generative adversarial network. In the process of adversarial training, the traditional generative adversarial network usually uses a fixed learning rate for training, which may cause the learning rate to be too large in the early stage of training and unable to be finely adjusted, or the learning rate to be too small in the later stage of training and slow convergence. In order to solve this problem, the present invention adopts a dynamic learning rate adjustment mechanism based on loss feedback to adjust the dynamic learning rate, which is expressed as:

[0150] In the formula, is the base learning rate, Represents the change in the discriminator loss between two consecutive iterations; is the learning rate adjustment factor of the generative adversarial network, which is used to control the speed at which the learning rate decreases. Preferably, Set to 0.01, Set to 0.3.

[0151] The fitness function of the generative adversarial network aims to encourage the generator to not only deceive the discriminator but also increase the diversity of generated samples. The calculation method is expressed as:

[0152] In the formula, is a parameter that controls the effect of fitness; is the mutation rate of the tth iteration; Preferably, Set to 0.3. Automatically adjust the mutation rate based on the training results of the previous iteration to adapt to dynamic changes in the training process, optimize the exploration and utilization balance of the algorithm, and prevent premature convergence to the local optimal solution. The specific adjustment strategy is as follows:

[0153] In the formula, is the mutation rate at the t+1th iteration, is the mutation rate adjustment coefficient; It is an evaluation indicator, which can be set as the diversity indicator of generated data. is the target value of the preset evaluation index. Preferably, Set to 0.2. Based on the fitness function, the best performing individuals are selected for reproduction to generate the new generation of network parameters, while the individuals with poor performance are eliminated, which is expressed as:

[0154] In the formula, The parameters of the generator corresponding to the individuals with good performance screened based on the fitness function; To select the function, the best performing parameter set is selected from the current population for update according to the fitness function.

[0155] Step S14, until the initial expanded sample meets the preset expansion requirement, a training sample set is constructed based on the initial sample set and the initial expanded sample.

[0156] Repeat the above steps until the preset stop iteration condition is met, which means that the model training is completed. In one embodiment, the preset stop iteration condition is to reach the preset maximum number of iterations. Preferably, the preset maximum number of iterations is set to 1000 times. After the data expansion model training is completed, the trained data expansion model is used to increase the number of samples. In one embodiment, assuming that the original collected samples are 800, and the data expansion model is expanded to generate 200 samples, the expanded data set contains 1000 samples. It should be noted that the training sample set can be used to train the feature extraction model, and further, the autoencoder can be trained after the features of the training sample set are extracted. The trained autoencoder can also be used to compress the data of its training sample set to further train the classifier model.

[0157] In summary, another artificial intelligence-based equipment predictive maintenance method provided by an embodiment of the present invention adopts a generative adversarial network in conjunction with an evolutionary algorithm for data expansion, and dynamically adjusts the learning process of the generator and the discriminator through an evolutionary strategy, thereby solving the problem of poor model generalization ability and insufficient precision caused by insufficient training data. In addition, the generator and discriminator of the generative adversarial network are optimized through a game learning strategy, so that the generated data is more in line with the complex distribution characteristics of actual equipment data, thereby improving the quality and diversity of generated samples.

[0158] On the basis of the above embodiments, the present invention also provides an artificial intelligence-based equipment predictive maintenance device. Figure 4 A schematic diagram of the structure of an artificial intelligence-based equipment predictive maintenance device provided by an embodiment of the present invention is shown, referring to Figure 4 The device includes: a data acquisition module 100, which is used to acquire the operating status monitoring data of the target device; a feature extraction module 200, which is used to analyze the statistical characteristics of the status monitoring data and identify the key features in the operating status monitoring data; a data processing module 300, which is used to simulate the quantum state of the key features, and based on the quantum state amplitude of the quantum state, the key features are compressed to generate the parameters to be measured; an execution module 400, which is used to classify and predict the parameters to be measured using a pre-built classifier model, and output the device status prediction results of the parameters to be measured; an output module 500, which is used to determine the predicted fault of the target device according to the device status prediction results.

[0159] An artificial intelligence-based equipment predictive maintenance device provided in an embodiment of the present invention has the same technical features as an artificial intelligence-based equipment predictive maintenance method provided in the above embodiment, so it can also solve the same technical problems and achieve the same technical effects.

[0160] The output module 500 is also used to determine the predicted fault of the target device according to the device state prediction result, including: obtaining the device predicted state indicated by the device state prediction result; and determining the device predicted state as the predicted fault of the target device.

[0161] The feature extraction module 200 is also used to calculate the information gain of each feature of the state monitoring data; based on the information gain, determine the statistical characteristics of the state monitoring data; and determine the features whose information gain exceeds the preset gain threshold as key features in the operating state monitoring data. The feature extraction module 200 is also used to train the preset neural network using the preset training sample set, determine the historical mean and standard deviation of the training sample set; determine the abnormal evaluation results of the training sample set based on the historical mean and standard deviation; and adjust the feature values ​​of the features of the training sample set based on the results of the abnormal evaluation.

