CN115018002A - Marine low-speed diesel engine air valve service health intelligent diagnosis system and method - Google Patents

Abstract

The invention relates to the technical field of monitoring of ultra-high-power low-speed diesel engines, in particular to a marine low-speed diesel engine air valve service health intelligent diagnosis system and method. According to the method and the device, massive data do not need to be collected for analysis, and a reliable abrasion loss analysis result can be obtained only by collecting key characteristic parameters of the rotating speed and the cylinder cover vibration signal, so that compared with the prior art, the collected data amount is greatly reduced. Besides, a fluid-solid-structure field dynamic coupling model is constructed for a cylinder cover, a cylinder body and a non-abrasion air valve, the fluid-solid-structure field dynamic coupling model is corrected through a cylinder cover vibration signal, a crank angle signal and an in-cylinder pressure signal of the air valve in a normal state (namely, when the air valve is not abraded), the corrected finite element model can simulate the in-cylinder pressure and the cylinder cover vibration signal along with the change rule of the crank angle, which are close to real data, and sufficient training data can be provided for an abrasion analysis model.

Description

Marine low-speed diesel engine air valve service health intelligent diagnosis system and method

Technical Field

The invention relates to the technical field of monitoring of ultra-high-power low-speed diesel engines, in particular to a marine low-speed diesel engine air valve service health intelligent diagnosis system and method.

Background

The ultra-high power low-speed diesel engine is used as the heart of an ocean vessel and needs to be broken through to support the development of the high-technology vessel industry in China.

Data show that the faults of parts of the diesel engine account for about 40-50% of the total faults of the ship, and particularly the fault rate of a gas valve of a gas distribution system accounts for 15% of the faults of the whole diesel engine. The marine diesel engine with the working speed of 60-120r/min is a low-speed diesel engine. In the working process of a low-speed diesel engine, the working environment of the valve face of the exhaust valve is the worst, particularly, when the low-speed diesel engine runs under full load, the valve face of the exhaust valve is abraded under high-frequency high-speed impact of the exhaust valve and the valve seat, the main failure mode in the service process of the exhaust valve is also the mode, and the specific gravity of the failure mode occupies more than half of the failure modes of all the exhaust valves.

Because the low-speed diesel engine is huge in size and overweight, the maintenance of the exhaust valve is generally carried out in an empirical mode, namely the maintenance of the ship is stopped after the ship sails for a period of time, and the problem can be really detected only after the air valve is damaged in the mode; moreover, for the air valve which needs to be repeatedly matched with other parts for use, the problem of the air outlet valve in operation, such as abnormal abrasion of the air valve, cannot be well detected by adopting a halt detection mode.

However, if the wear of the gas valve is analyzed by collecting the operating parameters of the gas valve, a more complicated analysis model is required to achieve the purpose due to more operating parameters. The analysis model needs a large amount of abrasion loss and corresponding operation data as training data, and the acquisition amount of the data is large, so that the time and the labor are consumed. In addition, the air valve cannot be used for ship operation after being worn to a certain extent, otherwise, the ship has potential safety hazards. Because the air valve with certain abrasion cannot be used for ship work, enough training data for training the analysis model cannot be obtained, the analysis model cannot complete training, and real-time analysis of the abrasion of the air valve is difficult to realize.

Therefore, how to conveniently, quickly and accurately complete the analysis of valve abrasion when the low-speed diesel engine valve works normally so as to improve the real-time monitoring effect of the large diesel engine valve, and the problem to be solved urgently is formed.

Disclosure of Invention

Aiming at the defects of the prior art, the invention provides the service health intelligent diagnosis system and method for the marine low-speed diesel engine air valve, which can conveniently, quickly and accurately finish the analysis of the air valve abrasion and improve the real-time monitoring effect of the large diesel engine air valve.

In order to solve the technical problems, the invention adopts the following technical scheme:

marine low-speed diesel engine air valve service health intelligent diagnosis system includes:

the acquisition unit is used for acquiring current operation data of the low-speed diesel engine air valve, wherein the current operation data comprises a current crank angle signal, a current rotating speed signal and a current cylinder cover vibration signal;

the signal processing unit is used for extracting and processing the characteristics of the current cylinder cover vibration signal and extracting and obtaining key characteristic parameters which are significant in relation to the air valve abrasion in the current cylinder cover vibration signal and a corresponding signal map;

the health evaluation unit is provided with a trained wear analysis model between the wear amount and the signal characteristic parameters, and the trained wear analysis model between the wear amount and the signal characteristic parameters is used for indicating the corresponding mapping relation between the key characteristic parameters of the cylinder cover vibration signals and the wear amount of the low-speed diesel engine air valve at different running rotating speeds; the health evaluation unit is used for determining the current abrasion loss of the low-speed diesel engine air valve according to the current rotating speed signal and the key characteristic parameter of the current cylinder cover vibration signal through an abrasion analysis model between the abrasion loss and the signal characteristic parameter, and diagnosing the health state of the low-speed diesel engine air valve according to the current abrasion loss;

and the database management unit is used for storing the attribute parameters, the operation data, the characteristic parameters and the signal spectrum of the cylinder cover vibration signals and the health state diagnosis result of the low-speed diesel engine air valve.

Preferably, the key characteristic parameters of the cylinder cover vibration signal comprise one or more of time domain characteristics, frequency domain characteristics and time-frequency domain-information entropy; the time domain features comprise dimensional features and dimensionless features, the dimensional features comprise exhaust amplitude, closed gas amplitude, impact amplitude and explosion amplitude, and the dimensionless features comprise kurtosis, a form factor, a peak factor, a pulse factor and a margin factor; the frequency domain features comprise center of gravity frequency and mean square frequency; the time-frequency domain-information entropy comprises power spectrum entropy, singular spectrum entropy and energy entropy.

