CN114576236A - State monitoring and fault diagnosis method and device for hydraulic system - Google Patents

State monitoring and fault diagnosis method and device for hydraulic system Download PDF

Info

Publication number
CN114576236A
CN114576236A CN202011378332.5A CN202011378332A CN114576236A CN 114576236 A CN114576236 A CN 114576236A CN 202011378332 A CN202011378332 A CN 202011378332A CN 114576236 A CN114576236 A CN 114576236A
Authority
CN
China
Prior art keywords
fault
hydraulic system
hydraulic
data
state
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202011378332.5A
Other languages
Chinese (zh)
Inventor
刘景轩
陈志军
苏晓岩
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Machinery Equipment Research Institute
Original Assignee
Beijing Machinery Equipment Research Institute
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Machinery Equipment Research Institute filed Critical Beijing Machinery Equipment Research Institute
Priority to CN202011378332.5A priority Critical patent/CN114576236A/en
Publication of CN114576236A publication Critical patent/CN114576236A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F15—FLUID-PRESSURE ACTUATORS; HYDRAULICS OR PNEUMATICS IN GENERAL
    • F15B—SYSTEMS ACTING BY MEANS OF FLUIDS IN GENERAL; FLUID-PRESSURE ACTUATORS, e.g. SERVOMOTORS; DETAILS OF FLUID-PRESSURE SYSTEMS, NOT OTHERWISE PROVIDED FOR
    • F15B19/00—Testing; Calibrating; Fault detection or monitoring; Simulation or modelling of fluid-pressure systems or apparatus not otherwise provided for
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F15—FLUID-PRESSURE ACTUATORS; HYDRAULICS OR PNEUMATICS IN GENERAL
    • F15B—SYSTEMS ACTING BY MEANS OF FLUIDS IN GENERAL; FLUID-PRESSURE ACTUATORS, e.g. SERVOMOTORS; DETAILS OF FLUID-PRESSURE SYSTEMS, NOT OTHERWISE PROVIDED FOR
    • F15B19/00—Testing; Calibrating; Fault detection or monitoring; Simulation or modelling of fluid-pressure systems or apparatus not otherwise provided for
    • F15B19/005—Fault detection or monitoring
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F15—FLUID-PRESSURE ACTUATORS; HYDRAULICS OR PNEUMATICS IN GENERAL
    • F15B—SYSTEMS ACTING BY MEANS OF FLUIDS IN GENERAL; FLUID-PRESSURE ACTUATORS, e.g. SERVOMOTORS; DETAILS OF FLUID-PRESSURE SYSTEMS, NOT OTHERWISE PROVIDED FOR
    • F15B2211/00—Circuits for servomotor systems
    • F15B2211/80—Other types of control related to particular problems or conditions
    • F15B2211/857—Monitoring of fluid pressure systems
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F15—FLUID-PRESSURE ACTUATORS; HYDRAULICS OR PNEUMATICS IN GENERAL
    • F15B—SYSTEMS ACTING BY MEANS OF FLUIDS IN GENERAL; FLUID-PRESSURE ACTUATORS, e.g. SERVOMOTORS; DETAILS OF FLUID-PRESSURE SYSTEMS, NOT OTHERWISE PROVIDED FOR
    • F15B2211/00—Circuits for servomotor systems
    • F15B2211/80—Other types of control related to particular problems or conditions
    • F15B2211/87—Detection of failures

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Fluid Mechanics (AREA)
  • Mechanical Engineering (AREA)
  • General Engineering & Computer Science (AREA)
  • Testing And Monitoring For Control Systems (AREA)
  • Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
  • Fluid-Pressure Circuits (AREA)

Abstract

The invention relates to a state monitoring and fault diagnosis method for a hydraulic system, belongs to the technical field of hydraulic systems, and solves the problems of difficulty in monitoring and inaccurate fault prediction in the prior art. The method comprises the following steps: arranging a multi-information sensor in a hydraulic system, and monitoring the running state of a corresponding part in real time; when any one operation state changes, extracting characteristic information corresponding to the state, inputting the characteristic information into a deep neural network trained in advance, and obtaining the probability of faults of all parts of the hydraulic system; and (3) carrying out normalization processing on the probability of each part with faults, inputting the normalization result into a D-S evidence decision model trained in advance, obtaining the part with the most possible faults and the fault grade, and pushing the part with the most possible faults and the fault grade to an operator. The invention realizes the automatic diagnosis of the fault of the hydraulic system and the positioning function thereof.