[0162] The above-mentioned data processing module 300 is also used to convert the key features into quantum bits, calculate the quantum noise level of the key features; determine the probability amplitude corresponding to the quantum bits; and compress the key features based on the probability amplitude and the quantum noise level. The above-mentioned data processing module 300 is also used to convert the key features into low-dimensional feature representations through a preset autoencoder and perform data compression on the key features; wherein the training method of the autoencoder includes: training the preset autoencoder using a preset training sample set; evaluating the data complexity of the training sample set based on the information entropy of the training sample set; adjusting the encoding layer dimension of the autoencoder according to the data complexity and the preset encoding dimension adjustment coefficient; and adjusting the output layer of the autoencoder based on the adjusted encoding layer dimension.

[0163] The above-mentioned execution module 400 is also used to train a preset neural decision tree using a preset training sample set, and calculate the feature influence parameter matrix of the preset training sample set; based on the feature influence parameter matrix and the learning status of the neural decision tree, calculate the feature influence weight of the training sample set; based on the feature influence weight, determine the classification error of each decision path of the neural decision tree; based on the classification error, optimize the decision rules of the neural decision tree, and determine the matrix update amount of the feature influence parameter matrix; until the neural decision tree meets the preset training accuracy, freeze the parameters of the neural decision tree according to the preset parameter ratio to construct a classifier model. The execution module 400 is also used to obtain the operation status monitoring samples of the preset equipment, annotate the operation status monitoring samples according to the fault status corresponding to the preset equipment, and construct an initial sample set; use the preset generator of the generative adversarial network to expand the samples of the initial sample set to generate initial expanded samples; use the discriminator of the generative adversarial network to discriminate the initial sample set and the initial expanded samples respectively; update the generator parameters of the generator based on the discrimination results of the initial sample set and the discrimination results of the initial expanded samples to update the initial expanded samples; until the initial expanded samples meet the preset expansion requirements, construct a training sample set based on the initial sample set and the initial expanded samples. The execution module 400 is also used to determine the loss function of the generator based on the discrimination results of the initial sample set and the discrimination results of the initial expanded samples; determine the fitness function of the generative adversarial network based on the dynamic changes corresponding to the generative adversarial network; and update the generator parameters of the generator based on the loss function and the fitness function.

[0164] An embodiment of the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above Figures 1 to 3 The embodiment of the present invention also provides a computer-readable storage medium on which a computer program is stored. The computer program is executed by a processor to perform the above Figures 1 to 3 The embodiment of the present invention also provides a structural diagram of an electronic device, such as Figure 5 FIG. 5 is a schematic diagram of the structure of the electronic device, wherein the electronic device includes a processor 51 and a memory 50, the memory 50 stores computer executable instructions that can be executed by the processor 51, and the processor 51 executes the computer executable instructions to implement the above Figures 1 to 3 The method shown. Figure 5In the illustrated embodiment, the electronic device further includes a bus 52 and a communication interface 53, wherein the processor 51, the communication interface 53 and the memory 50 are connected via the bus 52. The memory 50 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 53 (which may be wired or wireless), and the Internet, a wide area network, a local area network, a metropolitan area network, etc. may be used. The bus 52 may be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. It may also be an AMBA (Advanced Microcontroller Bus Architecture) bus, where AMBA defines three types of buses, including an APB (Advanced Peripheral Bus) bus, an AHB (Advanced High-performance Bus) bus, and an AXI (Advanced Xtensible Interface) bus. The bus 52 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5Only one bidirectional arrow is used in the diagram, but it does not mean that there is only one bus or one type of bus. The processor 51 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit in the processor 51 or the instructions in the form of software. The above processor 51 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor, or a combination of hardware and software modules in the decoding processor. The software module can be located in a random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, register or other mature storage media in the art. The storage medium is located in the memory, and the processor 51 reads the information in the memory and completes the above-mentioned Figures 1 to 3 Any of the methods shown.

[0165] A computer program product of a predictive maintenance method and device for equipment based on artificial intelligence provided by an embodiment of the present invention includes a computer-readable storage medium storing a program code, and the instructions included in the program code can be used to execute the method described in the previous method embodiment. The specific implementation can refer to the method embodiment, which will not be repeated here. A person skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the system described above can refer to the corresponding process in the above method embodiment, which will not be repeated here. In addition, in the description of the embodiment of the present invention, unless otherwise clearly specified and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected, or it can be indirectly connected through an intermediate medium, or it can be the internal connection of two components. For those skilled in the art, the specific meaning of the above terms in the present invention can be understood in specific circumstances. If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a disk or an optical disk, and other media that can store program codes. In the description of the present invention, it should be noted that the orientation or position relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inside", "outside", etc. is based on the orientation or position relationship shown in the accompanying drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", and "third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.Finally, it should be noted that the above embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above embodiments, those skilled in the art should understand that any person skilled in the art can still modify the technical solutions recorded in the above embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. An artificial intelligence-based equipment predictive maintenance method, characterized in that: The method comprises: Obtain the operating status monitoring data of the target device; Analyzing the statistical characteristics of the state monitoring data, and identifying key features in the operating state monitoring data based on the statistical characteristics; wherein the key features are identified using a pre-built feature extraction model, and a training sample set for training the feature extraction model is adjusted based on the result of the abnormality assessment; Simulating the quantum state of the key feature, and based on the quantum state amplitude of the quantum state, performing data compression on the key feature to generate a parameter to be measured; Using a pre-built classifier model to classify and predict the parameters to be measured, and output the device state prediction results of the parameters to be measured; The predicted fault of the target device is determined according to the device status prediction result.

