Hydraulic plunger pump fault online diagnosis method and system based on dynamic confidence distribution

Abstract

The invention discloses a hydraulic plunger pump fault online diagnosis method and system based on dynamic confidence distribution, and relates to the technical field of intelligent detection, and the method comprises the steps: carrying out the feature extraction of the preprocessed real-time operation state data, combining the extracted features, and obtaining a total feature vector; acquiring a real-time signal of the hydraulic plunger pump, and constructing a dynamic confidence interval based on the real-time signal; if the signal value of the current real-time signal exceeds the dynamic confidence interval, calculating the confidence probability of the current real-time signal, and obtaining the joint confidence probability of all the signals based on the confidence probability; inputting the total feature vector into a pre-trained fault diagnosis model for prediction to obtain a prediction probability of a fault category; and obtaining a posterior probability of the fault category based on the prediction probability and the joint confidence probability, and determining a fault diagnosis result based on the posterior probability. Common faults such as abrasion, blockage and internal leakage of the hydraulic plunger pump can be accurately recognized, and the real-time performance and reliability of diagnosis can be improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent detection technology, and in particular to a hydraulic plunger pump fault online diagnosis method and system based on dynamic confidence distribution. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

[0003] With the rapid advancement of technology, the intelligent control system of hydraulic pumps has become more comprehensive and automated. Due to the complex and changeable working environment, hydraulic pumps are easily affected by various factors during operation, resulting in failures. These failures will not only affect the normal operation of the hydraulic pump, but may also cause serious damage to the equipment and even cause safety accidents. Therefore, timely and accurate fault diagnosis of hydraulic pumps is of great significance. In the field of hydraulic pump fault prediction, the intelligent system integrates advanced sensing technology and data analysis methods, which can realize remote monitoring and rapid fault warning.

[0004] However, existing hydraulic pump fault prediction and diagnosis methods face huge challenges. In the process of fault model training, problems such as unstable training, model overfitting, low diagnostic accuracy under complex working conditions, and insufficient reliability and credibility of diagnostic results often occur. In addition, traditional optimization algorithms are prone to fall into local optimal solutions and it is difficult to find the global optimal solution. These challenges reduce the reliability and accuracy of the fault diagnosis process. Summary of the invention

[0005] In order to overcome the shortcomings of the above-mentioned prior art, the present invention provides an online diagnosis method and system for hydraulic plunger pump faults based on dynamic confidence distribution, which integrates dynamic confidence distribution and machine learning models to obtain fault diagnosis results and their credibility, and can accurately identify common faults such as hydraulic plunger pump wear, blockage, and internal leakage, thereby improving the real-time and reliability of diagnosis.

[0006] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0007] In a first aspect, the present invention provides a method for online fault diagnosis of a hydraulic plunger pump based on dynamic confidence distribution, comprising:

[0008] Acquire the real-time operation status data of the hydraulic piston pump and perform preprocessing to obtain the preprocessed real-time operation status data;

[0009] Perform feature extraction on the preprocessed real-time operation status data, and combine the extracted features to obtain a total feature vector;

[0010] Acquire a real-time signal of a hydraulic piston pump, and construct a dynamic confidence interval based on the real-time signal; if the signal value of the current real-time signal exceeds the dynamic confidence interval, calculate the confidence probability of the current real-time signal, and obtain the joint confidence probability of all signals based on the confidence probability;

[0011] Inputting the total feature vector into a pre-trained fault diagnosis model for prediction to obtain a predicted probability of the fault category;

[0012] A posterior probability of the fault category is obtained based on the predicted probability and the joint confidence probability, and a fault diagnosis result is determined based on the posterior probability.

[0013] As a further technical feature, the operating status data includes system pressure, output flow, pump body temperature, vibration signals in three directions of XYZ and control current.

[0014] Further technical features are used to obtain the total feature vector as follows: feature extraction is performed on the preprocessed operating status data, including the time domain features and frequency domain features of the signal. The time domain features include mean, standard deviation, kurtosis, skewness and root mean square value, and the frequency domain features include main frequency, spectral energy and spectral entropy. The time domain features and frequency domain features are combined to form a complete total feature vector.