Description

State monitoring and fault diagnosis method and device for hydraulic system
Technical Field
The invention relates to the technical field of hydraulic systems, in particular to a method and a device for state monitoring and fault diagnosis of a hydraulic system.
Background
At present, state monitoring and fault diagnosis technologies are increasingly applied to mechanical equipment, and the safe and reliable operation of the equipment is greatly guaranteed. The structure and principle of hydraulic systems are generally complex. When a hydraulic system fails, the fault location of the prior art is related to the professional level and experience of an operator, and measurement needs to be carried out by means of traditional instrument equipment during the fault elimination process.
The damage of any component can change the working state of the whole hydraulic system, and before the fault occurs, because the system state is not monitored in time, the factors which possibly cause the fault cannot be timely and accurately positioned and removed. The traditional instrument equipment measurement needs to consume a large amount of manpower, material resources and time.
Therefore, it is necessary to monitor the operating state of the hydraulic system on-line, and to perform maintenance and fault diagnosis in time when the operating state of the hydraulic system changes. In the prior art, a method for performing state detection and fault diagnosis on a hydraulic system by using a neural network is adopted, but problems of misclassification or uncertain classification and the like are easily caused due to selection of training samples and generalization of network structures, so that user experience is poor.
Disclosure of Invention
In view of the foregoing analysis, an embodiment of the present invention is directed to a method for monitoring a state and diagnosing a fault of a hydraulic system, so as to solve the problems of difficulty in monitoring and inaccurate fault prediction in the prior art.
In one aspect, an embodiment of the present invention provides a method for monitoring a state and diagnosing a fault of a hydraulic system, including the following steps:
arranging a multi-information sensor in a hydraulic system, and monitoring the running state of a corresponding part in real time;
when any one operation state changes, extracting characteristic information corresponding to the state, inputting the characteristic information into a deep neural network trained in advance, and obtaining the probability of faults of all parts of the hydraulic system;
and (3) carrying out normalization processing on the probability of each part with faults, inputting the normalization result into a D-S evidence decision model trained in advance, obtaining the part with the most possible faults and the fault grade, and pushing the part with the most possible faults and the fault grade to an operator.
The beneficial effects of the above technical scheme are as follows: the novel method for the online monitoring and fault diagnosis of the hydraulic system is provided, so that an accurate fault judgment conclusion can be automatically given when non-professional personnel operate the method, and sufficient information can be provided when fault elimination is carried out. By the method, potential problems can be found in time in the maintenance stage of the product, and the equipment is tracked in a full state in the product research and development stage, so that the product can be updated and iterated quickly. The D-S evidence theory is introduced, uncertainty caused by randomness can be well processed, the true range can be narrowed according to continuous accumulation of evidence, and uncertain and unknown states of events can be well processed and distinguished.
In a further improvement of the above method, the multiple information sensor includes at least one of a pressure sensor, a flow sensor, a vibration sensor, and a temperature sensor.
The beneficial effects of the above further improved scheme are: the type of the fusion information of the multi-information sensor is limited, and information such as vibration, pressure, flow, temperature and the like is obtained. For a hydraulic system, pressure and flow respectively correspond to output force (torque) and running speed of an actuating mechanism, vibration corresponds to running states of a key movement mechanism such as a hydraulic pump and a friction pair, and temperature corresponds to running efficiency of the hydraulic system.
Further, the laying of multiple information sensors in the hydraulic system, real-time monitoring of the operating state of the corresponding parts, further includes:
pressure sensors are distributed at the oil outlet of the hydraulic pump, the positive and negative cavities of the hydraulic cylinder, the reversing valve and the speed regulating valve;
a flow sensor is arranged at an oil outlet of the hydraulic pump;
distributing vibration sensors on the surfaces of a hydraulic pump, a motor and a hydraulic mechanical locking cylinder;
temperature sensors are distributed on the oil outlet of the hydraulic pump, the oil inlet of the hydraulic cylinder and the surface of the oil tank;
monitoring whether the running state of the hydraulic pump changes or not through vibration data acquired by a vibration sensor, and judging that the running state of the hydraulic pump changes when the change amount of the vibration data at the current moment and the last moment exceeds a threshold value;
monitoring whether the temperature states of the hydraulic pump, the hydraulic cylinder and the oil tank change or not through temperature data acquired by the temperature sensor, and judging that the temperature states of the corresponding hydraulic pump, the corresponding hydraulic cylinder and/or the corresponding oil tank change when the change amount of any one of the temperature data at the current moment and the temperature data at the previous moment exceeds a second threshold value;
monitoring whether the output torque of the actuating mechanism changes or not according to the pressure data acquired by the pressure sensor; when the pressure data change amount at the current moment and the previous moment exceeds a threshold value three, judging that the output torque of the actuating mechanism changes;
monitoring whether the running speed of the actuating mechanism changes or not through flow data acquired by a flow sensor; and when the flow data change amount at the current time and the last time exceeds the threshold value four times, judging that the running speed of the actuating mechanism is changed.
The beneficial effects of the above further improved scheme are: and the position of the hydraulic system which changes is pre-judged, and an operator is assisted to master the running state of the component in time. The real state of the interior of the hydraulic system is reflected in real time by integrating four physical quantities of pressure, flow, vibration and temperature, and conditions are provided for integrating various information to give the system state or fault level.
Further, when any one operation state changes, extracting characteristic information corresponding to the state, inputting the characteristic information into a deep neural network trained in advance, and acquiring the probability of faults of all parts of the hydraulic system, further comprising:
when any one operation state is changed, performing wavelet transformation and Kalman filtering on data output by a corresponding sensor in sequence;
and extracting the amplitude and the phase of the obtained filtering result to be used as characteristic information corresponding to the running state, inputting the characteristic information into a corresponding input end of the deep neural network trained in advance, and inputting standard values into other input ends to obtain the fault probability of each part of the hydraulic system.
The beneficial effects of the above further improved scheme are: only the information that the operating state has changed is useful for the fault determination, and therefore the above-described scheme extracts the characteristic information of the signal that the operating state has changed. Irrelevant information can be filtered through filtering, useful characteristic information is obtained, and then more accurate probability of faults of all parts of the hydraulic system can be obtained, and a foundation is laid for follow-up accurate determination of fault parts.
Further, the normalizing the probability of the failure of each part, and inputting the normalized result into a D-S evidence decision model to obtain the most likely failed part and the failure level, further includes:
the probability y of the fault of each part of the hydraulic system is calculated by the following formulaiNormalization processing is carried out to obtain a normalized vector yi′}，i＝1…N
Figure BDA0002808710360000041
In the formula, yiThe ith value of the output vector of the deep neural network represents the fault probability of the ith part of the hydraulic system, and N is the dimension of the output vector;
taking the normalization result as a first evidence, taking the set fault state distribution vector as a second evidence, inputting the second evidence into a D-S evidence decision model in the following formula, and obtaining decision confidence m (A) after the two evidences are fusedi)
Figure BDA0002808710360000042
Figure BDA0002808710360000051
Figure BDA0002808710360000052
In the formula, AiIndicating a fault event at the ith location of the hydraulic system, AjIndicating a fault event at the jth location of the hydraulic system, biIndicates the ith value, bel, in the assigned fault state vectoriTo representConfidence coefficient obtained by training, i is 1 … N, j is 1 … i-1i +1N, i is not equal to j, epsilon1、ε2For a preset threshold value, max { } is a maximum function;
identification { m (A)i) The position of the hydraulic system corresponding to the maximum value is used as a final fault position;
identification { m (A)i) The sensor output data x corresponding to the maximum valueiAccording to said xiAnd comparing with a preset threshold value to obtain the fault grade of the fault part.
The beneficial effects of the above further improved scheme are: the neural network is combined with the D-S evidence theory, so that the occurrence of misjudgment or an ill-conditioned condition is eliminated, and the accuracy of diagnosing a fault part is improved. beliObtained by prior training.
In another aspect, an embodiment of the present invention provides a state monitoring and fault diagnosing apparatus for a hydraulic system, including:
the multi-information sensor is used for monitoring the running state of each part of the hydraulic system in real time and sending the running state to the preprocessing module;
the preprocessing module is used for judging whether the operation states of all parts of the hydraulic system change or not, extracting the characteristic information corresponding to any one operation state when the operation state changes, inputting the characteristic information into a pre-trained deep neural network to obtain the probability of faults of all parts of the hydraulic system, and transmitting the probability to the fault diagnosis module;
and the fault diagnosis module is used for carrying out normalization processing on the probability of the faults of all the parts, inputting the normalization result into a D-S evidence decision model trained in advance, obtaining the part with the most possible fault and the fault grade, and pushing the part with the most possible fault and the fault grade to an operator.
The beneficial effects of the above technical scheme are as follows: the novel method for the online monitoring and fault diagnosis of the hydraulic system is provided, so that an accurate fault judgment conclusion can be automatically given when non-professional personnel operate the method, and sufficient information can be provided when fault elimination is carried out. By the method, potential problems can be found in time in the maintenance stage of the product, and the equipment is tracked in a full state in the product research and development stage, so that the product can be updated and iterated quickly. The D-S evidence theory is introduced, uncertainty caused by randomness can be well processed, the true range can be narrowed according to continuous accumulation of evidence, and uncertain and unknown states of events can be well processed and distinguished.
Based on the further improvement of the device, the multi-information sensor comprises:
the pressure sensors are arranged at the oil outlet of the hydraulic pump, the positive and negative cavities of the hydraulic cylinder, the reversing valve and the speed regulating valve position and are used for acquiring pressure data of the arranged position and monitoring the output torque of the actuating mechanism; and transmitting the pressure data to a preprocessing module;
the flow sensor is arranged at an oil outlet of the hydraulic pump and used for acquiring flow data of the arrangement position and monitoring the running speed of the actuating mechanism; and transmitting the flow data to a preprocessing module;
the vibration sensor is arranged on the surfaces of the hydraulic pump, the motor and the hydraulic mechanical locking cylinder and used for acquiring vibration data of the arranged position and monitoring the running state of the hydraulic pump; and transmitting the vibration data to a preprocessing module;
the temperature sensors are arranged at the oil outlet of the hydraulic pump, the oil inlet of the hydraulic cylinder and the surface of the oil tank and are used for acquiring temperature data of the arranged positions and monitoring the temperature states of the hydraulic pump, the hydraulic cylinder and the oil tank; and transmitting the temperature data to a preprocessing module.
The beneficial effect of adopting the above further improved scheme is: the type of the fusion information of the multi-information sensor is limited, and information such as vibration, pressure, flow, temperature and the like is obtained. For a hydraulic system, pressure and flow respectively correspond to output force (torque) and running speed of an actuating mechanism, vibration corresponds to running states of a key movement mechanism such as a hydraulic pump and a friction pair, and temperature corresponds to running efficiency of the hydraulic system.
Further, the preprocessing module executes the following program to judge whether the operation state of each part of the hydraulic system changes:
obtaining the variation of the vibration data at the current moment and the last moment; comparing the variation of the vibration data with a threshold value to judge whether the running state of the hydraulic pump changes, and when the variation exceeds the threshold value, judging that the running state of the hydraulic pump changes, otherwise, no variation exists;
obtaining the variation of the temperature data at the current moment and the previous moment, comparing the variation of the temperature data with a second threshold value to judge whether the temperature states of the hydraulic pump, the hydraulic cylinder and the oil tank change or not, and judging that the temperature states of the corresponding hydraulic pump, the corresponding hydraulic cylinder and/or the corresponding oil tank change when any variation exceeds the second threshold value, or else, no variation exists;
obtaining the change quantity of the pressure data at the current moment and the last moment; comparing the change amount of the pressure data with a third threshold value to judge whether the output torque of the actuating mechanism changes, and when the change amount exceeds the third threshold value, judging that the output torque of the actuating mechanism changes, otherwise, no change occurs;
obtaining the change quantity of the flow data at the current moment and the last moment; and comparing the change amount of the flow data with a threshold value four to judge whether the running speed of the execution mechanism changes, and when the change amount exceeds the threshold value four, judging that the running speed of the execution mechanism changes, otherwise, judging that no change exists.
The beneficial effect of adopting the above further improved scheme is: and prejudging the changed components of the hydraulic system, and assisting an operator to master the running states of the components in time. The real state of the interior of the hydraulic system is reflected in real time by integrating four physical quantities of pressure, flow, vibration and temperature, and conditions are provided for integrating various information to give the system state or fault level.
Further, the preprocessing module executes the following program to obtain the probability of the fault of each part of the hydraulic system:
when any one operation state is changed, performing wavelet transformation and Kalman filtering on data output by a corresponding sensor in sequence;
and extracting the amplitude and the phase of the obtained filtering result to be used as characteristic information corresponding to the running state, inputting the characteristic information into a corresponding input end of the deep neural network trained in advance, and inputting standard values into other input ends to obtain the fault probability of each part of the hydraulic system.
The beneficial effect of adopting the above further improved scheme is: only the information that the operating state has changed is useful for the fault determination, and therefore the above-described scheme extracts the characteristic information of the signal that the operating state has changed. Irrelevant information can be filtered through filtering, useful characteristic information is obtained, and then more accurate probability of faults of all parts of the hydraulic system can be obtained, and a foundation is laid for follow-up accurate determination of fault parts.
Further, the fault diagnosis module executes the following procedures to obtain the most probable fault position and the fault grade:
the probability y of the fault of each part of the hydraulic system is calculated by the following formulaiNormalization processing is carried out to obtain a normalized vector yi′}，i＝1…N
Figure BDA0002808710360000081
In the formula, yiThe ith value of the output vector of the deep neural network represents the fault probability of the ith part of the hydraulic system, and N is the dimension of the output vector;
taking the normalization result as a first evidence, taking the set fault state distribution vector as a second evidence, inputting the second evidence into a D-S evidence decision model in the following formula, and obtaining decision confidence m (A) after the two evidences are fusedi)
Figure BDA0002808710360000082
Figure BDA0002808710360000083
Figure BDA0002808710360000084
In the formula, AiIndicating a fault event at the ith location of the hydraulic system, AjIndicating a fault event at the jth location of the hydraulic system, biIndicates the ith value, bel, in the assigned fault state vectoriDenotes the confidence coefficient, i ≠ j, i ═ 1 … N, j ═ 1 … i-1i +1N, epsilon1、ε2For a preset threshold value, max { } is a maximum function;
identify { m (A)i) The position of the hydraulic system corresponding to the maximum value is used as a final fault position;
identification { m (A)i) The sensor output data x corresponding to the maximum valueiAccording to said xiAnd comparing with a preset threshold value to obtain the fault grade of the fault part.
The beneficial effect of adopting the above further improved scheme is: the neural network is combined with the D-S evidence theory, so that the occurrence of misjudgment or an ill-conditioned condition is eliminated, and the accuracy of diagnosing a fault part is improved. beliObtained by prior training.
In the invention, the technical schemes can be combined with each other to realize more preferable combination schemes. Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and drawings.
Drawings
The drawings are only for purposes of illustrating particular embodiments and are not to be construed as limiting the invention, wherein like reference numerals are used to designate like parts throughout.
Fig. 1 is a flow chart of a method for monitoring the state and diagnosing faults of a hydraulic system according to embodiment 1 of the present invention;
FIG. 2 is a schematic diagram of a hydraulic system;
fig. 3 is a schematic view of a state monitoring and fault diagnosis device of a hydraulic system according to embodiment 3 of the present invention;
fig. 4 is a schematic diagram of fault diagnosis of a certain hydraulic system obtained by the device in embodiment 4 of the invention.
Reference numerals:
1-an oil tank; 2-a motor; 3-a hydraulic pump; 4-a reversing valve; 5-one-way speed regulating valve; 6-hydraulic cylinder; 7-proportional overflow valve.
Detailed Description
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate preferred embodiments of the invention and together with the description, serve to explain the principles of the invention and not to limit the scope of the invention.
Example 1
A specific embodiment of the present invention discloses a method for monitoring a state and diagnosing a fault of a hydraulic system, as shown in fig. 1, including the following steps:
s1, arranging a multi-information sensor in a hydraulic system, and monitoring the running state of a corresponding part in real time;
s2, when any one operation state changes, extracting characteristic information corresponding to the state, inputting the characteristic information into a deep neural network trained in advance, and obtaining the probability of faults of all parts of the hydraulic system;
and S3, carrying out normalization processing on the probability of the fault of each part, inputting the normalization result into a D-S evidence decision model trained in advance, obtaining the part most likely to have the fault and the fault grade, and pushing the part and the fault grade to an operator.
In practice, the method of steps S1-S3 can monitor and detect the key elements (parts) of the hydraulic system.