Preferably, the wear analysis model is obtained by training as follows:

s1, constructing a three-dimensional model of a cylinder cover, a cylinder body and a wear-free air valve, then carrying out fluid-solid coupling, constructing a fluid-solid-structure field dynamic coupling model of the air valve service process, and simulating and analyzing the cycle dynamic response rule of an air valve thermal field to obtain the change rule of the pressure in the cylinder and the vibration signal of the cylinder cover along with the crank rotation angle;

s2, according to the measured cylinder cover vibration signal, crank angle signal and cylinder pressure signal of the non-abrasion air valve, correcting the fluid-solid-structure field dynamic coupling model to make the difference between the characteristic amplitude obtained by the simulation and the measured result less than the preset error value; the characteristic amplitude comprises in-cylinder pressure, exhaust amplitude, closed gas amplitude, impact amplitude and explosion amplitude;

s3, simulating and constructing air valves with different wear states based on the corrected fluid-solid-structure field dynamic coupling model, and establishing fluid-solid-structure field dynamic coupling models of the air valves in different wear states for simulating and analyzing cylinder cover vibration signals of the air valves at different rotating speeds and different wear states; the wear state includes an amount of wear;

s4, extracting characteristic parameters in a cylinder cover vibration signal in the simulation result of S3, carrying out abrasion correlation analysis, and selecting characteristic items with obvious relations in the characteristic parameters according to a preset standard as key characteristic parameters influencing air valve abrasion;

and S5, constructing and training a BP neural network model by taking the rotating speed and the selected key characteristic parameters as input variables and the abrasion loss as output variables to obtain an abrasion analysis model between the trained abrasion loss and the signal characteristic parameters.

Preferably, in S1, the process of constructing the fluid-solid-structure field dynamic coupling finite element model of the gas valve service process includes: constructing three-dimensional models of a cylinder cover, a cylinder body and an air valve and endowing the three-dimensional models with material properties; setting a gas flow equation, an oil injection parameter and a combustion equation in a fluid field, and calculating the pressure and the temperature in a cylinder; setting the motion process of the air valve and the constraint of the cylinder cover in a structural field, and carrying out vibration analysis and calculation; setting the interface of the fluid and the solid as a channel for mutual transmission of a fluid-solid field, wherein the interface is of a coupling type, so as to realize real-time interactive calculation; and establishing a dynamic coupling model of the fluid-solid-structure field in the service process of the gas valve.

Preferably, in S3, the wear state further includes a wear portion; the wear part comprises a valve surface and a disc bottom;

in the step S4, when the wear correlation analysis is performed, the relationship between the valve face wear amount and the characteristic parameter and the relationship between the disc bottom wear amount and the characteristic parameter are respectively analyzed; respectively selecting characteristic items with obvious relations in the characteristic parameters according to a preset standard as key characteristic parameters influencing the valve surface abrasion loss and the disc bottom abrasion loss;

in the step S5, when the BP neural network model is constructed, a digitized mapping relationship between the valve face wear amount and the rotational speed and between the cylinder head vibration signal characteristic parameters, and a digitized mapping relationship between the disc bottom wear amount and the rotational speed and between the disc bottom vibration signal characteristic parameters are respectively established.

Preferably, the wear amount includes a valve face wear amount and a disc bottom wear amount.

Preferably, when the health evaluation unit diagnoses the health state of the low-speed diesel engine air valve, if the valve surface abrasion loss or the disc bottom abrasion loss exceeds the corresponding early warning value, an early warning is sent out.

Preferably, the attribute parameters include a diesel engine model, a gas valve material and corresponding performance parameters thereof, and piston motion parameters.

Preferably, the performance parameters include mass density, specific heat, young's modulus and stress strain; the piston motion parameters comprise the crankshaft radius, the length of a connecting rod and the rotating speed of a main shaft; the characteristic parameters and the signal spectrum of the cylinder cover vibration signal comprise a real ship detection signal and a finite element simulation signal.

The application also provides a marine low-speed diesel engine air valve service health intelligent diagnosis method, which is implemented by using the marine low-speed diesel engine air valve service health intelligent diagnosis system and comprises the following steps:

acquiring current operation data of a low-speed diesel engine air valve through an acquisition unit, wherein the current operation data comprises a current crank angle signal, a current rotating speed signal and a current cylinder cover vibration signal;

the signal processing unit is used for extracting the characteristics of the current cylinder cover vibration signal to obtain key characteristic parameters which are significant in relation to the air valve abrasion in the current cylinder cover vibration signal and a corresponding signal map;

and the health evaluation unit determines the current abrasion loss of the low-speed diesel engine air valve according to the current rotating speed signal and the key characteristic parameter of the current cylinder cover vibration signal through an abrasion analysis model between the abrasion loss and the signal characteristic parameter, and diagnoses the health state of the low-speed diesel engine air valve according to the current abrasion loss.

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

1. compared with the prior art which needs to collect massive data, the inventor directly uses all operation data to analyze the abrasion loss of the low-speed diesel engine air valve without blindly according to a conventional mode, and diagnoses and detects the abrasion loss of the low-speed diesel engine air valve through the corresponding mapping relation between the key characteristic parameters of the cylinder cover vibration signals and the abrasion loss of the low-speed diesel engine air valve at different operation speeds after finding out the strong correlation between the rotation speed and the key characteristic parameters and the abrasion loss of the cylinder cover vibration signals. And through actual data verification, the strong correlation between the key characteristic parameters of the rotating speed and the cylinder cover vibration signal and the abrasion loss is checked.

In the detection mode, mass data are not needed to be analyzed, and a reliable abrasion loss analysis result can be obtained only by acquiring key characteristic parameters of the rotating speed and the cylinder cover vibration signal. Compared with the prior art, the method and the device have the advantages that the accuracy of the analysis result is guaranteed, and meanwhile the data demand when the abrasion loss is analyzed is greatly reduced.

2. Although the operation data after the valve is worn cannot be obtained as the training data, the data of the valve in the normal state can be obtained simply. Based on the basis, a fluid-solid-structure field dynamic coupling model is established for a cylinder cover, a cylinder body and a non-abrasion air valve, and is corrected through a cylinder cover vibration signal, a crank angle signal and an in-cylinder pressure signal of the air valve in a normal state (namely, when the air valve is not abraded). Therefore, the modified finite element model can simulate the change rule of the in-cylinder pressure and the cylinder cover vibration signal along with the crank angle, which is close to real data.