[0015] Further technical features, the joint confidence probability of all signals is obtained as follows:

[0016] Calculate real-time statistical parameters, including real-time mean and standard deviation;

[0017] constructing a dynamic confidence interval for each signal based on the real-time mean and standard deviation;

[0018] Calculating a confidence probability for each signal based on the dynamic confidence interval;

[0019] The confidence summaries of all signals are weighted and calculated to obtain the joint confidence probability.

[0020] Further technical features, constructing dynamic confidence intervals are expressed as:

[0021] CI(t)=[μ(t)-z·σ(t),μ(t)+z·σ(t)]

[0022] Where CI(t) represents the dynamic confidence interval of the t time window, μ(t) represents the real-time mean of the t time window, z represents the confidence coefficient, and σ(t) represents the standard deviation of the t time window.

[0023] A further technical feature is that if the signal value of the current real-time signal exceeds the dynamic confidence interval, the confidence probability of the current real-time signal is calculated, and the formula is expressed as:

[0024]

[0025] Among them, P(f i |D) represents the confidence probability, f i (t) represents the signal, μ(t) represents the real-time mean of the time window t, and σ(t) represents the standard deviation of the time window t.

[0026] According to a further technical feature, the fault diagnosis model includes a convolutional neural network module, a long short-term memory network module and an output layer in sequence, and the output layer uses a Softmax function to calculate the predicted probability of the fault category.

[0027] In a second aspect, the present invention provides an online fault diagnosis system for a hydraulic piston pump based on dynamic confidence distribution, comprising:

[0028] A data acquisition module is configured to: acquire real-time operating status data of the hydraulic piston pump and perform preprocessing to obtain preprocessed real-time operating status data;

[0029] A feature extraction module is configured to: extract features from the preprocessed real-time running status data, and combine the extracted features to obtain a total feature vector;

[0030] A dynamic confidence distribution module is configured to: obtain a real-time signal of a hydraulic piston pump, and construct a dynamic confidence interval based on the real-time signal; if the signal value of the current real-time signal exceeds the dynamic confidence interval, calculate the confidence probability of the current real-time signal, and obtain the joint confidence probability of all signals based on the confidence probability;

[0031] A model prediction module is configured to: input the total feature vector into a pre-trained fault diagnosis model for prediction to obtain a predicted probability of a fault category;

[0032] The fusion diagnosis decision module is configured to obtain a posterior probability of the fault category based on the predicted probability and the joint confidence probability, and determine the fault diagnosis result based on the posterior probability.

[0033] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the online fault diagnosis method for hydraulic piston pumps based on dynamic confidence distribution as described in the first aspect.

[0034] In a fourth aspect, the present invention provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the online diagnosis method for hydraulic piston pump faults based on dynamic confidence distribution as described in the first aspect are implemented.

[0035] One or more of the above technical solutions have the following beneficial effects:

[0036] The present invention adopts the fault diagnosis model CNN-LSTM to solve the key defects in the existing hydraulic pump fault prediction and diagnosis methods. Specifically, CNN can improve the stability of the training process by automatically extracting the spatial features of the signal; LSTM can effectively capture the long-term and short-term dependencies in the time series data, enhance the model's ability to understand dynamic signal changes, and further improve the stability of the overall training. In addition, the dynamic confidence distribution is combined with the CNN-LSTM model, and the confidence interval and weight distribution of the signal are adjusted in real time, which can further enhance the model's ability to identify faults under complex working conditions.

[0037] The present invention uses real-time mean and real-time standard deviation to construct dynamic confidence distribution, calculate the confidence probability of each signal, and combine the multi-signal joint confidence distribution to determine whether the system is abnormal, evaluate the credibility of the fault, reduce false alarms and missed alarms, and improve the accuracy of fault diagnosis. Dynamic confidence distribution can improve data quality and relevance, enhance model adaptability and robustness, and reduce overfitting.

[0038] The present invention combines dynamic confidence distribution with machine learning model fault diagnosis, comprehensively considers multiple factors, makes the final diagnosis result more reliable, and provides real-time, adaptive and robust fault detection for the hydraulic plunger pump.

[0039] The present invention constructs a dynamic confidence interval according to a real-time signal, can be adjusted as the operating state of the hydraulic plunger pump changes, can adapt to different working environments and working conditions, and has wide applicability.