Compared with the prior art, the method provided by the embodiment can automatically provide an accurate fault judgment conclusion when being operated by non-professional personnel, and can provide sufficient information when troubleshooting is performed. By the method, potential problems can be found in time in the maintenance stage of the product, and the equipment is tracked in a full state in the product research and development stage, so that the product can be updated and iterated quickly.
Example 2
The optimization is performed on the basis of the method in embodiment 1, and the multi-information sensor in step S1 includes at least one of a pressure sensor, a flow sensor, a vibration sensor, and a temperature sensor.
The hydraulic system comprises a hydraulic pump 3, a reversing valve 4, a (one-way) speed regulating valve 5, a hydraulic cylinder 6 and a proportional overflow valve 7, and the specific structure is shown in figure 2.
Preferably, the step S1 further includes:
s11, distributing pressure sensors at the oil outlet of the hydraulic pump 3, the positive and negative cavities of the hydraulic cylinder 6, the reversing valve 4, the speed regulating valve 5 and the like;
s12, distributing a flow sensor at an oil outlet of the hydraulic pump 3;
s13, distributing vibration sensors at key parts such as the surfaces of the hydraulic pump 3, the motor 2 and the hydraulic mechanical locking cylinder or friction pair parts;
s14, distributing temperature sensors on an oil outlet of the hydraulic pump 3, an oil inlet of the hydraulic cylinder 6 and the surface of the oil tank 1; the current popular sensors are sensors that integrate pressure, flow and temperature;
s15, monitoring whether the running state of the hydraulic pump 3 changes or not through vibration data acquired by a vibration sensor, and judging that the running state of the hydraulic pump 3 changes when the vibration data change amount of the current moment and the last moment exceeds a threshold value;
s16, monitoring whether the temperature states of the hydraulic pump 3, the hydraulic cylinder 6 and the oil tank change or not through temperature data acquired by the temperature sensors, judging that the temperature states of the corresponding hydraulic pump 3, the corresponding hydraulic cylinder 6 and/or the corresponding oil tank change when the change amount of any temperature data at the current moment and the previous moment exceeds a second threshold value, and then detecting parameters such as pressure, flow and the like of a hydraulic system;
s17, monitoring whether the output torque of the actuating mechanism changes or not through the pressure data acquired by the pressure sensor; when the pressure data change amount at the current moment and the previous moment exceeds a threshold value three, judging that the output torque of the actuating mechanism (the oil tank 1 and the motor 2) is changed;
s18, monitoring whether the running speed of the executing mechanism changes or not through flow data acquired by a flow sensor; and when the flow data change amount at the current time and the last time exceeds the threshold value four times, judging that the running speed of the actuating mechanism is changed.
Preferably, the step S2 further includes:
s21, when any one operation state changes, performing wavelet transformation and Kalman filtering on data output by a corresponding sensor in sequence;
and S22, extracting the amplitude and the phase of the obtained filtering result, using the amplitude and the phase as characteristic information corresponding to the running state, inputting the characteristic information into a corresponding input end of the deep neural network trained in advance, inputting standard values (normal working state values, and avoiding misjudgment caused by information redundancy) into other input ends, and obtaining the probability of faults of all parts of the hydraulic system.
Preferably, the step S3 further includes:
s31, the probability y of failure of each part of the hydraulic system is determined through the following formulaiNormalization processing is carried out to obtain a normalized vector yi′}，i＝1…N
Figure BDA0002808710360000121
In the formula, yiThe ith value of the output vector of the deep neural network represents the fault probability of the ith part of the hydraulic system, and N is the dimension of the output vector;
s32, taking the normalization result as a first evidence, and distributing a set fault state to vectors (obtained by a user according to a test, wherein the number of elements of the vectors is equal to that of the normalization result, and the element b isiRepresenting the typical state of the fault of the ith part of the hydraulic system) as a second evidence, and inputting the second evidence into a D-S evidence decision model in the following formula to obtain decision confidence m (A) after the two evidences are fusedi)
Figure BDA0002808710360000131
Figure BDA0002808710360000132
Figure BDA0002808710360000133
In the formula, AiIndicating a fault event at the ith location of the hydraulic system, AjIndicating a fault event at the jth location of the hydraulic system, biIndicates the ith value, bel, in the assigned fault state vectoriDenotes the confidence coefficient, i ≠ j, i ═ 1 … N, j ═ 1 … i-1i +1N, epsilon1、ε2For a preset threshold value, max { } is a maximum function;
s33, identifying m (A)i) The position of the hydraulic system corresponding to the maximum value is used as a final fault position;
s34, identifying m (A)i) The sensor output data x corresponding to the maximum valueiAccording to said xiAnd comparing with a preset threshold value to obtain the fault grade of the fault part.
Obtaining a trained deep neural network and a D-S evidence decision model through the following steps:
s01, acquiring training data covering all fault types of a hydraulic system, wherein input data in the training data are sensor data of the hydraulic system with faults, and output data are manually marked positions (codes) with the faults actually;
s02, for each group of training data, obtaining corresponding sensor data with the current time and the last time of the change of the operation state;
s03, extracting characteristic information corresponding to the operation state from the sensor data, inputting the characteristic information into a deep neural network, and obtaining the probability of faults of all parts of the hydraulic system;
s04, normalizing the probability of the fault of each part, and inputting a normalization result into a D-S evidence decision model to obtain the part most likely to have the fault;
and S05, judging whether the most possibly-failed part is consistent with the manually-labeled actually-failed part, if not, adjusting parameters in the deep neural network and the D-S evidence decision model, training again until the two parts are consistent, finishing the training of the group of training data, and performing the training of the next group of training data. The adjustment can be performed using a gradient descent method, as will be appreciated by those skilled in the art.
Example 3
The invention also discloses a state monitoring and fault diagnosis device of the hydraulic system corresponding to the method of the embodiment 1, which comprises a multi-information sensor, a preprocessing module and a fault diagnosis module which are connected in sequence, as shown in figure 3.