And then, simulating and constructing the gas valve with different wear states based on the corrected fluid-solid-structure field dynamic coupling model, and establishing the fluid-solid-structure field dynamic coupling model of the gas valve in different wear states. Through the fluid-solid-structure field dynamic coupling model of the air valve in different wear states, the vibration signals of the cylinder cover of the air valve in different rotating speeds and different wear states can be simulated nearly to be real, and simulation data which are similar to the real state are obtained. In this way, the problem that the training data volume of the analysis model is difficult to acquire can be solved. And then, constructing and training a BP neural network model by using the rotating speed and the cylinder cover vibration signal in the simulation data as input variables and using the abrasion loss as an output variable to obtain an abrasion analysis model.

Finally, when the air valve of the large diesel engine normally works, the current abrasion loss of the air valve (valve surface or disk bottom) can be calculated through an abrasion analysis model only by acquiring the current actual rotating speed and the actual cylinder cover vibration signal, so that whether the current air valve is in a healthy state or not can be known.

The application can conveniently, quickly and accurately complete the analysis of the valve abrasion when the low-speed diesel engine valve normally works.

3. In order to ensure the accuracy of the analysis model, after the simulation results of the fluid-solid-structure field dynamic coupling model of the air valve in different wear states are obtained, wear correlation analysis is carried out on the characteristic parameters in the cylinder head vibration signal, and the obvious relation between the characteristic parameters in the cylinder head vibration signal and the abrasion loss of the air valve (valve face or disc bottom) is known. By the method, parameter interference terms can be reduced, and the accuracy of the wear analysis model is ensured.

4. After the characteristic items are used as key characteristic parameters influencing air valve abrasion, a BP neural network model is constructed and trained, and a digital mapping relation between valve face abrasion loss and rotating speed and between cylinder cover vibration signal characteristic parameters and a digital mapping relation between disc bottom abrasion loss and rotating speed and between cylinder cover vibration signal characteristic parameters are respectively established. Therefore, when the wear analysis model is used, the wear amount can be accurately judged, and the wear position can be accurately identified, so that the current health state of the air valve can be better judged.

Drawings

For purposes of promoting a better understanding of the objects, aspects and advantages of the invention, reference will now be made in detail to the present invention as illustrated in the accompanying drawings, in which:

FIG. 1 is a logic block diagram of an intelligent diagnosis system for service health of a gas valve of a marine low-speed diesel engine according to a first embodiment;

FIG. 2 is a schematic diagram of a second embodiment of a system for intelligently diagnosing the gas valve service health of a marine low-speed diesel engine;

FIG. 3 is a flowchart illustrating an implementation of the intelligent diagnosis system for the service health of the gas valve of the marine low-speed diesel engine according to the second embodiment;

FIG. 4 is a schematic diagram showing the operation of the signal analysis software according to the second embodiment;

FIG. 5 is a schematic structural diagram of a database management system according to a second embodiment;

FIG. 6 is a schematic diagram of the fluid-solid-structure field dynamic coupling model, simulated and measured in-cylinder pressure and cylinder head vibration signals in the second embodiment;

fig. 7 is a schematic diagram of the marine low speed diesel valve in different wear states according to the second embodiment.

Detailed Description

The following is further detailed by the specific embodiments:

example one

As shown in fig. 1, the present embodiment discloses an intelligent diagnosis system for service health of a gas valve of a marine low-speed diesel engine, which includes an acquisition unit, a signal processing unit, a health assessment unit, and a database management unit.

The acquisition unit is used for acquiring current operation data of the low-speed diesel engine air valve, wherein the current operation data comprises a current crank angle signal, a current rotating speed signal and a current cylinder cover vibration signal;

the signal processing unit is used for extracting and processing the characteristics of the current cylinder cover vibration signal to extract and obtain key characteristic parameters which are obviously related to the air valve abrasion in the current cylinder cover vibration signal and a corresponding signal map;

the health evaluation unit is provided with a trained wear analysis model between the wear amount and the signal characteristic parameters, and the trained wear analysis model between the wear amount and the signal characteristic parameters is used for indicating a corresponding mapping relation between key characteristic parameters of a cylinder cover vibration signal and the wear amount of a low-speed diesel engine air valve at different operation rotating speeds; and the health evaluation unit is used for determining the current wear loss of the low-speed diesel engine air valve according to the current rotating speed signal and the key characteristic parameters of the current cylinder cover vibration signal through a wear analysis model between the wear loss and the signal characteristic parameters, and diagnosing the health state of the low-speed diesel engine air valve according to the current wear loss. The key characteristic parameters of the cylinder cover vibration signal comprise one or more of time domain characteristics, frequency domain characteristics and time-frequency domain-information entropy; the time domain features comprise dimensional features and dimensionless features, the dimensional features comprise exhaust amplitude, closed gas amplitude, impact amplitude and explosion amplitude, and the dimensionless features comprise kurtosis, a form factor, a peak factor, a pulse factor and a margin factor; the frequency domain features comprise center of gravity frequency and mean square frequency; the time-frequency domain-information entropy comprises power spectrum entropy, singular spectrum entropy and energy entropy; wherein the wear amount includes a valve face wear amount and a disc bottom wear amount.

The database management unit is used for storing attribute parameters, operation data, characteristic parameters and signal maps of cylinder head vibration signals and health state diagnosis results of the low-speed diesel engine air valve. The attribute parameters comprise the model of the diesel engine, the material of the air valve and corresponding performance parameters and piston motion parameters. The performance parameters comprise mass density, specific heat, Young modulus and stress strain; the piston motion parameters include crankshaft radius, connecting rod length, and spindle speed. The characteristic parameters and the signal spectrum of the cylinder cover vibration signal comprise a real ship detection signal and a finite element simulation signal.

The wear analysis model is obtained by training in the following way:

s1, constructing a three-dimensional model of the cylinder cover, the cylinder body and the wear-free air valve, then carrying out fluid-solid coupling, constructing a fluid-solid-structure field dynamic coupling model of the air valve service process, and simulating and analyzing the cycle dynamic response rule of the air valve thermal field to obtain the change rule of the cylinder pressure and the cylinder cover vibration signal along with the crank rotation angle. In specific implementation, the process of constructing the dynamic coupling model of the fluid-solid-structure field in the service process of the gas valve comprises the following steps: constructing three-dimensional models of a cylinder cover, a cylinder body and an air valve and endowing the three-dimensional models with material properties; setting a gas flow equation, an oil injection parameter and a combustion equation in a fluid field, and calculating the pressure and the temperature in a cylinder; setting the motion process of the air valve and the constraint of the cylinder cover in a structural field, and carrying out vibration analysis and calculation; the interface of the fluid and the solid is a channel for mutual transmission of fluid-solid fields (including temperature, pressure, displacement and the like), and the interface is of a coupling type, so that real-time interactive calculation is realized; and establishing a dynamic coupling model of the fluid-solid-structure field in the service process of the gas valve.

S2, correcting the fluid-solid-structure field dynamic coupling model according to the actually measured cylinder cover vibration signal, crank angle signal and in-cylinder pressure signal of the non-wear air valve, so that the difference value between the characteristic amplitude obtained by simulation and the actually measured result is smaller than a preset error value; the characteristic amplitude comprises in-cylinder pressure, exhaust amplitude, closed gas amplitude, impact amplitude and explosion amplitude;

s3, simulating and constructing air valves with different wear states based on the corrected fluid-solid-structure field dynamic coupling model, and establishing fluid-solid-structure field dynamic coupling models of the air valves in different wear states for simulating and analyzing cylinder cover vibration signals of the air valves at different rotating speeds and different wear states; the wear state includes an amount of wear; the wear state further comprises a wear location; the wear portion includes a valve face and a disc bottom.

S4, extracting characteristic parameters in a cylinder cover vibration signal in the simulation result of S3, carrying out abrasion correlation analysis, and selecting characteristic items with obvious relations in the characteristic parameters according to a preset standard as key characteristic parameters influencing air valve abrasion; specifically, when the wear correlation analysis is carried out, the relationship between the valve surface wear amount and the characteristic parameter and the relationship between the disc bottom wear amount and the characteristic parameter are respectively analyzed; and respectively selecting characteristic items with obvious relations in the characteristic parameters according to a preset standard as key characteristic parameters influencing the valve surface abrasion loss and the disc bottom abrasion loss.

And S5, constructing and training a BP neural network model by taking the rotating speed and the selected key characteristic parameters as input variables and the abrasion loss as output variables to obtain an abrasion analysis model between the trained abrasion loss and the signal characteristic parameters. Specifically, when a BP neural network model is constructed, a digital mapping relation between valve face abrasion loss and rotating speed and cylinder cover vibration signal characteristic parameters and a digital mapping relation between disc bottom abrasion loss and rotating speed and cylinder cover vibration signal characteristic parameters are respectively established.

When the health evaluation unit is used for diagnosing the health state of the low-speed diesel engine air valve in specific implementation, if the valve surface abrasion loss or the disc bottom abrasion loss exceeds the corresponding early warning value, early warning is sent out. In this embodiment, if the valve surface wear depth of the air valve exceeds 2mm, or the air valve disk bottom wear depth exceeds 15mm, an early warning is generated. In other embodiments, a person skilled in the art can set a specific warning value according to the specific type of the gas valve.

The invention also provides an intelligent diagnosis method for the service health of the gas valve of the marine low-speed diesel engine, which is implemented by adopting the intelligent diagnosis system for the service health of the gas valve of the marine low-speed diesel engine and comprises the following steps:

acquiring current operation data of a low-speed diesel engine air valve through an acquisition unit, wherein the current operation data comprises a current crank angle signal, a current rotating speed signal and a current cylinder cover vibration signal;

the signal processing unit is used for extracting the characteristics of the current cylinder cover vibration signal to obtain key characteristic parameters which are significant in relation to the air valve abrasion in the current cylinder cover vibration signal and a corresponding signal map;

and the health evaluation unit determines the current wear loss of the low-speed diesel engine air valve according to the current rotating speed signal and the key characteristic parameters of the current cylinder cover vibration signal through a wear analysis model between the wear loss and the signal characteristic parameters, and diagnoses the health state of the low-speed diesel engine air valve according to the current wear loss.

Compared with the prior art which needs to acquire massive data, the inventor does not blindly analyze the abrasion loss of the low-speed diesel engine air valve by directly using all operation data according to a conventional mode, but diagnoses and detects the abrasion loss of the low-speed diesel engine air valve through the corresponding mapping relation between the key characteristic parameters of the cylinder cover vibration signals and the abrasion loss of the low-speed diesel engine air valve under different operation speeds after finding out the strong correlation between the key characteristic parameters of the rotation speed and the cylinder cover vibration signals and the abrasion loss. And through actual data verification, the strong correlation between the key characteristic parameters of the rotating speed and the cylinder cover vibration signal and the abrasion loss is checked. In the detection mode, mass data are not needed to be analyzed, and a reliable abrasion loss analysis result can be obtained only by acquiring key characteristic parameters of the rotating speed and the cylinder cover vibration signal. Compared with the prior art, the method and the device have the advantages that the accuracy of the analysis result is guaranteed, and meanwhile the data demand when the abrasion loss is analyzed is greatly reduced.

On the other hand, although the operation data after the valve is worn cannot be acquired as the training data, the data of the valve in the normal state is very simply acquired. Based on the basis, a fluid-solid-structure field dynamic coupling model is established for a cylinder cover, a cylinder body and a non-abrasion air valve, and is corrected through a cylinder cover vibration signal, a crank angle signal and an in-cylinder pressure signal of the air valve in a normal state (namely, when the air valve is not abraded). Therefore, the modified finite element model can simulate the change rule of the in-cylinder pressure and the cylinder cover vibration signal along with the crank angle, which is close to real data. And then, simulating and constructing the gas valve with different wear states based on the corrected fluid-solid-structure field dynamic coupling model, and establishing the fluid-solid-structure field dynamic coupling model of the gas valve in different wear states. Through the fluid-solid-structure field dynamic coupling model of the air valve in different wear states, the vibration signals of the cylinder cover of the air valve in different rotating speeds and different wear states can be simulated nearly to be real, and simulation data which are similar to the real state are obtained. In this way, the problem that the training data volume of the analysis model is difficult to acquire can be solved. And then, constructing and training a BP neural network model by using the rotating speed and the cylinder cover vibration signal in the simulation data as input variables and using the abrasion loss as an output variable to obtain an abrasion analysis model.

In order to ensure the accuracy of the analysis model, after the simulation results of the fluid-solid-structure field dynamic coupling model of the air valve in different wear states are obtained, the wear correlation analysis can be performed on the characteristic parameters in the cylinder head vibration signal, and the obvious relation between the characteristic parameters in the cylinder head vibration signal and the wear amount of the air valve (valve surface or disc bottom) is known. By the method, parameter interference terms can be reduced, and the accuracy of the wear analysis model is ensured. After the characteristic items are used as key characteristic parameters influencing air valve abrasion, a BP neural network model is constructed and trained, and a digital mapping relation among valve face abrasion loss, rotating speed and cylinder cover vibration signal characteristic parameters and a digital mapping relation among disc bottom abrasion loss, rotating speed and cylinder cover vibration signal characteristic parameters are respectively established. Therefore, when the wear analysis model is used, the wear amount can be accurately judged, and the wear position can be accurately identified, so that the current health state of the air valve can be better judged. Finally, when the air valve of the large diesel engine normally works, the current abrasion loss of the air valve (valve surface or disk bottom) can be calculated through an abrasion analysis model only by acquiring the current actual rotating speed and the actual cylinder cover vibration signal, so that whether the current air valve is in a healthy state or not can be known. If the valve surface abrasion depth of the air valve exceeds 2mm or the air valve plate bottom abrasion depth exceeds 15mm, an early warning is sent out. By using the method, the valve wear analysis can be conveniently, quickly and accurately completed when the diesel engine valve normally works.

Example two

For better understanding of the content of the present application, the following describes a construction process of the marine low-speed diesel engine valve service health intelligent diagnosis system in the present application.

The specific construction process is shown in fig. 2, and mainly includes 4 parts, specifically, a low-speed engine running state real ship monitoring system (i.e., an acquisition unit and a signal processing unit), a low-speed engine running state finite element simulation, a low-speed engine air valve service prediction model (i.e., a wear analysis model), and a low-speed engine air valve service state health assessment system (i.e., a health assessment unit and a database management unit).

Firstly, constructing a real ship monitoring system (namely an acquisition unit and a signal processing unit) in the running state of the low-speed machine, and acquiring real ship monitoring signal data in the running process of the low-speed machine, wherein the real ship monitoring signal data comprises a crank angle signal, a rotating speed signal and a cylinder cover vibration signal; in order to realize vibration signal prediction in the service process of the low-speed engine air valve, a finite element analysis model of the running state of the low-speed engine is established and the finite element model is corrected based on the actually measured signal.

Then, a low-speed engine air valve service prediction model (namely, a wear analysis model) is built, finite element simulation of the air valve under different rotating speeds and different clearances, namely different wear quantities is carried out, simulated vibration signals of different wear quantities under different rotating speeds are obtained, a BP neural network model is built by taking the characteristic value and the rotating speed of the vibration signals as input variables and the wear quantity of the air valve as output variables, and then a digital mapping relation between the wear quantity of the air valve and the characteristic values of the rotating speed and the vibration signals is built.

And constructing a low-speed engine air valve service state health assessment system (namely a health assessment unit and a database management unit), realizing the query of real ship monitoring signal data and finite element simulation data, and simultaneously carrying out intelligent monitoring identification and health assessment on the air valve wear degradation state based on a low-speed engine air valve service prediction model contained in the low-speed engine air valve service state health assessment system.

The real ship monitoring signal data of the low-speed engine air valve full life cycle running state process can be obtained through the constructed real ship detection system of the low-speed engine running state, the obtained real ship monitoring signal data is subjected to signal analysis and is led into the low-speed engine air valve service state health assessment system, then the air valve wear degradation state is intelligently diagnosed, and finally the intelligent monitoring and identification of the air valve wear state and the effective and intelligent control of the air valve full life cycle in the low-speed engine running state are realized. The specific implementation process of the intelligent diagnosis system for the service health of the gas valve of the marine low-speed diesel engine is shown in fig. 3.

The functions and the construction process of each module are as follows:

1. real ship monitoring system (namely acquisition unit and signal processing unit) for low-speed machine running state

Acquiring signal data of a low-speed engine in the running process for real ship monitoring, mounting a vibration sensor on the surface of an engine cylinder cover, carrying out a series of physical tests under the normal running working condition of a low-speed engine air valve, and monitoring and acquiring a crank angle signal, a cylinder cover vibration signal and a rotating speed signal in the running process of the low-speed engine on line; further, the real ship monitoring signal data is analyzed and processed, and signal analysis software is developed as shown in fig. 4. And acquiring actually-measured signal characteristics and a signal map through signal analysis software, wherein the actually-measured signal characteristics and the signal map comprise key characteristic values and signal maps such as time domain characteristics, frequency domain characteristics, time-frequency domain-information entropy, exhaust amplitude values, closed gas amplitude values, impact amplitude values, explosion amplitude values and the like. The signal analysis software can analyze and process signal data acquired under different running states of the low-speed machine. After the processing is finished, the signal analysis software can also export the processed actually-measured signal characteristics and signal maps in a file form, and further serve as basic data of the low-speed engine air valve service state health assessment system.