[0040] The present invention can timely discover and handle potential problems by online diagnosis of hydraulic plunger pump faults, avoid further deterioration of the faults leading to equipment damage or shutdown, and help achieve preventive maintenance and extend the service life of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0042] Figure 1 is a flow chart of a method for online fault diagnosis according to an embodiment of the present invention;

[0043] Figure 2 is a flow chart of the fault diagnosis model and dynamic confidence interval feature vector acquisition according to an embodiment of the present invention;

[0044] Figure 3 It is a flow chart of calculating the joint confidence probability according to an embodiment of the present invention. DETAILED DESCRIPTION

[0045] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0046] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0047] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0048] Embodiment 1

[0049] like Figure 1 As shown, this embodiment discloses a hydraulic plunger pump fault online diagnosis method based on dynamic confidence distribution, which includes the following steps:

[0050] S1: acquiring real-time operating status data of the hydraulic piston pump and performing preprocessing to obtain preprocessed real-time operating status data;

[0051] In this embodiment, sensors are installed at the outlet, pump body, and control circuit of the axial piston pump to collect status data of the piston pump during operation, including system pressure, output flow, pump body temperature, vibration signals in three directions of XYZ, and control current.

[0052] Fault simulation is performed based on the collected operating status data, such as common faults such as wear, blockage, and internal leakage of hydraulic piston pumps. The sensor sampling frequency is 1kHz to ensure the capture of fast-fluctuating signal characteristics. Each sampling lasts for 1 minute and covers different operating states. The collected signals are stored as multi-channel time series data in a format that includes timestamps, signal values, and status labels.

[0053] The state data of the axial piston pump during operation, i.e., the operating state data, is preprocessed, including data cleaning, data segmentation, and data normalization. Data cleaning is to perform preliminary cleaning on the collected multi-channel time series data, eliminate interference signals and noise, and use linear interpolation to fill in missing data points. Data segmentation is to divide the signal according to a fixed time window to generate multiple subsequences of fixed length.

[0054] The historical operating status data of the hydraulic piston pump is obtained and preprocessed, and the preprocessed historical operating status data is divided into a training set and a test set according to the ratio of 80% and 20%.

[0055] S2: extracting features from the preprocessed real-time running status data, and combining the extracted features to obtain a total feature vector;

[0056] In this embodiment, the pre-processed running status data is subjected to feature extraction, including the time domain features and frequency domain features of the signal. The time domain features include mean μ, standard deviation σ, kurtosis K, skewness S and root mean square value RMS, and the frequency domain features include the main frequency f dom , spectrum energy E and spectrum entropy H. For the data in each time window, the time domain features and frequency domain features are combined to form a complete total feature vector, which is:

[0057] x p =[μ,σ,K,S,RMS,f dom ,E,H].

[0058] The total feature vector extracted from the preprocessed historical operating status data is used to train the fault diagnosis model CNN-LSTM.

[0059] S3: acquiring a real-time signal of the hydraulic piston pump, and constructing a dynamic confidence interval based on the real-time signal; if the signal value of the current real-time signal exceeds the dynamic confidence interval, calculating the confidence probability of the current real-time signal, and obtaining the joint confidence probability of all signals based on the confidence probability;

[0060] In this embodiment, the pre-processed real-time operation status data is constructed as a multi-channel feature vector, and the multi-channel feature vector is a real-time signal. The signal mean and standard deviation are calculated based on the real-time signal, and a dynamic confidence interval is constructed. If the signal value of the current real-time signal exceeds the dynamic confidence interval, its confidence probability is calculated. Combined with the dynamic confidence probability of multiple signals (system pressure, output flow, pump body temperature, vibration signals in three directions of XYZ, and control current), the joint confidence distribution is calculated to quantify the overall operation status of the system.

[0061] S301: Calculate real-time statistical parameters: The real-time statistical signal parameters are the real-time mean and standard deviation of each signal, which are specifically expressed as follows:

[0062]

[0063] Where t represents the current time, which indicates the time point of signal analysis; N represents the size of the sliding window; k represents the time index within the sliding window, which indicates the time point from t-N+1 to t; f i(k) represents the value of the signal at time point k, which is a specific sampling value in the signal data.