And the multi-information sensor is used for monitoring the running state of each part of the hydraulic system in real time and sending the running state to the preprocessing module.
The preprocessing module is used for judging whether the operation states of all parts of the hydraulic system change or not, extracting the characteristic information corresponding to the states when any one operation state changes, inputting the characteristic information into a pre-trained deep neural network, obtaining the probability of faults of all parts of the hydraulic system, and transmitting the probability to the fault diagnosis module.
And the fault diagnosis module is used for carrying out normalization processing on the probability of the faults of all the parts, inputting the normalization result into a D-S evidence decision model trained in advance, obtaining the part with the most possible fault and the fault grade, and pushing the part with the most possible fault and the fault grade to an operator.
Example 4
The optimization is carried out on the basis of the device in the embodiment 3, the invention also discloses a device corresponding to the method in the embodiment 2, and the multi-information sensor comprises a pressure sensor, a flow sensor, a vibration sensor and a temperature sensor.
The pressure sensors are arranged at the oil outlet of the hydraulic pump, the positive and negative cavities of the hydraulic cylinder, the reversing valve and the speed regulating valve position and are used for acquiring pressure data of the arranged position and monitoring the output torque of the actuating mechanism; and transmitting the pressure data to a preprocessing module.
The flow sensor is arranged at an oil outlet of the hydraulic pump and used for acquiring flow data of the arrangement position and monitoring the running speed of the actuating mechanism; and transmitting the flow data to a preprocessing module.
The vibration sensor is arranged on the surfaces of the hydraulic pump, the motor and the hydraulic mechanical locking cylinder and used for acquiring vibration data of the arranged position and monitoring the running state of the hydraulic pump; and transmitting the vibration data to a preprocessing module.
The temperature sensors are arranged at the oil outlet of the hydraulic pump, the oil inlet of the hydraulic cylinder and the surface of the oil tank and are used for acquiring temperature data of the arranged positions and monitoring the temperature states of the hydraulic pump, the hydraulic cylinder and the oil tank; and transmitting the temperature data to a preprocessing module.
Preferably, the preprocessing module executes the following program to judge whether the operation state of each part of the hydraulic system changes:
SS1, obtaining the change quantity of the vibration data at the current moment and the last moment; comparing the variation of the vibration data with a threshold value to judge whether the running state of the hydraulic pump changes, and when the variation exceeds the threshold value, judging that the running state of the hydraulic pump changes, otherwise, no variation exists;
SS2, obtaining the variation of the temperature data at the current moment and the previous moment, comparing the variation of the temperature data with a second threshold value to judge whether the temperature states of the hydraulic pump, the hydraulic cylinder and the oil tank are changed, when any variation exceeds the second threshold value, judging that the temperature states of the corresponding hydraulic pump, the corresponding hydraulic cylinder and/or the corresponding oil tank are changed, otherwise, no variation exists;
SS3, obtaining the change quantity of the pressure data at the current moment and the last moment; comparing the change amount of the pressure data with a third threshold value to judge whether the output torque of the actuating mechanism changes, and when the change amount exceeds the third threshold value, judging that the output torque of the actuating mechanism changes, otherwise, no change occurs;
SS4, obtaining the change quantity of the flow data at the current moment and the last moment; and comparing the change amount of the flow data with a threshold value four to judge whether the running speed of the execution mechanism changes, and when the change amount exceeds the threshold value four, judging that the running speed of the execution mechanism changes, otherwise, judging that no change exists.
Preferably, the preprocessing module executes the following program to obtain the probability of faults of all parts of the hydraulic system:
SS5, when any one operation state changes, the data output by the corresponding sensor is subjected to wavelet transformation and Kalman filtering in sequence;
and SS6, extracting the amplitude and the phase of the obtained filtering result, using the amplitude and the phase as characteristic information corresponding to the running state, inputting the characteristic information into a corresponding input end of the deep neural network trained in advance, and inputting standard values into other input ends to obtain the probability of faults of all parts of the hydraulic system.
Preferably, the fault diagnosis module executes the following procedures to obtain the most probable fault position and fault grade:
SS7. probability y of failure of each part of the hydraulic system by the following formulaiNormalization processing is carried out to obtain a normalized vector yi′}，i＝1…N
Figure BDA0002808710360000161
In the formula, yiThe ith value of the output vector of the deep neural network represents the fault probability of the ith part of the hydraulic system, and N is the dimension of the output vector;
SS8, taking the normalization result as a first evidence, taking the set fault state distribution vector as a second evidence, inputting the second evidence into a D-S evidence decision model in the following formula, and obtaining decision confidence m (A) after the two evidences are fusedi)
Figure BDA0002808710360000162
Figure BDA0002808710360000163
Figure BDA0002808710360000171
In the formula, AiIndicating a fault event at the ith location of the hydraulic system, AjIndicating a fault event at the jth location of the hydraulic system, biIndicates the ith value, bel, in the assigned fault state vectoriDenotes the confidence coefficient, i ≠ j, i ═ 1 … N, j ═ 1 … i-1i +1N, epsilon1、ε2For a preset threshold value, max { } is a maximum function;
SS9. identification { m (A)i) The position of the hydraulic system corresponding to the maximum value is used as a final fault position;
SS10. identification { m (A)i) The sensor output data x corresponding to the maximum valueiAccording to said xiAnd comparing with a preset threshold value to obtain the fault grade of the fault part.
For example, as shown in fig. 4, for a hydraulic pump failure, the oil leakage may be set to be severe at level 1 and the flow rate may be set to be insufficient at level 2.
Those skilled in the art will appreciate that all or part of the flow of the method implementing the above embodiments may be implemented by a computer program, which is stored in a computer readable storage medium, to instruct related hardware. The computer readable storage medium is a magnetic disk, an optical disk, a read-only memory or a random access memory, etc.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.