2. Finite element simulation of running state of low-speed machine

(1) Establishment and correction of dynamic coupling model of fluid-solid-structure field in gas valve service process

According to actual air valve parameters, a dynamic coupling model of a fluid-solid-structure field in an air valve service process is established, the cycle dynamic response rule of an air valve thermal field is analyzed, and the change rule of the in-cylinder pressure and the cylinder cover vibration signal along with the crank angle is obtained. And correcting the fluid-solid-structure field dynamic coupling model according to the actually measured cylinder cover vibration signal, so that the characteristic amplitudes of the in-cylinder pressure, the cylinder cover vibration signal diagram, the exhaust amplitude, the gas closing amplitude, the impact amplitude, the explosion amplitude and the like obtained by simulation are close to or identical with the actually measured result.

(2) Cylinder cover vibration signal of simulation air valve under different wearing and degrading states

And constructing finite element models of the air valve in different wear states based on the modified fluid-solid-structure field dynamic coupling model, and simulating to obtain cylinder cover vibration signals of the air valve in different wear degradation states at different rotating speeds.

3. Low-speed machine air valve abrasion loss prediction model (namely abrasion analysis model)

In order to realize vibration signal prediction in the service process of the air valve, based on finite element simulation results of the air valve in different wear degradation states at different rotating speeds, analyzing cylinder cover vibration signal data of the air valve in different wear loss at different rotating speeds, and extracting characteristic parameters such as a quasi-periodic signal time domain, a frequency domain, a time-frequency domain-information entropy and the like; and establishing a BP neural network model by taking the vibration signal characteristic value and the rotating speed as input variables and taking the air valve abrasion loss as an output variable, further establishing a digital mapping relation between the air valve abrasion loss and the rotating speed and vibration signal characteristic value, and further providing theoretical guidance and technical reference for the abrasion loss prediction of the air valve in different abrasion degradation states in the service process of the low-speed machine.

4. Low-speed engine air valve service state health assessment system (namely health assessment unit and database management unit)

The low-speed engine air valve service state health assessment system comprises two parts, namely a low-speed engine service state database management system (namely a database management unit) formed by real ship monitoring signals and finite element simulation data, and an air valve service state health assessment system (namely a health assessment unit) obtained based on a low-speed engine air valve service prediction model.

(1) Database management system (database management unit)

And carrying out systematic induction and arrangement on the low-speed machine real ship monitoring signals and the finite element simulation data to establish a low-speed machine service state database management system, wherein the database mainly comprises material parameters, real ship monitoring signal data and finite element simulation results. As shown in fig. 5.

The material parameters comprise diesel engine model, air valve material and corresponding performance parameters such as mass density, specific heat, Young modulus, stress strain and the like, and piston motion parameters such as crankshaft radius, connecting rod length, main shaft rotating speed and the like. The real ship monitoring signals comprise crank angle signals, cylinder cover vibration signals, rotating speed signals and characteristic signal values and signal maps obtained based on signal analysis software. The finite element simulation data comprises temperature, stress, strain, in-cylinder pressure, cylinder cover vibration signals in the service process of the air valve, and characteristic signal values and signal maps obtained by the finite element simulation data based on signal analysis software. The finite element simulation data also comprises temperature, stress, strain, in-cylinder pressure, cylinder cover vibration signals of the air valve in service process under different abrasion amounts of the valve surface, and characteristic signal values and signal maps obtained by the air valve based on signal analysis software. The characteristic signal values further comprise a discharge amplitude, a gas closing amplitude, an impact amplitude and a combustion and explosion amplitude corresponding to the cylinder cover vibration signals.

The database management system can also selectively compare and analyze the signal analysis results of the air valve under different abrasion quantities of the valve surface, including comparing real ship monitoring signals and finite element simulation signals of the air valve under a non-abrasion state and comparing finite element simulation signals of the air valve under different abrasion quantities of the valve surface.

The database system can also carry out query, entry, deletion and modification operations on material parameters, real ship monitoring signals and finite element simulation data.

(2) Air valve service state health assessment system (namely health assessment unit)

In order to further carry out health assessment on the service state of the air valve, a health assessment system of the service state of the air valve is built and a built low-speed machine air valve service prediction model is implanted, so that intelligent identification and diagnosis of the abrasion degradation state of the air valve in the running process of the low-speed machine are realized. If the abrasion loss of the current air valve is known, health evaluation is carried out on the current air valve through an abrasion threshold value; if an actual measurement signal of the operation state of the low-speed engine air valve is known, key characteristic values such as time domain characteristics, frequency domain characteristics, time-frequency domain-information entropy, exhaust amplitude, air-closing amplitude, impact amplitude and explosion amplitude and a signal map are obtained through signal processing software, then the wear degradation state of the air valve is intelligently diagnosed based on a low-speed engine air valve service prediction model, namely the wear of the current air valve is obtained through backward extrapolation of a BP neural network relation model between the wear of the air valve and the characteristic values of the rotating speed and vibration signals, and then the wear degradation state of the air valve is subjected to health assessment. When the valve abrasion amount exceeds a certain threshold value, the current abrasion deterioration state of the valve is considered to be serious, and an early warning is further sent out and an active forced scrapping criterion is made so as to avoid major faults of the diesel engine.

The specific construction process of the low-speed engine air valve service prediction model (namely, the wear analysis model) is as follows:

(1) establishment and correction of dynamic coupling model of fluid-solid-structure field in gas valve service process

As shown in fig. 6, according to actual gas valve parameters, a three-dimensional model of a cylinder cover, a cylinder body and a gas valve of the diesel engine is constructed, and then the three-dimensional model is introduced into ANSYS software, and a transient structure and a FLUENT module are adopted for fluid-solid coupling. Setting a gas flow equation, an oil injection parameter and a combustion equation in a fluid field, and calculating the pressure and the temperature in a cylinder; setting the motion process of the air valve and the constraint of the cylinder cover in a structural field, and carrying out vibration analysis and calculation; the interface of the fluid and the solid is a channel for mutual transmission of fluid-solid fields (including temperature, pressure, displacement and the like), and the interface is of a coupling type, so that real-time interactive calculation is realized; establishing a dynamic coupling model of a fluid-solid-structure field in the service process of the air valve, analyzing the cycle dynamic response rule of the thermal field of the air valve, and acquiring the change rule of the pressure in the cylinder and the vibration signal of the cylinder cover along with the crank angle. And correcting the fluid-solid-structure field dynamic coupling model according to the actually measured cylinder cover vibration signal, so that the characteristic amplitudes of the in-cylinder pressure, the cylinder cover vibration signal diagram, the exhaust amplitude, the gas closing amplitude, the impact amplitude, the explosion amplitude and the like obtained by simulation are close to or identical with the actually measured result.

Constructing three-dimensional models of a cylinder cover, a cylinder body and an air valve and endowing the three-dimensional models with material properties; setting a gas flow equation, an oil injection parameter and a combustion equation in a fluid field, and calculating the pressure and the temperature in a cylinder; setting the motion process of the air valve and the constraint of the cylinder cover in a structural field, and carrying out vibration analysis and calculation; the interface of the fluid and the solid is a channel for mutual transmission of fluid-solid fields (including temperature, pressure, displacement and the like), and the interface is of a coupling type, so that real-time interactive calculation is realized; and establishing a dynamic coupling model of the fluid-solid-structure field in the service process of the gas valve.

(2) Cylinder cover vibration signal of simulation air valve under different wearing and degrading states

And constructing air valves with different wear states based on the corrected fluid-solid-structure field dynamic coupling model, as shown in fig. 7, establishing fluid-solid-structure field dynamic coupling models of the air valves in different wear states, and simulating to obtain cylinder cover vibration signals of the air valves in different wear degradation states at different rotating speeds. Air valves with different wear degradation states mean that the valve faces of the air valves have different wear depths, the disc bottoms have different wear depths, and the disc bottoms and the valve faces both have different wear depths.

Low-speed engine air valve service prediction model (namely wear analysis model)

In order to realize vibration signal prediction in the air valve service process, a low-speed machine air valve service prediction model needs to be established, namely a digitalized mapping relation between the air valve surface abrasion loss and the disc bottom abrasion loss and the characteristic values of the rotating speed and the vibration signals. The vibration signal characteristic values comprise amplitudes corresponding to four waveforms of gas discharge, gas closing, impact and explosion in a time domain characteristic, kurtosis, a waveform factor, a peak factor, a pulse factor and a margin factor in a dimensionless characteristic, and a gravity center frequency, a mean square frequency and a power spectrum entropy, a singular spectrum entropy and an energy entropy in a time-frequency domain-information entropy.

The low-speed engine air valve service prediction model is established based on finite element simulation results of air valves under different rotation speeds in different wear degradation states. Firstly, extracting characteristic parameters such as a quasi-periodic signal time domain, a frequency domain, a time-frequency domain-information entropy and the like in a cylinder cover vibration signal based on finite element simulation results of different abrasion depths of a valve surface of the air valve, analyzing the relationship between the abrasion loss of the valve surface of the air valve and time domain characteristics and dimensionless characteristics, and selecting an item with a remarkable relationship in the time domain characteristics and the dimensionless characteristics as a key factor influencing the abrasion loss of the valve surface of the air valve; similarly, on the basis of finite element simulation results of different abrasion depths of the air valve plate bottom, characteristic parameters such as a quasi-periodic signal time domain, a frequency domain, a time-frequency domain-information entropy and the like in a cylinder cover vibration signal are extracted, the relation between the abrasion loss of the air valve plate bottom and time domain characteristics and dimensionless characteristics is analyzed, and items with obvious relation in the time domain characteristics and the dimensionless characteristics are selected as key factors influencing the abrasion loss of the air valve plate bottom. Further, based on finite element simulation data of different abrasion depths of the valve surface and the disc bottom of the air valve, time domain characteristics and dimensionless characteristics of a cylinder cover vibration signal of the air valve in the abrasion state at different rotating speeds are analyzed, key factors and rotating speeds in the established time domain characteristics and dimensionless characteristics are used as input variables, the valve abrasion amount is used as an output variable to construct a BP neural network model, digital mapping relations between the valve surface abrasion amount and the disc bottom abrasion amount of the air valve and characteristic values of the rotating speed and the vibration signal are respectively established, and theoretical guidance and technical reference are further provided for the abrasion amount prediction of the air valve in different abrasion degradation states of the air valve in the service process of the low-speed machine.

Finally, it should be noted that the above embodiments are only used for illustrating the technical solutions of the present invention and not for limiting the technical solutions, and those skilled in the art should understand that modifications or equivalent substitutions can be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions, and all that should be covered by the claims of the present invention.

Claims (10)

1. Marine low-speed diesel engine air valve service health intelligent diagnostic system which characterized in that includes:

the acquisition unit is used for acquiring current operation data of the low-speed diesel engine air valve, wherein the current operation data comprises a current crank angle signal, a current rotating speed signal and a current cylinder cover vibration signal;

the signal processing unit is used for extracting and processing the characteristics of the current cylinder cover vibration signal and extracting and obtaining key characteristic parameters which are significant in relation to the air valve abrasion in the current cylinder cover vibration signal and a corresponding signal map;

the health evaluation unit is provided with a trained wear analysis model between the wear amount and the signal characteristic parameters, and the trained wear analysis model between the wear amount and the signal characteristic parameters is used for indicating the corresponding mapping relation between the key characteristic parameters of the cylinder cover vibration signals and the wear amount of the low-speed diesel engine air valve at different running rotating speeds; the health evaluation unit is used for determining the current abrasion loss of the low-speed diesel engine air valve according to the current rotating speed signal and the key characteristic parameter of the current cylinder cover vibration signal through an abrasion analysis model between the abrasion loss and the signal characteristic parameter, and diagnosing the health state of the low-speed diesel engine air valve according to the current abrasion loss;

and the database management unit is used for storing the attribute parameters, the operation data, the characteristic parameters and the signal spectrum of the cylinder cover vibration signals and the health state diagnosis result of the low-speed diesel engine air valve.