2. The method according to claim 1, characterized in that The step of determining the predicted fault of the target device according to the device state prediction result includes: Acquire the predicted state of the device indicated by the device state prediction result; The predicted device state is determined as a predicted failure of the target device.

3. The method according to claim 1, characterized in that The step of analyzing the statistical characteristics of the condition monitoring data and identifying key features in the operating condition monitoring data based on the statistical characteristics comprises: Calculating the information gain of each feature of the condition monitoring data; Based on the information gain, determining the statistical characteristics of the condition monitoring data; The feature whose information gain exceeds a preset gain threshold is determined as a key feature in the operating status monitoring data.

4. The method according to claim 3, characterized in that The step of adjusting the training sample set for training the feature extraction model based on the result of the abnormality assessment comprises: Using a preset training sample set to train a preset neural network, and determining a historical mean and a standard deviation of the training sample set; Determining an abnormality assessment result of the training sample set based on the historical mean and standard deviation; Based on the result of the abnormality assessment, feature values ​​of the features of the training sample set are adjusted.

5. The method according to claim 1, characterized in that The step of simulating the quantum state of the key feature, compressing the data of the key feature based on the quantum state amplitude of the quantum state, and generating the parameter to be measured includes: Converting the key feature into a quantum bit and calculating the quantum noise level of the key feature; Determining a probability amplitude corresponding to the quantum bit; The key features are compressed based on the probability amplitude and the quantum noise level.

6. The method according to claim 5, characterized in that The step of compressing the key features based on the probability amplitude and the quantum noise level comprises: The key features are converted into low-dimensional feature representations by a preset autoencoder, and data compression is performed on the key features; The training method of the autoencoder includes: Using a preset training sample set to train a preset autoencoder; Based on the information entropy of the training sample set, evaluating the data complexity of the training sample set; Adjusting the encoding layer dimension of the autoencoder according to the data complexity and a preset encoding dimension adjustment coefficient; Based on the adjusted encoding layer dimension, an output layer of the autoencoder is adjusted.

7. The method according to claim 1, characterized in that The method for constructing the classifier model comprises: Using a preset training sample set to train a preset neural decision tree, and calculating a feature influence parameter matrix of the preset training sample set; Based on the feature influence parameter matrix and the learning status of the neural decision tree, calculating the feature influence weight of the training sample set; Determining the classification error of each decision path of the neural decision tree based on the feature influence weight; Based on the classification error, optimizing the decision rule of the neural decision tree, and determining the matrix update amount of the feature influence parameter matrix; Until the neural decision tree meets the preset training accuracy, the parameters of the neural decision tree are frozen according to the preset parameter ratio to construct a classifier model.

8. The method according to claim 1, characterized in that The method further comprises: Acquire operation status monitoring samples of preset devices, annotate the operation status monitoring samples according to the fault status corresponding to the preset devices, and construct an initial sample set; Using a preset generator of a generative adversarial network to perform sample expansion on the initial sample set to generate initial expanded samples; Using the discriminator of the generative adversarial network to discriminate the initial sample set and the initial expanded samples respectively; Based on the discrimination result of the initial sample set and the discrimination result of the initial expanded sample, the generator parameter of the generator is updated to update the initial expanded sample; Until the initial expanded samples meet the preset expansion requirements, a training sample set is constructed based on the initial sample set and the initial expanded samples.

9. The method according to claim 8, characterized in that The step of updating the generator parameters of the generator based on the discrimination result of the initial sample set and the discrimination result of the initial expanded sample comprises: Determining the loss function of the generator based on the discrimination result of the initial sample set and the discrimination result of the initial expanded sample; Determining a fitness function of the generative adversarial network based on dynamic changes corresponding to the generative adversarial network; Based on the loss function and the fitness function, the generator parameters of the generator are updated.

10. An artificial intelligence-based equipment predictive maintenance device, characterized in that: The device comprises: A data acquisition module is used to acquire the operating status monitoring data of the target device; A feature extraction module, used to analyze the statistical characteristics of the state monitoring data and identify key features in the operating state monitoring data; wherein the key features are identified using a pre-built feature extraction model, and a training sample set for training the feature extraction model is adjusted based on the result of anomaly assessment; A data processing module, used for simulating the quantum state of the key feature, and performing data compression on the key feature based on the quantum state amplitude of the quantum state to generate a parameter to be measured; An execution module, used to classify and predict the parameters to be measured using a pre-built classifier model, and output a device state prediction result of the parameters to be measured; The output module is used to determine the predicted fault of the target device according to the device status prediction result.

Images