Claims (10)

1. A method for monitoring the state and diagnosing the fault of a hydraulic system is characterized by comprising the following steps:
arranging a multi-information sensor in a hydraulic system, and monitoring the running state of a corresponding part in real time;
when any one operation state changes, extracting characteristic information corresponding to the state, inputting the characteristic information into a deep neural network trained in advance, and obtaining the probability of faults of all parts of the hydraulic system;
and (3) carrying out normalization processing on the probability of each part with faults, inputting the normalization result into a D-S evidence decision model trained in advance, obtaining the part with the most possible faults and the fault grade, and pushing the part with the most possible faults and the fault grade to an operator.
2. The method of claim 1, wherein the multi-information sensor comprises at least one of a pressure sensor, a flow sensor, a vibration sensor, and a temperature sensor.
3. The method for monitoring the state and diagnosing the fault of the hydraulic system according to claim 2, wherein a multi-information sensor is arranged in the hydraulic system to monitor the operation state of the corresponding part in real time, and further comprising:
pressure sensors are distributed at the oil outlet of the hydraulic pump, the positive and negative cavities of the hydraulic cylinder, the reversing valve and the speed regulating valve;
a flow sensor is arranged at an oil outlet of the hydraulic pump;
distributing vibration sensors on the surfaces of a hydraulic pump, a motor and a hydraulic mechanical locking cylinder;
temperature sensors are distributed on the oil outlet of the hydraulic pump, the oil inlet of the hydraulic cylinder and the surface of the oil tank;
monitoring whether the running state of the hydraulic pump changes or not through vibration data acquired by a vibration sensor, and judging that the running state of the hydraulic pump changes when the change amount of the vibration data at the current moment and the last moment exceeds a threshold value;
monitoring whether the temperature states of the hydraulic pump, the hydraulic cylinder and the oil tank change or not through temperature data acquired by the temperature sensor, and judging that the temperature states of the corresponding hydraulic pump, the corresponding hydraulic cylinder and/or the corresponding oil tank change when the change amount of any one of the temperature data at the current moment and the temperature data at the previous moment exceeds a second threshold value;
monitoring whether the output torque of the actuating mechanism changes or not according to the pressure data acquired by the pressure sensor; when the pressure data change amount at the current moment and the previous moment exceeds a threshold value three, judging that the output torque of the actuating mechanism changes;
monitoring whether the running speed of an actuating mechanism is changed or not through flow data acquired by a flow sensor; and when the flow data change amount at the current time and the last time exceeds the threshold value four times, judging that the running speed of the actuating mechanism is changed.
4. The method for monitoring the state and diagnosing the fault of the hydraulic system according to claim 3, wherein when any one of the operation states is changed, the characteristic information corresponding to the state is extracted and input into a deep neural network trained in advance to obtain the probability of the fault of each part of the hydraulic system, further comprising:
when any one operation state is changed, performing wavelet transformation and Kalman filtering on data output by a corresponding sensor in sequence;
and extracting the amplitude and the phase of the obtained filtering result to be used as characteristic information corresponding to the running state, inputting the characteristic information into a corresponding input end of the deep neural network trained in advance, and inputting standard values into other input ends to obtain the fault probability of each part of the hydraulic system.
5. The method for monitoring the state and diagnosing faults of a hydraulic system according to claim 4, wherein the normalizing process is performed on the probability of faults occurring at each part, and the normalized result is input into a D-S evidence decision model to obtain the part most likely to have faults and the fault level, further comprising:
the probability y of the fault of each part of the hydraulic system is calculated by the following formulaiNormalization processing is carried out to obtain a normalized vector yi′}，i＝1…N
Figure FDA0002808710350000021
In the formula, yiIs a deep neural netThe ith value of the output vector is calculated, the probability of the fault of the ith position of the hydraulic system is represented, and N is the dimension of the output vector;
taking the normalization result as a first evidence, taking the set fault state distribution vector as a second evidence, inputting the second evidence into a D-S evidence decision model in the following formula, and obtaining decision confidence m (A) after the two evidences are fusedi)
Figure FDA0002808710350000031
Figure FDA0002808710350000032
Figure FDA0002808710350000033
In the formula, AiIndicating a fault event at the ith location of the hydraulic system, AjIndicating a fault event at the jth location of the hydraulic system, biIndicates the ith value, bel, in the assigned fault state vectoriDenotes the confidence coefficient, i ≠ j, i ═ 1 … N, j ═ 1 … i-1i +1N, epsilon1、ε2For a preset threshold value, max { } is a maximum function;
identification { m (A)i) The position of the hydraulic system corresponding to the maximum value is used as a final fault position;
identification { m (A)i) The sensor output data x corresponding to the maximum valueiAccording to said xiAnd comparing with a preset threshold value to obtain the fault grade of the fault part.
6. A condition monitoring and fault diagnosis apparatus for a hydraulic system, comprising:
the multi-information sensor is used for monitoring the running state of each part of the hydraulic system in real time and sending the running state to the preprocessing module;
the preprocessing module is used for judging whether the operation states of all parts of the hydraulic system change or not, extracting the characteristic information corresponding to any one operation state when the operation state changes, inputting the characteristic information into a pre-trained deep neural network to obtain the probability of faults of all parts of the hydraulic system, and transmitting the probability to the fault diagnosis module;
and the fault diagnosis module is used for carrying out normalization processing on the probability of the faults of all the parts, inputting the normalization result into a D-S evidence decision model trained in advance, obtaining the part with the most possible fault and the fault grade, and pushing the part with the most possible fault and the fault grade to an operator.
7. The hydraulic system condition monitoring and fault diagnosis device according to claim 6, wherein the multiple information sensor includes:
the pressure sensors are arranged at the oil outlet of the hydraulic pump, the positive and negative cavities of the hydraulic cylinder, the reversing valve and the speed regulating valve position and are used for acquiring pressure data of the arranged position and monitoring the output torque of the actuating mechanism; and transmitting the pressure data to a pre-processing module;
the flow sensor is arranged at an oil outlet of the hydraulic pump and used for acquiring flow data of the arrangement position and monitoring the running speed of the actuating mechanism; and transmitting the flow data to a preprocessing module;
the vibration sensor is arranged on the surfaces of the hydraulic pump, the motor and the hydraulic mechanical locking cylinder and is used for acquiring vibration data of the arranged position and monitoring the running state of the hydraulic pump; and transmitting the vibration data to a preprocessing module;
the temperature sensors are arranged at the oil outlet of the hydraulic pump, the oil inlet of the hydraulic cylinder and the surface of the oil tank and are used for acquiring temperature data of the arranged positions and monitoring the temperature states of the hydraulic pump, the hydraulic cylinder and the oil tank; and transmitting the temperature data to a preprocessing module.
8. The apparatus according to claim 7, wherein the preprocessing module executes the following program to determine whether the operating status of each part of the hydraulic system changes:
obtaining the variation of the vibration data at the current moment and the last moment; comparing the variation of the vibration data with a threshold value to judge whether the running state of the hydraulic pump changes, and when the variation exceeds the threshold value, judging that the running state of the hydraulic pump changes, otherwise, no variation exists;
obtaining the variation of the temperature data at the current moment and the previous moment, comparing the variation of the temperature data with a second threshold value to judge whether the temperature states of the hydraulic pump, the hydraulic cylinder and the oil tank are changed, and when any variation exceeds the second threshold value, judging that the temperature states of the corresponding hydraulic pump, the corresponding hydraulic cylinder and/or the corresponding oil tank are changed, otherwise, judging that no variation exists;
obtaining the change quantity of the pressure data at the current moment and the last moment; comparing the change amount of the pressure data with a third threshold value to judge whether the output torque of the actuating mechanism changes, and when the change amount exceeds the third threshold value, judging that the output torque of the actuating mechanism changes, otherwise, no change occurs;
obtaining the change quantity of the flow data at the current moment and the last moment; and comparing the change amount of the flow data with a threshold value four to judge whether the running speed of the execution mechanism changes, and when the change amount exceeds the threshold value four, judging that the running speed of the execution mechanism changes, otherwise, judging that no change exists.
9. The apparatus according to claim 8, wherein the preprocessing module executes the following program to obtain the probability of the failure of each part of the hydraulic system:
when any one operation state is changed, performing wavelet transformation and Kalman filtering on data output by a corresponding sensor in sequence;
and extracting the amplitude and the phase of the obtained filtering result to be used as characteristic information corresponding to the running state, inputting the characteristic information into a corresponding input end of the deep neural network trained in advance, and inputting standard values into other input ends to obtain the fault probability of each part of the hydraulic system.
10. The hydraulic system condition monitoring and fault diagnosing apparatus according to claim 9, wherein the fault diagnosing module performs the following procedures to obtain a location where a fault is most likely to occur, and a fault level:
the probability y of the fault of each part of the hydraulic system is calculated by the following formulaiNormalization processing is carried out to obtain a normalized vector yi′}，i＝1…N
Figure FDA0002808710350000051
In the formula, yiThe ith value of the output vector of the deep neural network represents the fault probability of the ith part of the hydraulic system, and N is the dimension of the output vector;
taking the normalization result as a first evidence, taking the set fault state allocation vector as a second evidence, inputting the second evidence into a D-S evidence decision model in the following formula, and obtaining the decision confidence level m (A) after the two evidences are fusedi)
Figure FDA0002808710350000061
Figure FDA0002808710350000062
Figure FDA0002808710350000063
In the formula, AiIndicating a fault event at the ith location of the hydraulic system, AjIndicating a fault event at the jth location of the hydraulic system, biIndicates the ith value, bel, in the assigned fault state vectoriDenotes the confidence coefficient, i ≠ j, i ═ 1 … N, j ═ 1 … i-1i +1N, epsilon1、ε2For a preset threshold value, max { } is a maximum function;
identification { m (A)i) The position of the hydraulic system corresponding to the maximum value is used as a final fault position;
identification { m (A)i) The sensor output data x corresponding to the maximum valueiAccording to said xiAnd comparing with a preset threshold value to obtain the fault grade of the fault part.
CN202011378332.5A 2020-11-30 2020-11-30 State monitoring and fault diagnosis method and device for hydraulic system Pending CN114576236A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011378332.5A CN114576236A (en) 2020-11-30 2020-11-30 State monitoring and fault diagnosis method and device for hydraulic system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011378332.5A CN114576236A (en) 2020-11-30 2020-11-30 State monitoring and fault diagnosis method and device for hydraulic system

