CN113485302B - Fault diagnosis method and system for vehicle running process based on multivariate time series data - Google Patents

Abstract

The utility model discloses a vehicle operation process fault diagnosis method and system based on multivariate time series data, including: acquiring running state information in the running process of a vehicle; time sequence division is carried out on each running state information to obtain multivariate time sequence data; extracting correlation characteristics of the multivariate time sequence data from the multivariate time sequence data; extracting time dependency characteristics from the dependency characteristics of the multivariate time series data; and inputting the multivariate time sequence data and the time dependency characteristics into a trained fault detection and diagnosis model to obtain a fault detection and diagnosis result. The fault detection and diagnosis of the vehicle running process are realized.

Description

Fault diagnosis method and system for vehicle running process based on multivariate time series data

technical field

The invention relates to the technical field of vehicle running process fault detection, in particular to a vehicle running process fault diagnosis method and system based on multivariate time series data.

Background technique

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

The fault detection and diagnosis during vehicle operation is an important tool to remind the owner whether the vehicle is faulty, and to provide diagnostic information when the fault occurs. When the vehicle is running, there are some fault changes in the multivariate time series data returned by the component sensors. If the vehicle owner does not obtain the relevant information of the vehicle fault in time, it may lead to the loss of part of the vehicle or the stagnation of the entire operation. Therefore, the failure of the vehicle is accurately detected during the operation of the vehicle and the vehicle owner is reminded in time, which can avoid causing greater losses. At the same time, the vehicle operating system is complex. If the fault-related diagnostic information can be given at the same time as the fault is found to help find out the important factors causing the fault, it can provide very powerful help for the owner to troubleshoot the fault as soon as possible and restore the normal operation of the vehicle.

There are currently few methods for fault detection and diagnosis during vehicle operation. The time series data generated when the vehicle is running is random, and it is difficult to obtain the characteristic rules. There is a complex correlation between the time series data, and there is an imbalance of positive and negative samples. In addition, there are various types of vehicle operation failures, which may be caused by the effect of a single component. Or as a result of the joint action of multiple components, the cause of the failure is difficult to locate. The traditional methods based on manual and fault alarm devices have the problems of low efficiency, weak accuracy, and difficulty in locating the cause of the fault. Some fault detection based on traditional machine learning algorithm requires data to meet certain characteristic rules, and the commonly used classification-based fault detection algorithm is difficult to obtain high accuracy in the case of unbalanced positive and negative samples. Most of the fault detection algorithms based on deep learning only consider the time-dependent features of a single time series data, while ignoring the interaction between multiple time series data. In addition, the existing algorithms are more to make judgments on data failures, and rarely analyze the important factors that affect the occurrence of failures. It can be seen that the existing methods cannot effectively detect and diagnose faults during vehicle operation.

SUMMARY OF THE INVENTION

In order to solve the above problems, the present disclosure proposes a method and system for diagnosing faults during vehicle operation based on multivariate time series data, which realizes fault detection and diagnosis during vehicle operation.

To achieve the above object, the present disclosure adopts the following technical solutions:

In the first aspect, a fault diagnosis method for vehicle running process based on multivariate time series data is proposed, including:

Obtain the running status information during the running process of the vehicle;

Divide the time series of each operating state information to obtain multiple time series data;

Extract the correlation features of multivariate time series data from multivariate time series data;

Extract time-dependent features from correlation features of multivariate time series data;

Input multivariate time series data and time-dependent features into the trained fault detection and diagnosis model to obtain fault detection and diagnosis results.

In the second aspect, a fault diagnosis system for vehicle operation process based on multivariate time series data is proposed, including:

The data acquisition module is used to acquire the running status information during the running process of the vehicle;

The time sequence division module is used to divide the time sequence of each operating state information and obtain multiple time sequence data;

The correlation feature extraction module is used to extract the correlation features of the multivariate time series data from the multivariate time series data;

The time-dependent feature extraction module is used to extract time-dependent features from the correlation features of multivariate time series data;

The fault detection and diagnosis module is used to input multivariate time series data and time-dependent features into the trained fault detection and diagnosis model to obtain fault detection and diagnosis results.

In a third aspect, an electronic device is proposed, including a memory, a processor, and computer instructions stored in the memory and executed on the processor, and when the computer instructions are executed by the processor, complete a vehicle operation process based on multivariate time series data The steps described in the troubleshooting method.

In a fourth aspect, a computer-readable storage medium is provided, which is used for storing computer instructions, and when the computer instructions are executed by a processor, the steps described in the method for diagnosing faults in a vehicle running process based on multivariate time series data are completed.

Compared with the prior art, the beneficial effects of the present disclosure are:

1. The present disclosure also realizes the diagnosis of faults on the basis of detecting the faults in the running process of the vehicle by analyzing the multivariate time series data obtained by each sensor on the vehicle.

2. The present disclosure first extracts the correlation features of the multivariate time series data from the multivariate time series data, and then obtains the time dependence features from the correlation features of the multivariate time series data when performing fault detection and diagnosis on the vehicle running process. Time-dependent features include not only the time-dependent features of single time series data, but also the correlation features between multiple time-series data. The time-dependent features are used to detect and diagnose faults during vehicle operation, improving fault detection and diagnosis. accuracy.

Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will become apparent from the description which follows, or may be learned by practice of the invention.

Description of drawings

The accompanying drawings that form a part of the present application are used to provide further understanding of the present application, and the schematic embodiments and descriptions of the present application are used to explain the present application and do not constitute improper limitations on the present application.

1 is a flowchart of the method disclosed in Embodiment 1 of the present disclosure;

FIG. 2 is a model frame diagram disclosed in Embodiment 1 of the present disclosure;

FIG. 3 is a correlation feature extraction diagram of multivariate time series data disclosed in Embodiment 1 of the present disclosure;

FIG. 4 is a time-dependent feature extraction diagram disclosed in Embodiment 1 of the present disclosure;

FIG. 5 is a schematic diagram of the fault detection and diagnosis disclosed in Embodiment 1 of the present disclosure.

Detailed ways:

The present disclosure will be further described below with reference to the accompanying drawings and embodiments.

It should be noted that the following detailed description is exemplary and intended to provide further explanation of the application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

It should be noted that the terminology used herein is for the purpose of describing specific embodiments only, and is not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly dictates otherwise, the singular is intended to include the plural as well, furthermore, it is to be understood that when the terms "comprising" and/or "including" are used in this specification, it indicates that There are features, steps, operations, devices, components, and/or combinations thereof.

In this disclosure, terms such as "upper", "lower", "left", "right", "front", "rear", "vertical", "horizontal", "side", "bottom", etc. The orientation or positional relationship is based on the orientation or positional relationship shown in the drawings, and is only a relational word determined for the convenience of describing the structural relationship of each component or element of the present disclosure, and does not specifically refer to any component or element in the present disclosure, and should not be construed as a reference to the present disclosure. public restrictions.

In the present disclosure, terms such as "fixed connection", "connected", "connected", etc. should be understood in a broad sense, indicating that it may be a fixed connection, an integral connection or a detachable connection; it may be directly connected, or through an intermediate connection. media are indirectly connected. For the relevant scientific research or technical personnel in the field, the specific meanings of the above terms in the present disclosure can be determined according to specific situations, and should not be construed as limitations on the present disclosure.

Example 1

In order to realize the detection and diagnosis of vehicle running process faults, a method for vehicle running process fault diagnosis based on multivariate time series data is disclosed in this embodiment, including:

Obtain the running status information during the running process of the vehicle;

Divide the time series of each operating state information to obtain multiple time series data;

Extract the correlation features of multivariate time series data from multivariate time series data;

Extract time-dependent features from correlation features of multivariate time series data;

Input multivariate time series data and time-dependent features into the trained fault detection and diagnosis model to obtain fault detection and diagnosis results.

Further, the operating state information includes engine coolant temperature, engine speed, atmospheric pressure, engine voltage, engine torque, accelerator pedal opening, vehicle speed, and fuel consumption rate.

Further, before performing time sequence division on each operating state information, normalize each operating state information.

Further, input the multivariate time series data into the trained CNN network model to obtain the correlation features of the multivariate time series data.

Further, the Transformer Encoder module is used to extract time-dependent features from the correlation features of multivariate time series data.

Further, the fault detection and diagnosis model adopts generative adversarial network.

Further, the generative adversarial network includes a generator and a discriminator; the multivariate time series data and time-dependent features are input into the generator to obtain the reconstruction sequence and the reconstruction error; the reconstruction sequence and the multivariate time series data are input into the discriminator to obtain the discriminant. Obtain the fault score by using the discrimination score and reconstruction error; judge whether the multivariate time series is fault data according to the fault score; diagnose the fault according to the change of each operating state information in the multivariate time series data judged as fault data before and after reconstruction.

The method for detecting faults during vehicle operation based on multivariate time series data disclosed in this embodiment will be described in detail with reference to FIGS. 1-5 .

The vehicle running process fault detection method based on multivariate time series data, by analyzing the acquired multivariate time series data, realizes fault detection and diagnosis in the vehicle running process, as shown in Figures 1 and 2, mainly including: acquisition of multivariate time series data, multivariate time series data There are four stages of extraction of correlation features of time series data, extraction of time-dependent features, and fault detection and diagnosis.

Among them, the acquisition process of multivariate time series data is as follows:

Obtain the operating status information detected by the sensors of various components in the vehicle during vehicle operation. The operating status information includes engine coolant temperature, engine speed, atmospheric pressure, engine voltage, engine torque, accelerator pedal opening, vehicle speed, fuel consumption rate, etc. , which are time series data.

During specific implementation, the operating status information can be increased, decreased and replaced according to the actual needs of fault detection and diagnosis.

Since the dimensions of the values obtained by sensors of different components are different, the operating state information obtained by different sensors during vehicle operation is normalized, so that the information obtained by all sensors has the same scale.

For a certain vehicle running state information sequence {x ₁ ,x ₂ ,...,x _N }, the formula for normalizing the running state information is:

Use the sliding window to divide the running status information into time series to obtain multivariate time series data, which are as follows:

The operating status information returned by the sensors of various components during the operation of the vehicle contains continuous observations, which are usually collected at equidistant time stamps.

Define multivariate time series data as:

X={x ₁ , x ₂ , ..., x _N }

where N is the length of x, from 1 to N respectively representing a certain moment.

Define the observed value at time t as:

x _t is an M-dimensional vector, each dimension of the vector representing a component sensor, where t≤N, x∈R ^N×M .

The sliding window is the specified unit length to select time series data. Comparing the time series data to a scale, the sliding window is equivalent to a slider with a fixed length sliding on the scale, and the data in the slider is selected for each sliding unit. . Use x _{tT: t} (∈R ^(T+1)×M ) to define the observations from time tT to time t, that is, a sliding window of T+1 units:

{x _tT , x _t-T+1 , ..., x _t }

Expanding it is a time series matrix, which can be expressed as:

The extraction process of the correlation features of the multivariate time series data is as follows: using the CNN network to extract the correlation features of the multivariate time series data from the multivariate time series data.

The CNN network is adopted, and the collection of local information is checked with the help of the convolution in the CNN to extract the correlation features of multivariate time series data, as shown in Figure 3. There is a complex relationship between the various operating state information of the vehicle operation process data. For example, the opening of the accelerator pedal will affect the vehicle speed, and the speed is related to the engine speed, the fuel injection amount and the fuel consumption rate. The correlation features between the state information are extracted, that is, the correlation features between the dimensions of the multivariate time series data X are extracted.

Each convolutional layer in CNN consists of several convolutional units, and the parameters of each convolutional unit are optimized by the back-propagation algorithm. By setting multiple filters, different convolution units are used as different weight matrices to perform convolution operations on the data (the corresponding elements of tensors of the same specification are multiplied, and then added to obtain a new tensor). The purpose of the convolution operation is to capture local and detailed information, use different weight matrices to extract the correlation features between different dimensions of the input data, and finally combine and output a new feature matrix _{z c} , which is the obtained multivariate time series data. correlation characteristics.

The CNN network uses a one-dimensional convolutional layer and uses a linear rectification function (Rectified Linear Units, ReLU) as the activation function to non-linearly map the results of the convolutional layer. Compared with the traditional sigmoid activation function, ReLU can effectively alleviate the disappearance of gradients question. ReLU function It is a very simple function, when the input is less than zero, the output is zero, otherwise the output is equal to the input. ReLU function expression:

The specific process of extracting time-dependent features from correlation features of multivariate time series data is as follows: extracting time-dependent features from correlation features of multivariate time series data through Transformer Encoder module.

The running state information obtained during the running of the vehicle is the data sampled in time sequence, and contains many hidden information of the data in the time dimension. If a certain operating state information changes at a certain time, for example, the driver suddenly increases the opening of the accelerator pedal, then at a later time, other operating state information associated with the opening of the accelerator pedal, such as speed, engine speed , fuel consumption, etc. will vary. It can be seen that, by extracting the time-dependent feature of each operating state information, the change information of each operating state information can be obtained. The time-dependent feature extraction of time series data is realized by selecting multi-layer Transformer Encoder, and the specific structure of TransformerEncoder is shown in Figure 4. The specific process is as follows:

Perform position encoding on the correlation feature matrix z _c of the multivariate time series data extracted from the multivariate time series data to obtain the encoded feature matrix z _n , and use the encoded feature matrix z _n as the input of the Encoder, which can help determine the position of each vector in the sequence As well as the distance between different vectors, it also makes the model need to learn fewer parameters and the model trains faster.

Use the encoded feature matrix z _n added with the position encoding information as the input of the first Encoder, and then calculate the output directly as the input of the next Encoder, and continue until the last Encoder output, which is the final output, with This implements the extraction of time-dependent features. Encoder is mainly composed of two parts: the first part is the self-attention mechanism (Self-Attention), and the second part is the feedforward neural network. Residual connections are designed between sub-layers inside the unit, which can ensure that the information of the previous layer is completely transmitted to the next layer. The self-attention mechanism is the most important part of Transformer Encoder. The self-attention mechanism is improved from the attention mechanism. With the help of the self-attention mechanism, the dependence on external information can be reduced and the internal correlation characteristics of captured data can be enhanced. With the help of the self-attention mechanism, each position vector in the sequence can fuse the relevant information of each position vector before and after it.

First construct three weight matrices W _q , W _k , W _v , these three matrices are obtained through model training, and thus calculated:

Q=z _n W _q

K=z _n W _k

V=z _n W _v

According to the obtained Q, K, V matrix and then do the calculation:

Where d _k is the dimension of K, the final output of the self-attention mechanism is obtained, thereby realizing the extraction of time-dependent features.

Input multivariate time series data and time-dependent features into the trained fault detection and diagnosis model to obtain fault detection and diagnosis results.

In the specific implementation, the fault detection and diagnosis model adopts the Generative Adversarial Network (GAN), and the GNA includes a generator and a discriminator.

The generative adversarial network includes a generator (G) and a discriminator (D); input multivariate time series data and time-dependent features into the generator to obtain the reconstruction sequence and reconstruction error; input the reconstruction sequence and multivariate time series data into the discriminator , obtain the discriminant score; use the discriminant score and reconstruction error to obtain the fault score; judge whether the multivariate time series is fault data according to the fault score; Diagnose. As shown in Figure 5, the specific process is as follows:

(1) Use the GAN generator to get the reconstruction error.

The goal of the generator is to generate a reconstructed sequence G(z) that is as similar as possible to the real sample x _tT:t by learning the feature distribution of the real data, so that the discriminator cannot distinguish the real sample from the reconstructed sample.

Time-dependent feature matrix extracted from multivariate time series data:

Input to generator, output reconstruction sequence G(z):

The process of calculating the reconstruction error d _G is:

The matrix N is obtained by squaring each element in the matrix obtained by x-G(z):

Then, for the matrix N in units of columns, add up the elements and do the mean value to get:

N _tT : t = [n ¹ , n ² , ..., n ^M ]

Finally, each element of the matrix N _tT:t is averaged to obtain the reconstruction error d _G .

(2) Use the GAN discriminator to get the discriminant score.

The goal of the discriminator is to distinguish whether the input data is the real multivariate time series data x _tT:t or the reconstructed sequence G(z) of the generator. Ideally, the output of D is 1 if the input data is multivariate time series data, and 0 if the input data is the reconstructed sequence G(z) of the generator. The discrimination score d _D output by the discriminator is directly used as the output of the discriminator. The larger the value of d _D is, the more likely the data is to be the real data, and the smaller the value of d _D is, the more likely the data is to be reconstructed data.

(3) Use the reconstruction error and the discriminant score to obtain the fault score, and judge the sequence exceeding the threshold as normal data, otherwise it is fault data.

Take the difference between the output value d _D of the discriminator and the reconstruction error d _G of the generator as the failure score d _score :

d _score = d _D -d _G

的平均值为阈值th，然后在测试阶段对数据是否为故障数据做出判断：The failure score calculated by selecting the training phase

The average value of is the threshold th, and then judge whether the data is fault data in the test phase:

Where 0 represents fault data and 1 represents normal data.

(4) The multivariate time series data determined as fault data is used to diagnose the fault by comparing the change values of each dimension before and after data reconstruction, where the change value of each dimension is the change value of each operating state information.

The fault detection can be provided by calculating the fault score d _score . In the testing stage, if fault data is input, a large reconstruction error d _G will be obtained, and the fault is caused by the joint influence of each component sensor, that is, each operating state information. Therefore, the N _tT:t matrix obtained in the process of calculating the reconstruction error can be used to obtain the change of each operating state information before and after reconstruction. If the operating state information changes greatly before and after reconstruction, it means that the operating state information The impact of failure is greater. Based on this, the change of each operating state information before and after reconstruction is used as the basis for fault diagnosis, and the dimension where the change before and after reconstruction exceeds the set value is regarded as an important factor affecting the occurrence of the fault, so as to diagnose the fault.

The fault diagnosis method disclosed in this embodiment realizes the detection and diagnosis of faults in the running process of the vehicle by analyzing the multivariate time series data obtained by various sensors on the vehicle. The correlation features of multivariate time series data are extracted from multivariate time series data, and then time-dependent features are obtained from the correlation features of multivariate time series data. The time-dependent features not only include the time-dependent features of single time series data, but also include The correlation feature between multivariate time series data can be used to detect and diagnose faults during vehicle operation through the time-dependent features, which improves the accuracy of fault detection and diagnosis.

Example 2

In this embodiment, a fault diagnosis system for vehicle running process based on multivariate time series data is disclosed, including:

The data acquisition module is used to acquire the running status information during the running process of the vehicle;

The time sequence division module is used to divide the time sequence of each operating state information and obtain multiple time sequence data;

The correlation feature extraction module is used to extract the correlation features of the multivariate time series data from the multivariate time series data;

The time-dependent feature extraction module is used to extract time-dependent features from the correlation features of multivariate time series data;

The fault detection and diagnosis module is used to input multivariate time series data and time-dependent features into the trained fault detection and diagnosis model to obtain fault detection and diagnosis results.

Example 3

In this embodiment, an electronic device is disclosed, which includes a memory and a processor, and computer instructions stored in the memory and executed on the processor. When the computer instructions are executed by the processor, the based on The steps are described in the method for diagnosing vehicle operation process faults based on multivariate time series data.

Example 4

In this embodiment, a computer-readable storage medium is disclosed, which is used to store computer instructions, and when the computer instructions are executed by a processor, complete the method for diagnosing faults in a vehicle running process based on multivariate time series data disclosed in Embodiment 1. the steps described.

The above are only preferred embodiments of the present application, and are not intended to limit the present application. For those skilled in the art, the present application may have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this application shall be included within the protection scope of this application.

As will be appreciated by those skilled in the art, the embodiments of the present application may be provided as a method, a system, or a computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It will be understood that each process and/or block in the flowchart illustrations and/or block diagrams, and combinations of processes and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to the processor of a general purpose computer, special purpose computer, embedded processor or other programmable data processing device to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing device produce Means for implementing the functions specified in a flow or flow of a flowchart and/or a block or blocks of a block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory result in an article of manufacture comprising instruction means, the instructions The apparatus implements the functions specified in the flow or flow of the flowcharts and/or the block or blocks of the block diagrams.

These computer program instructions can also be loaded on a computer or other programmable data processing device to cause a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process such that The instructions provide steps for implementing the functions specified in the flow or blocks of the flowcharts and/or the block or blocks of the block diagrams.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: the present invention can still be Modifications or equivalent replacements are made to the specific embodiments of the present invention, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention shall be included within the protection scope of the claims of the present invention.

Claims (6)

1. The vehicle operation process fault diagnosis method based on the multivariate time series data is characterized by comprising the following steps of:

acquiring running state information in the running process of a vehicle;

time sequence division is carried out on each running state information to obtain multivariate time sequence data;

inputting the multivariate time sequence data into a trained CNN network model to obtain the correlation characteristics of the multivariate time sequence data;

extracting time dependency characteristics from the correlation characteristics of the multivariate time sequence data by using a Transformer Encoder module;

inputting the multivariate time sequence data and the time dependency characteristics into a trained fault detection and diagnosis model to obtain a fault detection and diagnosis result;

wherein, the fault detection and diagnosis model adopts a generation countermeasure network; the generation countermeasure network comprises a generator and an arbiter; inputting the multivariate time sequence data and the time-dependent characteristics into a generator to obtain a reconstruction sequence and a reconstruction error; inputting the reconstructed sequence and the multivariate time sequence data into a discriminator to obtain a discrimination score; obtaining a fault score by using the discrimination score and the reconstruction error; judging whether the multivariate time sequence is fault data or not according to the fault score; and diagnosing the fault according to the change before and after the reconstruction of each operation state information in the multi-element time sequence data of the fault data.

2. The method of claim 1, wherein the operating condition information includes an engine coolant temperature, an engine speed, a barometric pressure, an engine voltage, an engine torque, an accelerator pedal opening, a vehicle speed, and a fuel consumption rate.

3. The multivariate time series data-based vehicle operational process fault diagnosis method as defined in claim 1, wherein the operational state information is normalized before time-series division of the operational state information.

4. Vehicle operation process fault diagnosis system based on many units time series data, its characterized in that includes:

the data acquisition module is used for acquiring running state information in the running process of the vehicle;

the time sequence dividing module is used for carrying out time sequence division on each running state information to obtain multivariate time sequence data;

the correlation characteristic extraction module is used for inputting the multivariate time sequence data into the trained CNN network model to obtain the correlation characteristics of the multivariate time sequence data;

the time dependency characteristic extraction module is used for extracting time dependency characteristics from the correlation characteristics of the multivariate time sequence data by using the Transformer Encoder module; wherein, the fault detection and diagnosis model adopts a generation countermeasure network; the generation countermeasure network comprises a generator and an arbiter; inputting the multivariate time sequence data and the time-dependent characteristics into a generator to obtain a reconstruction sequence and a reconstruction error; inputting the reconstructed sequence and the multivariate time sequence data into a discriminator to obtain a discrimination score; obtaining a fault score by using the discrimination score and the reconstruction error; judging whether the multivariate time sequence is fault data or not according to the fault score; diagnosing the fault according to the change before and after the reconstruction of each operation state information in the multi-element time sequence data which is judged as fault data;

and the fault detection and diagnosis module is used for inputting the multivariate time sequence data and the time dependency characteristics into the trained fault detection and diagnosis model to obtain a fault detection and diagnosis result.

5. An electronic device comprising a memory and a processor and computer instructions stored in the memory and executed on the processor, wherein the computer instructions, when executed by the processor, perform the steps of the multivariate time series data based vehicle operation process fault diagnosis method as claimed in any one of claims 1-3.

6. A computer readable storage medium storing computer instructions which, when executed by a processor, perform the steps of the multivariate sequential data based vehicle operation process fault diagnosis method as claimed in any one of claims 1-3.

Images