[0064] S302: constructing a dynamic confidence interval based on the real-time mean and standard deviation: constructing a dynamic confidence interval for each signal, which is specifically expressed as follows:

[0065] CI(t)=[μ(t)-z·σ(t),μ(t)+z·σ(t)]

[0066] Where CI(t) represents the dynamic confidence interval of the t time window, μ(t) represents the real-time mean of the t time window, z represents the confidence coefficient, and σ(t) represents the standard deviation of the t time window.

[0067] S303: Calculate confidence probability based on dynamic confidence interval: Each signal f i (t) The confidence probability P(f) in the dynamic confidence interval i |D) characterizes the "normality" of the signal and is calculated as follows:

[0068] If the signal value f i (t) falls within the dynamic confidence interval:

[0069] P(f i D)=1;

[0070] If the signal value f i (t) exceeds the dynamic confidence interval:

[0071]

[0072] Among them, P(f i |D) represents the confidence probability, f i (t) represents the signal, μ(t) represents the real-time mean of the time window t, and σ(t) represents the standard deviation of the time window t.

[0073] The more the signal deviates from the mean, the higher the confidence probability P(f i |D)The lower.

[0074] S304: Joint confidence distribution of multiple signals: weighted calculation is performed on the confidence probabilities of all signals to obtain the joint confidence probability of the system:

[0075]

[0076] Where F represents the fault category, D represents the given observation data, n represents the total number of signals, i represents the i-th signal, j represents the index used to traverse the signal in the weight calculation, and w i Represents the weight of the i-th signal, which is dynamically adjusted according to signal fluctuations and confidence.

[0077] S4: inputting the total feature vector into a pre-trained fault diagnosis model for prediction to obtain a predicted probability of the fault category;

[0078] In this embodiment, the fault diagnosis model includes a convolutional neural network (CNN) module, a long short-term memory (LSTM) network module, and an output layer in sequence. The CNN module extracts the local spatiotemporal features of the total feature vector through one-dimensional convolution and reduces the dimension through a pooling layer. The local spatiotemporal features are input into the LSTM module for extraction. The LSTM module captures the long-term and short-term dependencies in the local spatiotemporal features and outputs the hidden state representing the comprehensive spatiotemporal features, i.e., the multi-scale features. The multi-scale features are input into the output layer, which uses Softmax to calculate the predicted probability of each fault category. That is to say, the fault diagnosis model adopts the CNN-LSTM model, which combines the respective advantages of the convolutional neural network CNN and the long short-term memory network LSTM network. It can also reduce the problem of gradient vanishing or gradient exploding when LSTM processes long sequences through convolution operations; CNN-LSTM can capture the multi-scale features of data at different levels and improve the accuracy of fault diagnosis.

[0079] The preprocessed historical operating status data is divided into a training set and a test set in the ratio of 80% and 20%. The training set is used to train the fault diagnosis model, and the test set is used to evaluate the performance of the fault diagnosis model.

[0080] Furthermore, the fault categories include common faults such as hydraulic piston pump wear, blockage, and internal leakage. The data in the training set is labeled to train the fault diagnosis model. The specific operation of data labeling is: for each sampling time period, check whether the equipment has a fault (or artificially simulate a fault), confirm the fault type in the time period, and assign a fault category label to each sample.

[0081] Here is a simple example of fault diagnosis model training:

[0082] Suppose there is a vibration signal dataset of a hydraulic pump, which contains common faults such as wear, blockage, and internal leakage. The following is a simple training process:

[0083] Collect data: Collect vibration signals of hydraulic pumps under different fault conditions.

[0084] Feature extraction: extract time domain features (mean, standard deviation, skewness, etc.) and frequency domain features (main frequency, spectrum energy, etc.) from each signal segment, and combine the features to obtain the total feature vector.

[0085] Training model: The extracted features and corresponding labels are input into the CNN-LSTM model for training. The model is optimized using the cross entropy loss function to learn the relationship between signal features and fault categories.

[0086] Prediction and output probability: During the test phase, a new hydraulic pump signal is input, and the model outputs the probability of each fault category. Assume that the model outputs: the probability of wear is 0.72, the probability of blockage is 0.15, and the probability of internal leakage is 0.13. Based on these probability values, the diagnostic system can determine that the hydraulic pump is currently in a state of wear and tear, and take corresponding measures in subsequent processing or maintenance.

[0087] Continuously collect operating status data to update the fault diagnosis model. Through parameter optimization, continuously improve the adaptability and diagnostic capabilities of the model to achieve long-term and intelligent online monitoring.

[0088] S5: Obtaining a posterior probability of the fault category based on the predicted probability and the joint confidence probability, and determining a fault diagnosis result based on the posterior probability.

[0089] In this embodiment, the predicted probability Combined with the joint confidence probability P(FD), the final posterior probability (i.e. confidence) is calculated, and the formula is expressed as:

[0090]

[0091] Among them, y represents the fault category and α is the weight coefficient, which is used to balance the influence of the fault diagnosis model and the dynamic confidence distribution.

[0092] According to the posterior probability P (F = yD), the fault category with the largest probability is selected as the fault diagnosis result. The fault category is such as piston pump wear, blockage, internal leakage, etc. The formula is expressed as:

[0093] y * = argmax y P(F=yD)

[0094] Among them, y * For the diagnosis result.

[0095] The final output diagnosis result y * and its confidence P(F=yD). When the confidence is lower than the threshold (such as 0.5), further monitoring is prompted.

[0096] The present invention combines dynamic confidence distribution with the fault diagnosis model CNN-LSTM to solve the problem that traditional optimization algorithms are prone to fall into local optimality and are difficult to find the global optimal solution. The present invention can realize online diagnosis of hydraulic piston pump faults. Real-time online diagnosis is mainly reflected in (1) the online monitoring system will continuously collect the operating data of the hydraulic piston pump. These signals are collected in real time by sensors. These data are dynamically changing and reflect the operating status of the hydraulic system in real time. (2) When the system detects a potential fault, it will comprehensively evaluate the current state based on the confidence distribution and CNN-LSTM prediction results. The system combines the probability value of fault prediction and the confidence of the signal through posterior probability fusion to give a reliability assessment of the fault, helping operators make quick decisions. (3) As the system continues to collect more data, the model will be updated and optimized in real time to adapt to different working conditions. This enables the system to continuously improve the accuracy of prediction and the sensitivity of fault identification during long-term operation.

[0097] Embodiment 2

[0098] This embodiment discloses a hydraulic plunger pump fault online diagnosis system based on dynamic confidence distribution, including:

[0099] A data acquisition module is configured to: acquire real-time operating status data of the hydraulic piston pump and perform preprocessing to obtain preprocessed real-time operating status data;

[0100] A feature extraction module is configured to: extract features from the preprocessed real-time running status data, and combine the extracted features to obtain a total feature vector;

[0101] A dynamic confidence distribution module is configured to: obtain a real-time signal of a hydraulic piston pump, and construct a dynamic confidence interval based on the real-time signal; if the signal value of the current real-time signal exceeds the dynamic confidence interval, calculate the confidence probability of the current real-time signal, and obtain the joint confidence probability of all signals based on the confidence probability;

[0102] A model prediction module is configured to: input the total feature vector into a pre-trained fault diagnosis model for prediction to obtain a predicted probability of a fault category;

[0103] The fusion diagnosis decision module is configured to obtain a posterior probability of the fault category based on the predicted probability and the joint confidence probability, and determine the fault diagnosis result based on the posterior probability.

[0104] Embodiment 3

[0105] The purpose of this embodiment is to provide a computing device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method of embodiment 1 when executing the program.

[0106] Embodiment 4

[0107] The purpose of this embodiment is to provide a computer-readable storage medium, a computer-readable storage medium having a computer program stored thereon, and when the program is executed by a processor, the steps of the method of embodiment 1 are performed.

[0108] The steps involved in the apparatus of the above embodiments 3 and 4 correspond to the method embodiment 1, and the specific implementation method can refer to the relevant description part of embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood to include any medium that can store, encode or carry an instruction set for execution by a processor and enable the processor to execute any method in the present invention.

[0109] Those skilled in the art should understand that the modules or steps of the present invention described above can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0110] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0111] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A hydraulic piston pump fault online diagnosis method based on dynamic confidence distribution, characterized in that: include: Acquire the real-time operation status data of the hydraulic piston pump and perform preprocessing to obtain the preprocessed real-time operation status data; Perform feature extraction on the preprocessed real-time operation status data, and combine the extracted features to obtain a total feature vector; Acquire a real-time signal of a hydraulic piston pump, and construct a dynamic confidence interval based on the real-time signal; if the signal value of the current real-time signal exceeds the dynamic confidence interval, calculate the confidence probability of the current real-time signal, and obtain the joint confidence probability of all signals based on the confidence probability; Inputting the total feature vector into a pre-trained fault diagnosis model for prediction to obtain a predicted probability of the fault category; A posterior probability of the fault category is obtained based on the predicted probability and the joint confidence probability, and a fault diagnosis result is determined based on the posterior probability.

2. The method for online fault diagnosis of hydraulic piston pump based on dynamic confidence distribution according to claim 1, characterized in that: The operating status data includes system pressure, output flow, pump body temperature, vibration signals in the XYZ directions and control current.

3. The method for online fault diagnosis of hydraulic piston pump based on dynamic confidence distribution according to claim 1, characterized in that: The specific steps to obtain the total feature vector are as follows: feature extraction is performed on the preprocessed operating status data, including the time domain features and frequency domain features of the signal. The time domain features include mean, standard deviation, kurtosis, skewness and root mean square value, and the frequency domain features include main frequency, spectrum energy and spectrum entropy. The time domain features and frequency domain features are combined to form a complete total feature vector.

4. The method for online fault diagnosis of hydraulic piston pump based on dynamic confidence distribution according to claim 1, characterized in that: The joint confidence probability of all signals is obtained as follows: Calculate real-time statistical parameters, including real-time mean and standard deviation; constructing a dynamic confidence interval for each signal based on the real-time mean and standard deviation; Calculating a confidence probability for each signal based on the dynamic confidence interval; The confidence summaries of all signals are weighted and calculated to obtain the joint confidence probability.

5. The method for online fault diagnosis of hydraulic piston pump based on dynamic confidence distribution according to claim 4, characterized in that: The dynamic confidence interval is constructed as: CI(t)=[μ(t)-z·σ(t),μ(t)+z·σ(t)] Where CI(t) represents the dynamic confidence interval of the t time window, μ(t) represents the real-time mean of the t time window, z represents the confidence coefficient, and σ(t) represents the standard deviation of the t time window.

6. The method for online fault diagnosis of hydraulic piston pump based on dynamic confidence distribution according to claim 4, characterized in that: If the signal value of the current real-time signal exceeds the dynamic confidence interval, the confidence probability of the current real-time signal is calculated, and the formula is expressed as: Among them, P(f i| D) represents the confidence probability, f i (t) represents the signal, μ(t) represents the real-time mean of the time window t, and σ(t) represents the standard deviation of the time window t.

7. The method for online fault diagnosis of hydraulic piston pump based on dynamic confidence distribution according to claim 1, characterized in that: The fault diagnosis model includes a convolutional neural network module, a long short-term memory network module and an output layer in sequence, and the output layer uses a Softmax function to calculate the predicted probability of the fault category.

8. The hydraulic piston pump fault online diagnosis system based on dynamic confidence distribution is characterized by: include: A data acquisition module is configured to: acquire real-time operating status data of the hydraulic piston pump and perform preprocessing to obtain preprocessed real-time operating status data; A feature extraction module is configured to: extract features from the preprocessed real-time running status data, and combine the extracted features to obtain a total feature vector; A dynamic confidence distribution module is configured to: obtain a real-time signal of a hydraulic piston pump, and construct a dynamic confidence interval based on the real-time signal; if the signal value of the current real-time signal exceeds the dynamic confidence interval, calculate the confidence probability of the current real-time signal, and obtain the joint confidence probability of all signals based on the confidence probability; A model prediction module is configured to: input the total feature vector into a pre-trained fault diagnosis model for prediction to obtain a predicted probability of a fault category; The fusion diagnosis decision module is configured to obtain a posterior probability of the fault category based on the predicted probability and the joint confidence probability, and determine the fault diagnosis result based on the posterior probability.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the online fault diagnosis method for hydraulic piston pump based on dynamic confidence distribution as described in any one of claims 1 to 7 are implemented.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps in the online diagnosis method for hydraulic piston pump fault based on dynamic confidence distribution as described in any one of claims 1 to 7 are implemented.

Images