Publications (1)

Publication Number Publication Date
CN114576236A true CN114576236A (en) 2022-06-03

Family

ID=81767349

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011378332.5A Pending CN114576236A (en) 2020-11-30 2020-11-30 State monitoring and fault diagnosis method and device for hydraulic system

Country Status (1)

Country Link
CN (1) CN114576236A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116379043A (en) * 2023-04-11 2023-07-04 浙江勤鹏科技股份有限公司 Fault detection method and system for spin-on oil filter

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116379043A (en) * 2023-04-11 2023-07-04 浙江勤鹏科技股份有限公司 Fault detection method and system for spin-on oil filter
CN116379043B (en) * 2023-04-11 2024-03-08 浙江勤鹏科技股份有限公司 Fault detection method and system for spin-on oil filter

Similar Documents

Publication Publication Date Title
WO2023071217A1 (en) Multi-working-condition process industrial fault detection and diagnosis method based on deep transfer learning
CN100595710C (en) Data processing method, data processing device, diagnosis method, and diagnosis device
US7539597B2 (en) Diagnostic systems and methods for predictive condition monitoring
US7496798B2 (en) Data-centric monitoring method
JP5016519B2 (en) Diagnostic system and method for predictive condition monitoring
US7941701B2 (en) Fuzzy classification approach to fault pattern matching
US8204697B2 (en) System and method for health assessment of downhole tools
EP1960853B1 (en) Evaluating anomaly for one-class classifiers in machine condition monitoring
JP5179086B2 (en) Industrial process monitoring method and monitoring system
CN112508105B (en) Fault detection and retrieval method for oil extraction machine
CN111538311B (en) Flexible multi-state self-adaptive early warning method and device for mechanical equipment based on data mining
CN109298633A (en) Chemical production process fault monitoring method based on adaptive piecemeal Non-negative Matrix Factorization
CN110757510A (en) Method and system for predicting remaining life of robot
CN108803555B (en) Sub-health online identification and diagnosis method based on performance monitoring data
CN113574480A (en) Apparatus for predicting equipment damage
CN114576236A (en) State monitoring and fault diagnosis method and device for hydraulic system
Liu et al. A machine learning-based clustering approach to diagnose multi-component degradation of aircraft fuel systems
CN112418460A (en) Fault diagnosis method and fault diagnosis device for engineering vehicle
CN114371677B (en) Industrial process state monitoring method based on spectrum radius-interval principal component analysis
Song et al. A Shape-constrained Transfer Temporal Transformer Network for remaining useful life prediction of rotating machines
CN116702030B (en) Blast furnace state monitoring method and device based on sensor reliability analysis
Fountoulaki et al. Using neural networks for mean shift identification and magnitude of bivariate autocorrelated processes
CN108960654A (en) A kind of the equipment part health control method and system of Machine oriented study
CN117671393B (en) Fault monitoring method and system for electrical mechanical equipment
Agung’J et al. A multi-domain tool state classifier model for computer numerically controlled machining

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination