Multi-way valve fault detection method and system based on deep learning model

Abstract

The invention relates to the technical field of multi-way valve fault diagnosis, and discloses a multi-way valve fault detection method and system based on a deep learning model, and the method comprises the steps: collecting the flow information of an oil inlet of a first multi-way valve; pressure information of an oil inlet and an oil outlet of the first multi-way valve is collected; configuration information of the first multi-way valve is obtained; and according to the flow information, the pressure information and the configuration information, the operation state and the fault type of the first multi-way valve are judged based on a CNN + LSTM model. By means of the system, pressure signals, flow signals and configuration information of an oil inlet and an oil outlet of the first multi-way valve can be collected, then the running state and the fault type of the first multi-way valve can be judged based on the LSTM + CNN model obtained through training, and the multi-way valve fault diagnosis and classification efficiency is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-way valve fault diagnosis, and in particular to a multi-way valve fault detection method and system based on a deep learning model. Background Art

[0002] The multi-way valve is the core hydraulic control component of some hydraulic system equipment, such as excavators. Its main function is to control the movement of the hydraulic actuator. The multi-way valve determines the working performance of the equipment, improves the degree of automation of the equipment, and effectively reduces the labor intensity. Due to the harsh working environment of equipment such as excavators, the working performance of the multi-way valve is easily affected by the environment and becomes unstable, and the possibility of multi-way valve failure also increases. Therefore, there is an urgent need for an automated detection system that can improve the efficiency of multi-way valve fault diagnosis. Summary of the invention

[0003] The present invention aims to at least solve the technical problem in the prior art that there is a lack of an automated detection system capable of improving the efficiency of multi-way valve fault diagnosis.

[0004] To this end, an object of the present invention is to propose a multi-way valve fault detection system based on a deep learning model, comprising:

[0005] Collecting flow information of the oil inlet of the first multi-way valve;

[0006] Collecting pressure information of the oil inlet and the oil outlet of the first multi-way valve;

[0007] Acquire configuration information of the first multi-way valve;

[0008] According to the flow information, the pressure information and the configuration information, based on the CNN+LSTM model, the operating status and fault type of the first multi-way valve are determined.

[0009] Furthermore, the CNN+LSTM model is obtained by training the initial neural network model, and the training process includes:

[0010] Collect the corresponding multi-way valve oil circuit data under different configurations and different faults;

[0011] Inputting the multi-way valve oil circuit data into the initial neural network model for model training to obtain a trained model;

[0012] The trained model is evaluated and optimized to obtain the final CNN+LSTM model.

[0013] Furthermore, the trained model is evaluated and optimized to finally obtain the CNN+LSTM model, including:

[0014] Using evaluation indicators to evaluate the trained model;

[0015] If the evaluation is qualified, the CNN+LSTM model is obtained; if the evaluation is unqualified, the trained model is optimized by adjusting model parameters, optimizing algorithms or improving model structures.

[0016] Furthermore, the multi-way valve oil circuit data includes flow data, pressure data and configuration data, and the configuration data includes the number of multi-way valve plates and pilot pressure reduction pressure data.

[0017] Furthermore, the CNN+LSTM model includes an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a Reshape layer, an LSTM layer, a hidden layer and an output layer which are arranged in sequence.

[0018] Furthermore, the multi-way valve oil circuit data is input into the initial neural network model for model training to obtain a trained model, including:

[0019] Cleaning the oil circuit data of the multi-way valve;

[0020] Normalizing the cleaned oil circuit data of the multi-way valve;

[0021] The normalized data is input into the initial neural network model for model training.

[0022] Furthermore, the evaluation indicators include accuracy, precision, recall, F1 score and AUC-ROC.

[0023] Furthermore, the fault types include main overflow valve fault, main pressure reducing valve fault, shuttle valve fault, proportional pressure reducing valve fault, pilot valve oil circuit blockage fault, pressure compensation valve fault and reversing joint stuck fault.

[0024] The present invention provides a multi-way valve fault detection system based on a deep learning model, which is implemented using the multi-way valve fault detection method based on a deep learning model, and includes a first flow meter, a first pressure gauge, a first data acquisition card, a diagnostic controller, and a first system configuration module;

[0025] The first flow meter is arranged on the pipeline between the power source and the oil inlet of the first multi-way valve, and is used to collect flow information of the oil inlet and transmit it to the data acquisition card; the first pressure gauges are respectively arranged at the oil inlet and the oil outlet of the oil circuit of the first multi-way valve, and are used to collect pressure information at the oil inlet and the oil outlet, and transmit it to the first data acquisition card; the first flow meter and the first pressure sensor are both electrically connected to the data acquisition card; the diagnosis controller is electrically connected to the first data acquisition card; the first system configuration module is electrically connected to the diagnosis controller;

[0026] The first data acquisition card transmits the flow information and the pressure information to the diagnostic controller; the first system configuration module transmits the configuration information of the first multi-way valve to the diagnostic controller, and the diagnostic controller determines the operating status and fault type of the first multi-way valve based on the flow information, the pressure information and the configuration information.

[0027] Further, the initial neural network model is trained using a fault data acquisition system to obtain the CNN+LSTM model, wherein the fault data acquisition system includes a second flow meter, a second pressure gauge, a second multi-way valve, a second data acquisition card, a data acquisition computer, a data storage device, and a second system configuration module;

[0028] The second flow meter, the second pressure gauge and the second system configuration module are used to collect the multi-way valve oil circuit data corresponding to the second multi-way valve under different configurations and different faults, and transmit them to the second data acquisition card;

[0029] The second data acquisition card transmits the multi-way valve oil circuit data to the data acquisition computer; the data acquisition computer transmits the multi-way valve oil circuit data to the data storage device for storage;

[0030] The data acquisition computer inputs the multi-way valve oil circuit data into the initial neural network model for model training to obtain a trained model, evaluates and optimizes the trained model, and finally obtains the CNN+LSTM model.

[0031] The present invention proposes a multi-way valve fault detection method and system based on a deep learning model, which has the following beneficial effects:

[0032] The multi-way valve fault diagnosis system designed by the present invention can effectively collect the pressure signals, flow signals and configuration information of the oil inlet and outlet of the first multi-way valve, and then judge the operating state and fault type of the first multi-way valve based on the LSTM+CNN model trained by the present invention, which can effectively improve the efficiency of multi-way valve fault diagnosis and classification;

[0033] The multi-way valve fault data acquisition system designed by the present invention is used for the training of the LSTM+CNN model. The multi-way valve fault data acquisition system can effectively collect the pressure data of the oil inlet and outlet of the second multi-way valve, the flow data of the oil circuit, and the configuration information, providing data support for model training;

[0034] According to the characteristics of multi-way valve fault diagnosis, the present invention designs the input layer of the LSTM+CNN model; the CNN layer structure, and the single-layer pooling layer are selected to retain more feature information; the Reshape layer performs data conversion; the LSTM layer captures the long-term dependencies in the time series data, and finally outputs the diagnosis result probability through two fully connected layers. The LSTM+CNN model structure design takes into account the characteristics of the spatial dimension of the collected multi-way valve oil circuit data, the dependency of the time dimension, and retains more feature information, so that the LSTM+CNN model has better diagnostic performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0036] Figure 1 is a schematic diagram of a multi-way valve fault detection system based on a deep learning model in an embodiment of the present invention;

[0037] Figure 2 It is a schematic diagram of a fault data acquisition system of a multi-way valve fault detection method based on a deep learning model in an embodiment of the present invention.

[0038] Figure 3 It is a flow chart of a multi-way valve fault detection method based on a deep learning model in an embodiment of the present invention. DETAILED DESCRIPTION

[0039] Various aspects and features of the present invention are described herein with reference to the accompanying drawings.

[0040] It should be understood that various modifications may be made to the embodiments of the present application. Therefore, the above description should not be considered as limiting, but only as an example of the embodiments. Other modifications within the scope and spirit of the present invention will occur to those skilled in the art.

[0041] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the invention and, together with the general description of the invention given above and the detailed description of the embodiments given below, serve to explain the principles of the invention.

[0042] These and other characteristics of the invention will become apparent from the following description of a preferred form of embodiment given as a non-limiting example, with reference to the accompanying drawings.

[0043] It should also be understood that, although the invention has been described with reference to some specific examples, those skilled in the art will be able to realize many other equivalent forms of the invention that have the characteristics recited in the claims and are therefore within the scope of protection defined thereby.

[0044] The above and other aspects, features and advantages of the present invention will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings.

[0045] Specific embodiments of the present invention are described hereinafter with reference to the accompanying drawings; however, it should be understood that the embodiments claimed are merely examples of the present invention, which may be implemented in a variety of ways. Well-known and / or repeated functions and structures are not described in detail to avoid obscuring the present invention with unnecessary or redundant details. Therefore, the specific structural and functional details claimed herein are not intended to be limiting, but merely serve as a basis and representative basis for the claims to teach those skilled in the art to use the present invention in a variety of ways with substantially any suitable detailed structure.

[0046] In the description of the present invention, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0047] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

[0048] Example

[0049] like Figure 3 As shown, this embodiment provides a multi-way valve fault detection system based on a deep learning model, comprising the following steps:

[0050] Step S1: collecting flow information of the oil inlet of the first multi-way valve;

[0051] Step S2: collecting pressure information of the oil inlet and the oil outlet of the first multi-way valve;

[0052] Step S3: obtaining configuration information of the first multi-way valve;

[0053] Step S4: According to the flow information, the pressure information and the configuration information, based on the CNN+LSTM model, determine the operating status and fault type of the first multi-way valve.

[0054] like Figure 1 As shown, in order to implement the above-mentioned multi-way valve fault detection method based on deep learning model, the present invention provides a multi-way valve fault detection system based on deep learning model, including a first flow meter, a first pressure gauge, a first data acquisition card, a diagnostic controller and a first system configuration module;

[0055] The first flow meter is arranged on the pipeline between the power source and the oil inlet of the first multi-way valve, and is used to collect flow information of the oil inlet and transmit it to the data acquisition card;

[0056] The plurality of first pressure gauges are respectively arranged at the oil inlet and the oil outlet of the first multi-way valve oil circuit, including the oil inlet P, the oil return port T, the LS port and the Z port of the first multi-way valve oil circuit, for collecting the pressure information of the oil inlet P, the pressure information of the oil return port T, the pressure information of the LS port and the inlet pressure of the proportional pressure reducing valve at the Z port, and transmitting them to the first data acquisition card; the first flow meter and the first pressure sensor are both electrically connected to the data acquisition card; the diagnostic controller is electrically connected to the first data acquisition card; the first system configuration module is electrically connected to the diagnostic controller;

[0057] The first data acquisition card transmits the flow information and the pressure information to the diagnostic controller; the first system configuration module transmits the configuration information of the first multi-way valve to the diagnostic controller, and the diagnostic controller determines the operating status and fault type of the first multi-way valve based on the flow information, the pressure information and the configuration information.

[0058] According to another specific embodiment of the present invention, on the basis of step S4, according to the flow information, the pressure information and the configuration information, based on the CNN+LSTM model, the operating state and the fault type of the first multi-way valve are judged, wherein the CNN+LSTM model is obtained by training the initial neural network model, and the training process includes:

[0059] Step S41: collecting corresponding multi-way valve oil circuit data under different configurations and different faults;

[0060] Step S42: After processing the multi-way valve oil circuit data, the data is input into the initial neural network model for model training, and the weights and parameters of the model are continuously adjusted to obtain a trained model;

[0061] Step S43: Evaluate and optimize the trained model to finally obtain the CNN+LSTM model.

[0062] According to another specific embodiment of the present invention, based on the step S43, the trained model is evaluated and optimized to finally obtain the CNN+LSTM model, including:

[0063] Step S431: Evaluate the trained model using evaluation indicators. If the evaluation is qualified, the CNN+LSTM model is obtained.

[0064] Step S432: If the evaluation is unqualified, the trained model is optimized by adjusting model parameters, optimizing algorithms or improving model structures.

[0065] According to another specific embodiment of the present invention, on the basis of step S41, the multi-way valve oil circuit data corresponding to different configurations and different faults are collected, and the multi-way valve oil circuit data includes flow data, pressure data and configuration data, and the configuration data includes the number of multi-way valve plates and pilot pressure reducing pressure data.

[0066] According to another specific embodiment of the present invention, on the basis of the above embodiment, the initial neural network model is trained using a fault data acquisition system to obtain the CNN+LSTM model, and the fault data acquisition system includes a second flow meter, a second pressure gauge, a second multi-way valve, a second data acquisition card, a data acquisition computer, a data storage device and a second system configuration module;

[0067] like Figure 2 As shown, the power source provides high-pressure oil to the second multi-way valve to drive the entire system to operate.

[0068] The second flow meter is set at the flow rate of the oil inlet P of the second multi-way valve;

[0069] The second pressure gauge is installed at the oil inlet P of the second multi-way valve to measure the inlet pressure of the second multi-way valve, and a pressure gauge 2 is installed behind the T port of the multi-way valve to measure the outlet pressure;

[0070] The second pressure gauge is installed at the LS port of the second multi-way valve to measure the LS pressure;

[0071] The second pressure gauge is installed at the Z port of the second multi-way valve to measure the inlet pressure of the proportional pressure reducing valve;

[0072] The second flow meter and the plurality of the second pressure gauges are electrically connected to the second data acquisition card, and the second data acquisition card is connected to the data acquisition computer. According to the interface output by the above-mentioned sensor, the data of the sensor is converted and acquired, and transmitted to the data acquisition computer through the communication protocol.

[0073] The second system configuration module is connected to the data acquisition computer and provides the data acquisition computer with configuration data of the second multi-way valve; the second system configuration module can be an independent system that communicates with the data acquisition computer, or a module running on the data acquisition computer.

[0074] The second flow meter, the second pressure gauge and the second system configuration module are used to collect the multi-way valve oil circuit data corresponding to the second multi-way valve under different configurations and different faults, and transmit them to the second data acquisition card;

[0075] The second data acquisition card transmits the multi-way valve oil circuit data to the data acquisition computer; the data acquisition computer transmits the multi-way valve oil circuit data to the data storage device for storage;

[0076] The data acquisition computer is used to process the multi-way valve oil circuit data, input it into the initial neural network model for model training, continuously adjust the weights and parameters of the model to obtain the trained model, evaluate and optimize the trained model, and finally obtain the CNN+LSTM model.

[0077] According to another specific embodiment of the present invention, on the basis of the above embodiment, the CNN+LSTM model includes an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a Reshape layer, an LSTM layer, a hidden layer and an output layer which are arranged in sequence.

[0078] More specifically, the Reshape layer is used between CNN and LSTM to reshape the data to achieve data transfer between the CNN layer and the LSTM layer; finally, the classification results are output through two fully connected layers.

[0079] The input data of the input layer includes 4 channels of pressure data, 1 channel of flow data, and configuration data (preprocessed into numerical form). The input data enters the CNN+LSTM model through the input layer, and the input shape is (batch_size, time_steps = 100, num_features), where num_features = 5 channels of signals + number of configuration data.

[0080] The first convolution layer of the CNN layer extracts local features from the input signal. The number of convolution kernels is filters = 64, and the convolution kernel_size is 3. The activation function is ReLU.

[0081] The first pooling layer uses MaxPooling to reduce feature dimensions, reduce computational effort, and prevent overfitting. The pooling window size is 2 and the step size is 2.

[0082] The second convolutional layer further extracts deeper features. The number of convolution kernels is filters = 128, the convolution kernel size is 3, and the activation function is ReLU.

[0083] In order to retain more feature information, no pooling is performed after the second convolutional layer.

[0084] Reshape the data between CNN and LSTM through the Reshape layer to achieve data transfer between the CNN layer and the LSTM layer;

[0085] LSTM captures long-term dependencies in time series data. The number of LSTM units = 512, and the activation functions are tanh and sigmoid. Tanh is used to update the cell state inside the LSTM unit and calculate the output gate. The sigmoid function is used to calculate the forget gate, input gate, and output gate to control the flow of information.

[0086] The hidden layer of the fully connected layer (Dense) maps the output of LSTM to the label of the fault classification. The number of units = 64, and the activation function is ReLU.

[0087] The output layer of the fully connected layer (Dense) outputs the fault category probability. The number of units is 7, the number of fault categories, and the activation function is Softmax.

[0088] According to another specific embodiment of the present invention, on the basis of step S42, after processing the multi-way valve oil circuit data, the data is input into the initial neural network model for model training, and the weights and parameters of the model are continuously adjusted to obtain a trained model, including:

[0089] The multi-way valve oil circuit data is first cleaned and then normalized, and the normalized data is input into the initial neural network model for model training.

[0090] Data normalization is to scale the data so that it falls into a small specific interval (such as [0,1]) to reduce the impact of dimensional differences between different features on model training. In the present invention, the following formula is used to normalize the pressure or flow data collected from each channel:

[0091]

[0092] Where p is the original value of a certain pressure or flow, p min is the minimum value of a certain pressure or flow, p max is the maximum value of a certain pressure or flow, and p' is the normalized pressure or flow value of a certain road.

[0093] According to another specific embodiment of the present invention, based on the step S43, the evaluation indicators include accuracy, precision, recall, F1 score and AUC-ROC.

[0094] According to another specific embodiment of the present invention, based on the step S431, the fault types include main overflow valve failure, main pressure reducing valve failure, shuttle valve failure, proportional pressure reducing valve failure, pilot valve oil circuit blockage failure, pressure compensation valve failure and reversing joint stuck failure.

[0095] The first system configuration module and the second system configuration module are configured according to the structure and composition of the multi-way valve. The configuration information and configuration data include the number of multi-way valve slices, pilot pressure reduction pressure, etc.

[0096] The present invention proposes a multi-way valve fault detection method and system based on a deep learning model, which has the following beneficial effects:

[0097] The multi-way valve fault diagnosis system designed by the present invention can effectively collect the pressure signals, flow signals and configuration information of the oil inlet and outlet of the first multi-way valve, and then judge the operating state and fault type of the first multi-way valve based on the LSTM+CNN model trained by the present invention, which can effectively improve the efficiency of multi-way valve fault diagnosis and classification;

[0098] The multi-way valve fault data acquisition system designed by the present invention is used for the training of the LSTM+CNN model. The multi-way valve fault data acquisition system can effectively collect the pressure data of the oil inlet and outlet of the second multi-way valve, the flow data of the oil circuit, and the configuration information, providing data support for model training;

[0099] According to the characteristics of multi-way valve fault diagnosis, the present invention designs the input layer of the LSTM+CNN model; the CNN layer structure, and the single-layer pooling layer are selected to retain more feature information; the Reshape layer performs data conversion; the LSTM layer captures the long-term dependencies in the time series data, and finally outputs the diagnosis result probability through two fully connected layers. The LSTM+CNN model structure design takes into account the characteristics of the spatial dimension of the collected multi-way valve oil circuit data, the dependency of the time dimension, and retains more feature information, so that the LSTM+CNN model has better diagnostic performance.

[0100] The present disclosure also provides an electronic device, comprising at least a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned multi-way valve fault detection method based on a deep learning model when executing the computer program on the memory, and the method comprises the following steps:

[0101] Step S1: collecting flow information of the oil inlet of the first multi-way valve;

[0102] Step S2: collecting pressure information of the oil inlet and the oil outlet of the first multi-way valve;

[0103] Step S3: obtaining configuration information of the first multi-way valve;

[0104] Step S4: According to the flow information, the pressure information and the configuration information, based on the CNN+LSTM model, determine the operating status and fault type of the first multi-way valve.

[0105] In some embodiments, the processor executing the computer program may be a processing device including one or more general-purpose processing devices, such as a microprocessor, a central processing unit (CPU), a graphics processing unit (GPU), etc. More specifically, the processor may be a complex instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a processor running other instruction sets, or a processor running a combination of instruction sets. The processor may also be one or more special-purpose processing devices, such as an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a digital signal processor (DSP), a system on a chip (SoC), etc.

[0106] The memory may be a read-only memory (ROM), a random access memory (RAM), a phase-change random access memory (PRAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), an electrically erasable programmable read-only memory (EEPROM), other types of random access memory (RAM), a flash disk or other form of flash memory, a cache, a register, a static memory, a compact disk read-only memory (CD-ROM), a digital versatile disk (DVD) or other optical storage, a cassette or other magnetic storage device, or any other possible non-temporary medium used to store information or instructions that can be accessed by a computer device.

[0107] The electronic devices of the embodiments of the present disclosure may include but are not limited to fixed terminal devices such as MCU controllers, servers, desktop computers, digital TVs, etc., and mobile terminal devices such as vehicle-mounted devices (such as head-up display devices), handheld devices (such as mobile phones, tablet computers, etc.), wearable devices (such as smart watches, smart bracelets, etc.), etc.

[0108] The present disclosure also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the multi-way valve fault detection method based on a deep learning model is implemented. The method includes the following steps:

[0109] Step S1: collecting flow information of the oil inlet of the first multi-way valve;

[0110] Step S2: collecting pressure information of the oil inlet and the oil outlet of the first multi-way valve;

[0111] Step S3: obtaining configuration information of the first multi-way valve;

[0112] Step S4: According to the flow information, the pressure information and the configuration information, based on the CNN+LSTM model, determine the operating status and fault type of the first multi-way valve.

[0113] The present invention has the following beneficial effects:

[0114] The multi-way valve fault diagnosis system designed by the present invention can effectively collect the pressure signals, flow signals and configuration information of the oil inlet and outlet of the first multi-way valve, and then judge the operating state and fault type of the first multi-way valve based on the LSTM+CNN model trained by the present invention, which can effectively improve the efficiency of multi-way valve fault diagnosis and classification;

[0115] The multi-way valve fault data acquisition system designed by the present invention is used for the training of the LSTM+CNN model. The multi-way valve fault data acquisition system can effectively collect the pressure data of the oil inlet and outlet of the second multi-way valve, the flow data of the oil circuit, and the configuration information, providing data support for model training;

[0116] According to the characteristics of multi-way valve fault diagnosis, the present invention designs the input layer of the LSTM+CNN model; the CNN layer structure, and the single-layer pooling layer are selected to retain more feature information; the Reshape layer performs data conversion; the LSTM layer captures the long-term dependencies in the time series data, and finally outputs the diagnosis result probability through two fully connected layers. The LSTM+CNN model structure design takes into account the characteristics of the spatial dimension of the collected multi-way valve oil circuit data, the dependency of the time dimension, and retains more feature information, so that the LSTM+CNN model has better diagnostic performance.

[0117] The computer-readable storage medium of the embodiments of the present disclosure may adopt any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. In the embodiments of the present disclosure, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, device or device, for example, the memory described above.

[0118] The computer programs of the embodiments of the present disclosure may be organized into one or more computer executable components or modules. Any number and combination of such components or modules may be used to implement the various aspects of the present disclosure. For example, the various aspects of the present disclosure are not limited to the specific computer executable instructions or specific components or modules shown in the drawings and described herein. Other embodiments may include different computer executable instructions or components with more or less functions than those shown and described herein.

[0119] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

Claims

1. A multi-way valve fault detection method based on deep learning model, characterized in that: include: Collecting flow information of the oil inlet of the first multi-way valve; Collecting pressure information of the oil inlet and the oil outlet of the first multi-way valve; Acquire configuration information of the first multi-way valve; According to the flow information, the pressure information and the configuration information, based on the CNN+LSTM model, the operating status and fault type of the first multi-way valve are determined.

2. The multi-way valve fault detection method based on deep learning model according to claim 1 is characterized in that: The CNN+LSTM model is obtained by training the initial neural network model. The training process includes: Collect the corresponding multi-way valve oil circuit data under different configurations and different faults; Inputting the multi-way valve oil circuit data into the initial neural network model for model training to obtain a trained model; The trained model is evaluated and optimized to obtain the final CNN+LSTM model.

3. The multi-way valve fault detection method based on deep learning model according to claim 2 is characterized in that: The trained model is evaluated and optimized to finally obtain the CNN+LSTM model, including: Using evaluation indicators to evaluate the trained model; If the evaluation is qualified, the CNN+LSTM model is obtained; if the evaluation is unqualified, the trained model is optimized by adjusting model parameters, optimizing algorithms or improving model structures.

4. The multi-way valve fault detection method based on deep learning model according to claim 2 is characterized in that: The multi-way valve oil circuit data includes flow data, pressure data and configuration data, and the configuration data includes the number of multi-way valve plates and pilot pressure reduction pressure data.

5. The multi-way valve fault detection method based on deep learning model according to claim 1 is characterized in that: The CNN+LSTM model includes an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a Reshape layer, an LSTM layer, a hidden layer and an output layer which are arranged in sequence.

6. The multi-way valve fault detection method based on deep learning model according to claim 2 is characterized in that: Inputting the multi-way valve oil circuit data into the initial neural network model for model training to obtain a trained model, including: Cleaning the oil circuit data of the multi-way valve; Normalizing the cleaned oil circuit data of the multi-way valve; The normalized data is input into the initial neural network model for model training.

7. The multi-way valve fault detection method based on deep learning model according to claim 2 is characterized in that: The evaluation indicators include accuracy, precision, recall, F1 score and AUC-ROC.

8. The multi-way valve fault detection method based on deep learning model according to claim 1 is characterized in that: The fault types include main relief valve fault, main pressure reducing valve fault, shuttle valve fault, proportional pressure reducing valve fault, pilot valve oil circuit blockage fault, pressure compensation valve fault and reversing joint stuck fault.

9. A multi-way valve fault detection system based on a deep learning model, characterized in that: The method is implemented by using a multi-way valve fault detection method based on a deep learning model as described in any one of claims 1 to 8, comprising a first flow meter, a first pressure gauge, a first data acquisition card, a diagnostic controller and a first system configuration module; The first flow meter is arranged on the pipeline between the power source and the oil inlet of the first multi-way valve, and is used to collect flow information of the oil inlet and transmit it to the data acquisition card; the first pressure gauges are respectively arranged at the oil inlet and the oil outlet of the oil circuit of the first multi-way valve, and are used to collect pressure information at the oil inlet and the oil outlet, and transmit it to the first data acquisition card; the first flow meter and the first pressure sensor are both electrically connected to the data acquisition card; the diagnosis controller is electrically connected to the first data acquisition card; the first system configuration module is electrically connected to the diagnosis controller; The first data acquisition card transmits the flow information and the pressure information to the diagnostic controller; the first system configuration module transmits the configuration information of the first multi-way valve to the diagnostic controller, and the diagnostic controller determines the operating status and fault type of the first multi-way valve based on the flow information, the pressure information and the configuration information.

10. The multi-way valve fault detection system based on deep learning model according to claim 9, characterized in that: The initial neural network model is trained by using a fault data acquisition system to obtain the CNN+LSTM model, wherein the fault data acquisition system includes a second flow meter, a second pressure gauge, a second multi-way valve, a second data acquisition card, a data acquisition computer, a data storage device, and a second system configuration module; The second flow meter, the second pressure gauge and the second system configuration module are used to collect the multi-way valve oil circuit data corresponding to the second multi-way valve under different configurations and different faults, and transmit them to the second data acquisition card; The second data acquisition card transmits the multi-way valve oil circuit data to the data acquisition computer; the data acquisition computer transmits the multi-way valve oil circuit data to the data storage device for storage; The data acquisition computer inputs the multi-way valve oil circuit data into the initial neural network model for model training to obtain a trained model, evaluates and optimizes the trained model, and finally obtains the CNN+LSTM model.

Images