2. The intelligent diagnosis system for marine low-speed diesel engine gas valve service health according to claim 1, wherein key characteristic parameters of the cylinder head vibration signal comprise one or more of time domain characteristics, frequency domain characteristics and time-frequency domain-information entropy; the time domain features comprise dimensional features and dimensionless features, the dimensional features comprise exhaust amplitude, closed gas amplitude, impact amplitude and explosion amplitude, and the dimensionless features comprise kurtosis, a form factor, a peak factor, a pulse factor and a margin factor; the frequency domain features comprise center of gravity frequency and mean square frequency; the time-frequency domain-information entropy comprises power spectrum entropy, singular spectrum entropy and energy entropy.

3. The intelligent diagnosis system for the service health of the gas valve of the marine low-speed diesel engine according to claim 1, wherein the wear analysis model is obtained by training in the following way:

s1, constructing a three-dimensional model of a cylinder cover, a cylinder body and a wear-free air valve, then carrying out fluid-solid coupling, constructing a fluid-solid-structure field dynamic coupling model of the air valve service process, and simulating and analyzing the cycle dynamic response rule of an air valve thermal field to obtain the change rule of the pressure in the cylinder and the vibration signal of the cylinder cover along with the crank rotation angle;

s2, correcting the fluid-solid-structure field dynamic coupling model according to the actually measured cylinder cover vibration signal, crank angle signal and in-cylinder pressure signal of the non-wear air valve, so that the difference value between the characteristic amplitude obtained by simulation and the actually measured result is smaller than a preset error value; the characteristic amplitude comprises in-cylinder pressure, exhaust amplitude, closed gas amplitude, impact amplitude and explosion amplitude;

s3, simulating and constructing air valves with different wear states based on the corrected fluid-solid-structure field dynamic coupling model, and establishing fluid-solid-structure field dynamic coupling models of the air valves in different wear states for simulating and analyzing cylinder cover vibration signals of the air valves at different rotating speeds and different wear states; the wear state includes an amount of wear;

s4, extracting characteristic parameters in a cylinder cover vibration signal in the simulation result of S3, performing abrasion correlation analysis, and selecting characteristic items with remarkable relations in the characteristic parameters according to a preset standard as key characteristic parameters influencing the abrasion of the air valve;

and S5, constructing and training a BP neural network model by taking the rotating speed and the selected key characteristic parameters as input variables and the abrasion loss as output variables to obtain an abrasion analysis model between the trained abrasion loss and the signal characteristic parameters.

4. The intelligent diagnosis system for marine low-speed diesel engine valve service health according to claim 3, wherein in the step S1, the process of constructing the fluid-solid-structure field dynamic coupling finite element model of the valve service process includes: constructing three-dimensional models of a cylinder cover, a cylinder body and an air valve and endowing the three-dimensional models with material properties; setting a gas flow equation, an oil injection parameter and a combustion equation in a fluid field, and calculating the pressure and the temperature in a cylinder; setting the motion process of the air valve and the constraint of the cylinder cover in a structural field, and carrying out vibration analysis and calculation; setting the interface of the fluid and the solid as a channel for mutual transmission of a fluid-solid field, wherein the interface is of a coupling type, so as to realize real-time interactive calculation; and establishing a dynamic coupling model of the fluid-solid-structure field in the service process of the gas valve.

5. The intelligent diagnosis system for the service health of the gas valve of the marine low-speed diesel engine according to claim 4, wherein:

in S3, the wear state further includes a wear portion; the wear part comprises a valve surface and a disc bottom;

in the step S4, when performing wear correlation analysis, the relationship between the valve face wear amount and the characteristic parameter, and the relationship between the disc bottom wear amount and the characteristic parameter are respectively analyzed; respectively selecting characteristic items with obvious relations in the characteristic parameters according to a preset standard as key characteristic parameters influencing the valve surface abrasion loss and the disc bottom abrasion loss;

in the step S5, when the BP neural network model is constructed, a digitized mapping relationship between the valve face wear amount and the rotational speed and between the cylinder head vibration signal characteristic parameters, and a digitized mapping relationship between the disc bottom wear amount and the rotational speed and between the disc bottom vibration signal characteristic parameters are respectively established.

6. The intelligent diagnosis system for the service health of the gas valve of the marine low-speed diesel engine according to claim 1, which is characterized in that: the wear amounts include a valve face wear amount and a disc bottom wear amount.

7. The intelligent diagnosis system for the service health of the gas valve of the marine low-speed diesel engine according to claim 6, wherein: when the health evaluation unit diagnoses the health state of the low-speed diesel engine air valve, if the valve surface abrasion loss or the disc bottom abrasion loss exceeds the corresponding early warning value, an early warning is sent out.

8. The marine low-speed diesel engine gas valve service health intelligent diagnosis system of claim 1, characterized in that: the attribute parameters comprise the model of the diesel engine, the material of the air valve and the corresponding performance parameters and the motion parameters of the piston.

9. The marine low-speed diesel engine gas valve service health intelligent diagnosis system of claim 8, characterized in that: the performance parameters comprise mass density, specific heat, Young modulus and stress strain; the piston motion parameters comprise the crankshaft radius, the length of a connecting rod and the rotating speed of a main shaft; the characteristic parameters and the signal spectrum of the cylinder cover vibration signal comprise a real ship detection signal and a finite element simulation signal.

10. The intelligent diagnosis method for the service health of the gas valve of the marine low-speed diesel engine is characterized by adopting the intelligent diagnosis system for the service health of the gas valve of the marine low-speed diesel engine according to claim 1 to execute processing, and comprises the following steps:

acquiring current operation data of a low-speed diesel engine air valve through an acquisition unit, wherein the current operation data comprises a current crank angle signal, a current rotating speed signal and a current cylinder cover vibration signal;

the signal processing unit is used for extracting the characteristics of the current cylinder cover vibration signal to obtain key characteristic parameters which are significant in relation to the air valve abrasion in the current cylinder cover vibration signal and a corresponding signal map;

and the health evaluation unit determines the current wear loss of the low-speed diesel engine air valve according to the current rotating speed signal and the key characteristic parameters of the current cylinder cover vibration signal through a wear analysis model between the wear loss and the signal characteristic parameters, and diagnoses the health state of the low-speed diesel engine air valve according to the current wear loss.

Images