CN115526211A - Fault diagnosis method for valve-controlled hydraulic cylinder system based on load port independent control - Google Patents

Abstract

The invention provides a fault diagnosis method for a valve-controlled hydraulic cylinder system independently controlled by a load port, which comprises the following steps: step S1: establishing a valve control hydraulic cylinder system; step S2: fault injection is carried out to generate eight types of faults; and step S3: acquiring signals to obtain nine types of signals and generate a sample; and step S4: preprocessing a sample; step S5: after the pretreatment is finished, dividing the generated sample into a training set, a verification set and a test set according to a proportion; step S6: the training set and the verification set are used for training a fault diagnosis algorithm; step S7: and judging the performance of the trained algorithm model after the test of the test set. According to the invention, the element characteristics under long time span are excavated by a deep learning algorithm according to the sensor information of a plurality of components of the valve-controlled hydraulic cylinder system which is independently controlled towards the load port, and the specific element with a fault of the system is identified, so that the fault diagnosis work of the valve-controlled hydraulic cylinder system which is independently controlled towards the load port is realized.

Description

Fault diagnosis method for valve-controlled hydraulic cylinder system oriented to independent control of load ports

technical field

The present invention relates to the technical field of industrial process fault diagnosis and pattern recognition, in particular, to a valve-controlled hydraulic cylinder system fault diagnosis method oriented to the independent control of the load port, especially to a method based on one-dimensional large convolution kernel and residual network A fault diagnosis method for a valve-controlled hydraulic cylinder system with independent control of the load port.

Background technique

The valve-controlled hydraulic cylinder system has the characteristics of high power, high precision, and fast response. It is often used in typical hydraulic systems such as tractor power shift transmission systems and hybrid electric vehicle drive systems. The traditional valve-controlled hydraulic cylinder servo control system uses a three-position four-way valve to complete the control of the hydraulic cylinder. The low flexibility and high energy loss brought about by the coupled mechanical structure limit its control accuracy and efficiency. In contrast, the valve-controlled hydraulic cylinder system for independent control of the load port has the characteristics of independent control of two chambers, so that the system has high precision and high flexibility while achieving low energy consumption.

At present, some research has been carried out at home and abroad on the valve-controlled hydraulic cylinder system oriented to the independent control of the load port. The existing technical scheme designs a load port independent control system with continuously adjustable oil return pressure. By adding an electric proportional relief valve with adjustable pressure on the oil return line and connecting a check valve in parallel, the control system is solved. Cavitation problems in the low-pressure chamber of the actuator in low-pressure regenerative mode and pressure loss in normal mode. The existing technical scheme designs a hydraulic system and control method based on two-stage energy supply and independent valve control of the load port. Through the fuzzy sliding mode variable structure control strategy, the control accuracy of the hydraulic drive unit and the efficiency of the system are improved. At the same time, a certain amount of research has been carried out on the fault diagnosis of the valve-controlled hydraulic cylinder system at home and abroad. The existing technical scheme designs a pneumatic control valve fault diagnosis method based on a low-deviation random configuration network, combined with the principal component analysis method to realize the fault diagnosis of the pneumatic control valve, and has a high diagnostic accuracy. The existing technical scheme designs a fault diagnosis model based on a high-dimensional nonlinear classifier, completes the fault diagnosis of the simulated hydraulic cylinder model established in the AMESim software, solves the problem of nonlinearity and small samples, and improves the fault diagnosis model fault identification capability.

There are certain limitations in the prior art, as follows:

1. The research on the valve-controlled hydraulic cylinder system oriented to the independent control of the load port mainly focuses on the structural design of the system, the design of the controller and the application of the system. The overall performance of the system has a greater impact.

2. The fault diagnosis of the existing valve-controlled hydraulic cylinder system is mainly aimed at a single component in the system, such as a valve or a hydraulic cylinder, ignoring the system-level signal differences brought about by the multi-sensor feedback and control strategy of the system, and lacking the ability to diagnose system-level faults. Diagnostic research.

3. Existing research has carried out active fault-tolerant control research on hydraulic valves with independent control of the valve port. Based on the known fault situation, the control strategy design has been carried out for hydraulic valves with independent valve ports, but there is a lack of control strategies for this type of hydraulic valve. Diagnosis and identification of faults.

Therefore, need to propose a kind of new technical scheme.

Contents of the invention

In view of the defects in the prior art, the object of the present invention is to provide a fault diagnosis method for a valve-controlled hydraulic cylinder system oriented to the independent control of the load port.

According to a valve-controlled hydraulic cylinder system fault diagnosis method for independent control of the load port provided by the present invention, the method includes the following steps:

Step S1: establishing a valve-controlled hydraulic cylinder system;

Step S2: Perform fault injection to generate eight types of faults;

Step S3: Acquire signals, obtain nine types of signals, and generate samples;

Step S4: Preprocessing the samples;

Step S5: After the preprocessing is completed, the generated samples are divided into training set, verification set and test set according to the proportion;

Step S6: the training set and the verification set are used for the training of the fault diagnosis algorithm;

Step S7: After the trained algorithm model is tested by the test set, the performance is judged.

Preferably, the faults in step S2 include four high-speed switching valves, two hydraulic reversing valves, hydraulic cylinders and displacement sensors.

Preferably, the PWM driving voltage of the high-speed switching valve is provided by supporting software and hardware.

Preferably, the failure of the hydraulic reversing valve is: the failure of the oil particles and the impact of the oil to the spool, the leakage and the damage of the sealing ring, and the failure of the spring.

Preferably, the displacement sensor is installed on the valve body.

Preferably, the signals in step S3 are composed of nine categories of faults and system normal states in step S2, and each category collects two main valve control chamber pressure signals and spool displacement signals, and collects hydraulic cylinder The pressure signal of the control chamber and the displacement signal of the rod are sampled at a sampling frequency of 1000Hz for a total of 2s; each category of normal data and fault data contains 900 sets of data, each set of data includes 9 features, and each feature includes 2000 data points to generate samples .

为预处理后结果：Preferably, the step S4 preprocesses the sample by the following formula, where x _m represents the mth feature,

For the preprocessed result:

Preferably, the ratio in step S5 is 6:2:1.

Compared with the prior art, the present invention has the following beneficial effects:

1. According to the sensor information of multiple components of the valve-controlled hydraulic cylinder system independently controlled by the load port, the present invention will design a deep learning algorithm to mine the component characteristics under a long time span, identify the specific component that the system fails, and realize the load-oriented Fault diagnosis of valve-controlled hydraulic cylinder system with independent control;

2. The present invention constructs a valve-controlled hydraulic cylinder system with independent control of the load port, collects sensor information of multiple components to express system characteristics; designs a deep learning algorithm, especially a one-dimensional large convolution kernel and residual network The advanced deep learning algorithm performs multi-source signal fusion and realizes the feature extraction of the valve-controlled hydraulic cylinder system under a long time span;

3. For the constructed valve-controlled hydraulic cylinder system, the present invention distinguishes normal data and fault data of the valve-controlled hydraulic cylinder system; at the same time, when the system fails, it can identify the specific components that have failed in the system, which will help the system in the event of a failure. Carry out repair work quickly in case of failure;

4. The present invention uses a deep learning algorithm to locate faults to specific components inside the system. On the one hand, it solves the problem of relying on manual experience to complete system fault detection, and realizes rapid diagnosis of system faults. On the other hand, it also uses the method of identifying specific faulty components , complete the maintenance and replacement of faulty parts, and reduce the maintenance cost of the system while ensuring the safe operation of the system.

Description of drawings

Other characteristics, objects and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

Fig. 1 is the principle diagram of the valve-controlled hydraulic cylinder system facing the independent control of the load port of the present invention;

Fig. 2 is the control block diagram of the valve-controlled hydraulic cylinder system facing the independent control of the load port of the present invention;

Fig. 3 is the flow chart of fault diagnosis of the system of the present invention;

Fig. 4 is the fault diagnosis algorithm diagram of the system of the present invention;

FIG. 5 is a confusion matrix diagram of the system fault diagnosis of the present invention.

detailed description

The present invention will be described in detail below in conjunction with specific embodiments. The following examples will help those skilled in the art to further understand the present invention, but do not limit the present invention in any form. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

Example 1:

According to a method for diagnosing faults of a valve-controlled hydraulic cylinder system oriented to independent control of a load port provided by the present invention, the method includes the following steps:

Step S1: establishing a valve-controlled hydraulic cylinder system;

Step S2: Perform fault injection to generate eight types of faults; faults include four high-speed switching valves, two hydraulic directional valves, hydraulic cylinders, and displacement sensors; the PWM driving voltage of the high-speed switching valves is provided by supporting software and hardware; hydraulic directional valves The most common faults are: oil particles and oil impact to the spool wear and leakage and seal ring damage and spring failure; the displacement sensor is installed on the valve body.

Step S3: Acquire signals, obtain nine types of signals, and generate samples; the signals are composed of nine types of faults and system normal states in step S2, and each type collects two main valve control chamber pressure signals and spool displacement signals , and collect the pressure signal of the hydraulic cylinder control chamber and the rod displacement signal, with a sampling frequency of 1000Hz for a total of 2s; normal data and fault data each contain 900 sets of data, each set of data includes 9 features, each feature includes 2000 data points, generating samples.

为预处理后结果：Step S4: Preprocess the sample; preprocess the sample by the following formula, where x _m represents the mth feature,

For the preprocessed result:

Step S5: After the preprocessing is completed, the generated samples are divided into training set, verification set and test set according to the ratio; the ratio is 6:2:1.

Step S6: the training set and the verification set are used for the training of the fault diagnosis algorithm;

Step S7: After the trained algorithm model is tested by the test set, the performance is judged.

Example 2:

Embodiment 2 is a preferred example of Embodiment 1 to describe the present invention more specifically.

Figure 1 and Figure 2 respectively show the schematic diagram and control block diagram of a valve-controlled hydraulic cylinder system for independent control of the load port involved in this proposal. The system uses a digital hydraulic pilot programmable valve to independently control the pressure of the two chambers of the hydraulic cylinder. The pilot stage of the digital hydraulic pilot programmable valve is composed of four identical two-position three-way slide valve high-speed switching valves, and the main stage is composed of two identical three-position three-way slide valve hydraulic reversing valves, and two A constant pressure source supplies oil to the main stage and the pilot stage respectively. In addition to the LVDT sensor sensing the displacement of the main spool and the displacement of the hydraulic cylinder rod, the system also installs a pressure sensor to sense the pressure of the main valve and the control chamber of the hydraulic cylinder. The control strategy adopts a two-stage PID feedback structure. Under the input reference displacement of the system, the load displacement driven by the oil cylinder can track the reference displacement as required. The first-stage PID1 is composed of controllers PID1-1 and PID1-2, and uses cylinder displacement as a feedback signal to generate reference displacements for the left main valve and right main valve spool. The second-stage PID2 is composed of controllers PID2-1, PID2-2, PID2-3 and PID2-4, which respectively use the left spool displacement and right spool displacement as feedback signals to generate and control the opening and closing characteristics of the four pilot valves. control signal.

The failure of the valve-controlled hydraulic cylinder system for independent control of the load port involves 8 components: four pilot-stage high-speed switching valves, two main-stage hydraulic reversing valves, hydraulic cylinders, and displacement sensors. The PWM driving voltage of the high-speed switching valve is provided by supporting software and hardware. On the one hand, the response time difference of the hardware end will cause the time lag of the control signal time of the high-speed switching valve; allow. At the same time, the high-frequency opening and closing characteristics of the high-speed switching valve repeatedly compress the spring, which may easily cause fatigue failure of the high-speed switching valve spring. The main failure of the hydraulic reversing valve is the wear of the valve core caused by the oil particles and the impact of the oil, which further leads to leakage and damage to the sealing ring, and causes fatigue failure of the spring. The displacement sensor is installed on the valve body, and the valve body will generate high-frequency vibration due to the high-speed opening and closing characteristics of the pilot stage high-speed switching valve, which will cause vibration of the displacement sensor probe installed in the system, thus affecting the feedback displacement of the displacement sensor accuracy. The hydraulic cylinder pushes the load to move as required. During the operation, the main failure is the hydraulic cylinder failure caused by the increase of leakage.

为预处理后结果。完成数据预处理后，将生成的样本按照6:2:1的比例划分训练集、验证集和测试集。训练集和验证集用于故障诊断算法的训练，训练完成的算法模型后经测试集测试后判断性能。Figure 3 shows the fault diagnosis process of the valve-controlled hydraulic cylinder system facing the independent control of the load port. After the establishment of the valve-controlled hydraulic cylinder system facing the independent control of the load port is completed, fault injection is performed to generate four high-speed switching valves and two hydraulic cylinders. There are 8 types of faults in the dynamic reversing valve, hydraulic cylinder and sensors. These 8 types of faults and the normal state of the system form a total of 9 categories. The fault diagnosis of the system is realized for these 9 categories. For each category, two main valve control chamber pressure signals and spool displacement signals are collected, and hydraulic cylinder control chamber pressure signals and rod displacement signals are collected, and the sampling frequency is 1000Hz for a total of 2s. Each category of normal data and fault data contains 900 sets of data, each set of data includes 9 features, and each feature includes 2000 data points. In order to avoid the impact of signals of different magnitudes on model training, the data is preprocessed by the following formula, where x _m represents the mth feature,

is the result after preprocessing. After the data preprocessing is completed, the generated samples are divided into training set, verification set and test set according to the ratio of 6:2:1. The training set and verification set are used for the training of the fault diagnosis algorithm, and the trained algorithm model is tested by the test set to judge its performance.

分别为指定批次样本t时刻的

的均值向量和方差向量，ε为避免分母为0的给定常数向量，γ和β分别为待训练的缩放和平移因子，

为输入

经过批归一化层的输出。通过引入残差结构使得算法可以通过增加深度提高准确率，并缓解了在网络中增加深度带来的梯度消失问题。6个残差块的输出经过平均池化层产生一个特征数固定的一维特征向量，该特征向量经过一个全连接层后得到特征数为9的特征向量，该向量通过式(3)所示的softmax函数得到概率分布，式中y_i表示第i个样本经过全连接层后的输出，y_ij表示输出y_i中第j个变量的取值，m表示分类器的分类数目，即为9。于是y_ij中最大的值对应的j即为第i个样本对应的类别。This proposal designs a fault diagnosis algorithm based on one-dimensional large convolution kernel and residual network for the fault diagnosis problem of the valve-controlled hydraulic cylinder system. The structure diagram of the algorithm is shown in Figure 4. First, the long-term input sequence of a valve-controlled hydraulic cylinder system is operated through a one-dimensional large convolution kernel, and a single large convolution kernel traverses the entire input time sequence by padding zeros and moving a certain step size, and then through the maximum pooling operation The result of the one-dimensional large convolution kernel convolution operation is processed and input to the residual block. In this algorithm, 6 residual blocks are designed, and each residual block uses the local information of the small convolution kernel to sense the sequence, and the batch normalization layer represented by formula (1) makes the output and input have distribution similarity To increase the training speed and stability of the network. At the same time, the ReLU of formula (2) is used as the activation function, and for each input variable x, the large value compared with 0 is output to speed up the convergence and solve the gradient problem. μ _B in formula (1) and

Respectively for the specified batch of samples at time t

The mean vector and variance vector of , ε is a given constant vector to avoid the denominator being 0, γ and β are the scaling and translation factors to be trained, respectively,

for input

The output of the batch normalization layer. By introducing the residual structure, the algorithm can improve the accuracy by increasing the depth, and alleviate the gradient disappearance problem caused by increasing the depth in the network. The output of the 6 residual blocks passes through the average pooling layer to generate a one-dimensional feature vector with a fixed number of features. After the feature vector passes through a fully connected layer, a feature vector with a feature number of 9 is obtained. This vector is shown in formula (3) The softmax function of the probability distribution is obtained, where y _i represents the output of the i-th sample after passing through the fully connected layer, y _ij represents the value of the j-th variable in the output y _i , and m represents the number of classifications of the classifier, which is 9 . Then the j corresponding to the largest value in y _ij is the category corresponding to the i-th sample.

relu(x)=max(0,x)#(2)

表示第i个样本的指示变量。对于一个M分类问题，

为包含M个元素的向量，样本的第j个值为1代表该样本的真实分类为第j类，其余值为0。

为该样本的预测类别概率值，也是一个包含M个元素的向量，其构成与

类似。For the algorithm shown in Figure 4, it is a multi-classification problem, and the cross-entropy loss function shown in formula (4) is used as the evaluation index of the network during training. where N is the sample size,

Denotes the indicator variable for the i-th sample. For an M classification problem,

is a vector containing M elements, the jth value of the sample is 1, which means that the true classification of the sample is the jth class, and the other values are 0.

is the predicted category probability value of the sample, which is also a vector containing M elements, and its composition is the same as

similar.

After the verification of the fault diagnosis process of the valve-controlled hydraulic cylinder system oriented to the independent control of the load port in Figure 3 and the corresponding fault diagnosis algorithm in Figure 4, the accuracy of the test set can be obtained as 95.4%, and the corresponding confusion matrix is shown in Figure 5, which can be seen It is found that the fault diagnosis algorithm proposed by the present invention can better distinguish the normal data and fault data of the valve-controlled hydraulic cylinder system. At the same time, when the system fails, identifying the specific component that fails in the system will help the system to quickly carry out maintenance work when the system fails.

1) A complete set of fault diagnosis process is proposed for the valve-controlled hydraulic cylinder system oriented to the independent control of the load port, and the fault of the system is located to the specific components inside the system, which should be protected.

2) The idea of one-dimensional large convolution kernel and residual network is applied to the fault diagnosis of the valve-controlled hydraulic cylinder system oriented to the independent control of the load port, and the fault diagnosis work is completed, which should be protected.

Those skilled in the art can understand this embodiment as a more specific description of Embodiment 1.

Those skilled in the art know that, in addition to realizing the system provided by the present invention and its various devices, modules, and units in a purely computer-readable program code mode, the system provided by the present invention and its various devices can be completely programmed by logically programming the method steps. , modules, and units implement the same functions in the form of logic gates, switches, ASICs, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by the present invention can be regarded as a hardware component, and the devices, modules, and units included in it for realizing various functions can also be regarded as hardware components. The structure; the device, module, and unit for realizing various functions can also be regarded as a software module for realizing the method or a structure in a hardware component.

Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art may make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. In the case of no conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims (8)

1. A fault diagnosis method for a valve-controlled hydraulic cylinder system independently controlled by a load port is characterized by comprising the following steps:

step S1: establishing a valve control hydraulic cylinder system;

step S2: fault injection is carried out to generate eight types of faults;

and step S3: acquiring signals to obtain nine types of signals and generating a sample;

and step S4: preprocessing a sample;

step S5: after the pretreatment is finished, dividing the generated sample into a training set, a verification set and a test set according to a proportion;

step S6: the training set and the verification set are used for training a fault diagnosis algorithm;

step S7: and judging the performance of the trained algorithm model after the test of the test set.

2. The method for diagnosing the fault of the valve-controlled hydraulic cylinder system facing the load port independent control according to claim 1, wherein the fault in the step S2 comprises four high-speed switching valves, two hydraulic reversing valves, a hydraulic cylinder and a displacement sensor.

3. The method for diagnosing the fault of the valve-controlled hydraulic cylinder system facing the load port independent control according to claim 2, wherein the PWM driving voltage of the high-speed switch valve is provided by matched software and hardware.

4. The method for diagnosing the fault of the valve-controlled hydraulic cylinder system facing the load port independent control according to claim 2, wherein the fault of the hydraulic reversing valve is as follows: oil particles and oil impact wear the valve element and leak out with seal ring damage and failure of the spring.

5. The method of claim 2, wherein the displacement sensor is mounted on the valve body.

6. The method for diagnosing the fault of the valve-controlled hydraulic cylinder system facing to the load port independent control, according to claim 1, wherein the signals in the step S3 are composed of the fault in the step S2 and the normal state of the system into nine categories, each category is used for collecting two main valve control cavity pressure signals and two spool displacement signals, and collecting a hydraulic cylinder control cavity pressure signal and a rod displacement signal, and sampling is carried out for 2S at a sampling frequency of 1000 Hz; the normal data and fault data each contained 900 sets of data, each set of data contained 9 signatures, each signature consisting of 2000 data points, resulting in a sample.

7. The method for diagnosing the fault of the valve-controlled hydraulic cylinder system facing the independent control of the load port according to claim 1, wherein the step S4 is to preprocess the sample by the following formula, wherein x is _m It is shown that the m-th feature,

as a result after pretreatment:

8. the method for diagnosing the fault of the valve-controlled hydraulic cylinder system facing the load port independent control according to claim 1, wherein the ratio in the step S5 is 6:2:1.

Images