CN113806860B - Fault feature extraction system, method, storage medium and equipment based on simulation - Google Patents
Fault feature extraction system, method, storage medium and equipment based on simulation Download PDFInfo
- Publication number
- CN113806860B CN113806860B CN202111014924.3A CN202111014924A CN113806860B CN 113806860 B CN113806860 B CN 113806860B CN 202111014924 A CN202111014924 A CN 202111014924A CN 113806860 B CN113806860 B CN 113806860B
- Authority
- CN
- China
- Prior art keywords
- fault
- simulation
- extraction
- module
- model
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/15—Vehicle, aircraft or watercraft design
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
- F02K—JET-PROPULSION PLANTS
- F02K9/00—Rocket-engine plants, i.e. plants carrying both fuel and oxidant therefor; Control thereof
- F02K9/96—Rocket-engine plants, i.e. plants carrying both fuel and oxidant therefor; Control thereof characterised by specially adapted arrangements for testing or measuring
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Geometry (AREA)
- Evolutionary Computation (AREA)
- Computer Hardware Design (AREA)
- Data Mining & Analysis (AREA)
- Chemical & Material Sciences (AREA)
- Mechanical Engineering (AREA)
- Mathematical Optimization (AREA)
- Aviation & Aerospace Engineering (AREA)
- Automation & Control Theory (AREA)
- Mathematical Analysis (AREA)
- Combustion & Propulsion (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a fault feature extraction system, a fault feature extraction method, a storage medium and fault feature extraction equipment based on simulation, which are used for solving the problems that the current engine fault diagnosis depends on a large amount of fault data, the real-time performance is relatively insufficient, and the dynamic fault development detection and diagnosis are difficult to meet. The system comprises an engine system simulation model generation module, an extraction condition setting module, a model simulation solving module, a post-simulation result processing module, a feature extraction module and a fault feature library. The method comprises the following steps: 1. acquiring a normal simulation model and a fault simulation model of an engine system and performing simulation solution; 2. extracting real-time data in simulation or result data after simulation; 3. carrying out dynamic characteristic data extraction or steady-state characteristic data extraction on the extracted data in the step 2; 4. processing the data extracted in the step 3 to obtain a dynamic characteristic extraction result and a steady-state characteristic extraction result; 5. and processing and storing the extraction result.
Description
Technical Field
The invention relates to simulation research of faults of a liquid rocket engine, in particular to a simulation-based liquid rocket engine fault feature extraction system, a simulation-based liquid rocket engine fault feature extraction method, a storage medium and equipment.
Background
The liquid rocket engine works under the extremely environment conditions of high temperature, high pressure and strong corrosion, on one hand, the duration of the fault occurrence and development process is very short, but the influence results are serious; on the other hand, the engine type is different, the structure, the working condition and the fault expression form are different, and the fault mode is completely different even for the same type or the same engine, so that the duplication and reproduction of the engine fault are difficult to realize by a physical simulation and test method. Therefore, compared with sample data with huge engine test and flight data, the size and proportion of fault samples are very small, which also results in that in the condition of lacking sufficient sample data and lacking sufficient fault modes and characteristic descriptions thereof, judging criteria suitable for fault detection and diagnosis of different models of engines are difficult to obtain.
In recent years, with the rapid development of computer simulation technology, model-based fault simulation research of the engine working process has become an effective way for analyzing the engine fault mechanism and the transient process characteristics thereof and obtaining the engine fault mode characteristics, diagnosis knowledge and fault sample data.
The liquid rocket engine is a time-varying system with a complex structure, mainly shows time-varying performance (difference between normal working modes and conversion between normal working modes and fault working modes) of a system, and the range of observation parameter values changes along with the change of the working modes and the operation time, so that great difficulty exists in predetermining time-varying fault diagnosis criteria of the system, a principal component feature method, a BP neural network method, a fuzzy information optimization processing-based method, a mutual information entropy-based method, a wavelet analysis-based method and the like exist in the conventional feature extraction method, the methods depend on a large amount of fault data to a certain extent, the instantaneity is relatively insufficient, and the attention to the fault transition process characteristics is relatively less, so that the early stage of fault occurrence, particularly the detection and diagnosis of the dynamic process fault are difficult to meet.
Disclosure of Invention
The invention aims to solve the problems that the fault diagnosis of a liquid rocket engine depends on a large amount of fault data, the real-time performance is relatively insufficient, and the detection and diagnosis in the dynamic process of fault development are difficult to meet at present, and provides a simulation-based liquid rocket engine fault feature extraction system, a simulation-based liquid rocket engine fault feature extraction method, a simulation-based storage medium and simulation-based liquid rocket engine fault feature extraction equipment.
The technical scheme provided by the invention is as follows:
the fault characteristic extraction system based on simulation is characterized in that:
the system comprises an engine system simulation model generation module, an extraction condition setting module, a model simulation solving module, a post-simulation result processing module, a feature extraction module and a fault feature library;
the engine system simulation model generation module is used for constructing an engine system normal simulation model and an engine system fault simulation model and outputting the engine system normal simulation model and the engine system fault simulation model to the model simulation solving module;
the extraction condition setting module is used for setting feature extraction conditions and is respectively connected with the model simulation solving module and the feature extraction module;
one output of the model simulation solving module is connected with one input of the feature extraction module, the other output of the model simulation solving module is connected with the input of the simulated result processing module, and the output of the simulated result processing module is connected with the other input of the feature extraction module; the model simulation solving module is used for compiling and solving the normal simulation model of the engine system and the fault simulation model of the engine system, and outputting real-time data in simulation or simulated result data processed by the simulated result processing module to the feature extraction module according to the feature extraction conditions;
the feature extraction module comprises a real-time data receiving module in simulation, a result data reading module after simulation, a dynamic feature data processing module, a steady-state feature data processing module and an extraction result processing module;
the input of the real-time data receiving module in the simulation is connected with one output of the model simulation solving module, and the real-time data receiving module in the simulation is used for receiving and storing the real-time data in the simulation;
the input of the simulated result data reading module is connected with the output of the simulated result processing module and is used for reading and storing the simulated result data;
the dynamic characteristic data processing module is used for receiving dynamic characteristic data in the real-time data receiving module in simulation or the simulated result data reading module according to the characteristic extraction conditions, carrying out dynamic amplitude evaluation and derivative evaluation on the dynamic characteristic data, and judging whether the dynamic characteristic data reach a steady state or not according to an evaluation result; one output of the dynamic characteristic data processing module is connected with one input of the extraction result processing module, and the other output of the dynamic characteristic data processing module is connected with one input of the steady-state characteristic data processing module;
one input of the steady-state characteristic data processing module receives the dynamic characteristic data reaching steady state output by the dynamic characteristic data processing module; the other input receives steady-state characteristic data of a real-time data receiving module in simulation or a simulated result data reading module according to characteristic extraction conditions, and carries out steady-state amplitude evaluation on the steady-state characteristic data and the dynamic characteristic data reaching steady state, and the output of a steady-state characteristic data processing module is connected with the other input of the extraction result processing module;
the extraction result processing module is used for generating a fault feature table from the obtained dynamic feature extraction result and steady-state feature extraction result and outputting the fault feature table to the fault feature library;
the fault feature library is used for storing the fault feature table generated by the feature extraction module.
Further, the engine system simulation model generation module comprises a fault model library, a model editing module, a fault mode library, a fault mode management module and a parameter injection module;
the fault model library is used for storing fault models of engine system components, and the output of the fault model library is connected with one input of the model editing module;
the model editing module is used for constructing a normal simulation model of the engine system and a fault simulation model of the engine system, and one output of the model editing module is connected with one input of the fault mode management module;
the other input of the fault mode management module is connected with the fault mode library and is used for reading information in the fault mode library, editing values and injection conditions of fault parameters and outputting the values and the injection conditions to the parameter injection module;
the fault mode library is used for storing fault modes;
the parameter injection module is used for writing the fault parameter values in the fault mode library read by the fault mode management module into the engine system simulation model, and outputting the fault parameter values to the model editing module to generate the engine system fault simulation model.
Further, the fault model is divided into three layers of a subsystem, a component and a component according to the physical topological structure of the engine system, each component or component comprises parameters, variables and mathematical equations capable of describing the behavior of the component or component, and parameters related to the fault characteristics of the component or component are added to support subsequent fault parameter injection;
the fault mode comprises standard fault mode information and information of a mapping relation between the fault mode and the model;
the standard fault mode information comprises a fault mode name, a fault reason, a fault result, fault influence analysis, fault severity and fault occurrence probability;
the information of the mapping relation between the fault mode and the model comprises a fault parameter name and a fault mode triggering condition.
Further, the feature extraction conditions comprise an extraction feature variable, an extraction start schedule, an extraction frequency, an extraction number and an extraction mode;
the extracted feature variable condition is to extract dynamic feature data or extract steady-state feature data;
the extraction mode condition is real-time data extraction in simulation or result data extraction after simulation.
The invention also provides a fault feature extraction method based on simulation, which is characterized by comprising the following steps of:
step 1, loading a fault model library, and constructing a normal simulation model of an engine system through a model editing module based on the model library;
step 2, loading a fault mode library, using a fault mode management module to read information in the fault mode library, editing values and injection conditions of fault parameters, and using a parameter injection module to perform fault parameter injection on a normal model of the simulation engine system;
step 3, generating a corresponding engine system fault simulation model by using a model editing module according to the parameters injected in the step 2, and respectively carrying out simulation solution on the engine system normal simulation model and the engine system fault simulation model by using a model simulation solution module;
step 4, setting feature extraction conditions according to an extraction condition setting module, wherein the feature extraction conditions comprise an extraction feature variable, an extraction starting schedule, an extraction frequency, an extraction number and an extraction mode; the extracted feature variable condition is to extract dynamic feature data or extract steady-state feature data; the extraction mode condition is real-time data extraction in simulation or result data extraction after simulation;
judging an extraction mode according to the feature extraction conditions, and if real-time data extraction is performed in the simulation, entering a step 5;
if the simulated result data is extracted, the simulated result data is transmitted to a simulated result processing module for screening processing, and is output to a simulated result data reading module in a characteristic extraction module for reading and storing, and the step 6 is entered;
step 5, real-time data in the simulation in the process of simulating and solving the normal simulation model and the fault simulation model of the engine system are extracted in real time according to the characteristic extraction conditions in the step 4, and are transmitted to a real-time data receiving module in the simulation to be received and stored;
step 6, judging feature variables extracted from real-time data in simulation or simulation result data according to the feature extraction conditions in the step 4, if the dynamic feature data are extracted, entering the step 7, and if the steady-state feature data are extracted, entering the step 9;
step 7, extracting dynamic characteristic data, and carrying out dynamic amplitude evaluation and derivative evaluation on the dynamic characteristic data by using a dynamic characteristic data processing module;
step 8, carrying out steady state judgment according to the amplitude evaluation result and the derivative evaluation result in the step 7,
if a steady state is reached, step 9 is entered,
if the steady state is not reached, carrying out moving average treatment on the amplitude value evaluation result and the derivative evaluation result in the step 7 to obtain a dynamic characteristic extraction result, and entering a step 10;
step 9, using a steady-state characteristic data processing module to evaluate steady-state amplitude values of the steady-state characteristic data extracted in the step 6 and the dynamic characteristic data which reach a steady state after steady-state judgment in the step 8, and performing the moving average processing on the steady-state amplitude value evaluation result to obtain a steady-state characteristic extraction result;
and step 10, using an extraction result processing module, merging and processing according to the dynamic characteristic extraction result and the steady-state characteristic extraction result and the change rule of the obtained fault characteristic variable to generate a fault characteristic table, and storing the fault characteristic table in a fault characteristic library.
Further, the bases for steady state judgment in the step 8 include:
a. if the derivative evaluation result is continuously reduced in multiple steps, the amplitude evaluation result is monotonously changed, and then a steady state is achieved when the relative change of two continuous steps of derivatives is less than 1%;
b. if the derivative evaluation result shows positive-to-negative or negative-to-positive variation, reaching a steady state when the relative variation of the two continuous derivative evaluation results is less than 1%;
c. if the derivative estimate exhibits an oscillatory change, then steady state is reached when the continuous two-step value estimate changes by less than 0.01%.
The steady state judgment provided by the invention judges whether the data reaches the steady state or not by combining the dynamic amplitude evaluation result and the derivative evaluation result, the judgment mode is comprehensive, and the judgment result is more accurate.
Further, the fault parameter injection is single or batch injection.
Further, in step 10, the processing of the dynamic feature extraction result and the steady-state feature extraction result specifically includes:
the processing mode of the dynamic feature extraction result is that the positive and negative deviations of the amplitude evaluation results and the derivative evaluation results of all feature variables are recorded in sequence according to the time sequence aiming at each fault mode; the positive and negative deviation is that the difference between the amplitude evaluation result or the derivative evaluation result of the characteristic variable of the fault working condition and the normal working condition is a positive value or a negative value;
the steady-state feature extraction result processing mode is that for each fault mode, the amplitudes of all feature variables are compared in sequence from the positive deviation and the negative deviation and the influence sequencing so as to describe the change rule of each feature variable in different fault modes;
the positive and negative deviation is that the difference between the characteristic variable amplitude evaluation results of the fault working condition and the normal working condition is a positive value or a negative value;
and the influence sorting is the sorting of absolute difference values of the characteristic variable amplitude evaluation results of the fault working condition and the normal working condition.
The present invention also provides a computer-readable storage medium having stored thereon a computer program, characterized in that: the computer program, when executed by a processor, implements the steps of the simulation-based fault feature extraction method described above.
The invention also provides a computer device comprising a processor, a memory connected with the processor, and a computer program capable of running on the processor, which is characterized in that: the processor, when executing the computer program, implements the steps of the simulation-based fault feature extraction method described above.
The invention has the beneficial effects that:
1. according to the simulation-based fault feature extraction method provided by the invention, the fault features in the liquid rocket engine system are extracted based on a simulation mode, the extraction mode can be real-time data extraction in the simulation or result data extraction after the simulation, the real-time data extraction in the simulation can meet the application scene with high real-time requirements, the fault information in the engine system can be acquired more timely, serious influence on the engine system is avoided, and the dynamic process of fault transition can be acquired through real-time simulation data, so that the analysis of faults is facilitated; the simulated result data extraction mode can meet the application scene of fault offline processing.
2. The invention can realize two characteristic extraction modes aiming at real-time data in simulation and simulated result data: (1) individual steady state feature extraction and processing; (2) combining dynamic characteristics with steady-state characteristics; the steady-state feature extraction mode is beneficial to finding fault features, but the combination of the dynamic feature extraction mode and the steady-state feature extraction mode is beneficial to locating the fault features. The mode of combining the dynamic characteristics and the steady-state characteristics covers the dynamic characteristics of a common engine system, and the dynamic amplitude evaluation and the derivative evaluation are carried out to judge whether the engine system reaches a steady state or not, so that the automatic switching from the dynamic characteristics to the steady-state characteristics is realized.
3. According to the fault feature extraction system, single or multiple fault modes are managed and fault injection is achieved according to different liquid rocket engine models, dynamic features and steady-state features of the system are extracted based on simulation, so that different fault detection and diagnosis requirements are met, and instantaneity and versatility of fault feature extraction are improved.
4. The fault feature extraction result processing method comprises two processing modes of a dynamic feature extraction result and a steady-state feature extraction result, and the fault feature library generated by the two processing modes can meet different fault detection and diagnosis application requirements.
Drawings
FIG. 1 is a schematic diagram of a simulation-based fault feature extraction system of the present invention;
FIG. 2 is a schematic diagram of a feature extraction module according to the present invention;
FIG. 3 is a flow chart of a simulation-based fault feature extraction method of the present invention.
Detailed Description
As shown in fig. 1 and fig. 2, the present embodiment provides a simulation-based liquid rocket engine fault feature extraction system, which includes an engine system simulation model generation module, an extraction condition setting module, a model simulation solving module, a post-simulation result processing module, a feature extraction module and a fault feature library;
the engine system simulation model generation module comprises a fault model library, a model editing module, a fault mode library, a fault mode management module and a parameter injection module; the engine system simulation model generation module is used for constructing an engine system normal simulation model and an engine system fault simulation model and outputting the models to the model simulation solving module.
See China patent with publication number CN105260555A for the specific content of an engine system simulation model generation module, and discloses a fault injection system based on a Modelica model and a method thereof; and Chinese patent with publication number CN112799900A, a model fault injection method and system based on Modelica are disclosed.
The fault model library is used for storing a fault model of an engine system component, the fault model is built based on a multi-domain unified modeling language Modelica, system simulation of the electro-mechanical and hydraulic control multi-domain is supported, and the output of the fault model library is connected with one input of the model editing module;
the fault model is divided into three layers of a subsystem, a component and a component according to the physical topological structure of the engine system, each component or component comprises parameters, variables and mathematical equations capable of describing the behavior of the component or component, and parameters related to the fault characteristics of the component or component are added to support subsequent fault parameter injection.
One output of the model editing module is connected with one input of the fault mode management module and is used for constructing a normal simulation model of the engine system and a fault simulation model of the engine system;
the other input of the fault mode management module is connected with the fault mode library and is used for reading information in the fault mode library, editing values and injection conditions of fault parameters and outputting the values and the injection conditions to the parameter injection module;
the fault mode library is used for storing fault modes; the fault mode comprises standard fault mode information and information of a mapping relation between the fault mode and the model;
the standard fault mode information comprises a fault mode name, a fault reason, a fault result, fault influence analysis, fault severity and fault occurrence probability;
the information of the mapping relation between the fault mode and the model comprises a fault parameter name and a fault mode triggering condition.
The parameter injection module receives the value and injection condition of the edited fault parameters output by the fault mode management module, finds out the corresponding fault parameter identification in the model, writes the value of the fault parameters into the engine system simulation model, completes single or batch fault parameter injection, and outputs the fault parameter injection to the model editing module to generate the engine system fault simulation model.
The extraction condition setting module is used for setting feature extraction conditions; the feature extraction conditions comprise extraction feature variables, an extraction starting schedule, extraction frequency, extraction number and extraction mode; the extracted feature variable condition is to extract dynamic feature data or steady-state feature data; the extraction mode condition is real-time data extraction in simulation or result data extraction after simulation.
The model simulation solving module is used for compiling and solving a normal simulation model of the engine system and a fault simulation model of the engine system to solve unknown variables in the system; when the feature extraction condition meets the first condition, real-time data in the simulation of the model simulation solving module is output to one input of the feature extraction module; when the feature extraction condition meets the second condition, the simulated result data of the model simulation solving module is output to the simulated result processing module;
the first condition is real-time data extraction in simulation, and the second condition is result data extraction after simulation.
The post-simulation result processing module is used for screening post-simulation result data and outputting the result data to the other input of the feature extraction module;
the feature extraction module comprises a real-time data receiving module in simulation, a result data reading module after simulation, a dynamic feature data processing module, a steady-state feature data processing module and an extraction result processing module;
the real-time data receiving module in the simulation is used for receiving and storing real-time data in the simulation;
the simulated result data reading module is used for reading and storing simulated result data;
the dynamic characteristic data processing module is used for receiving dynamic characteristic data in the real-time data receiving module in simulation or the simulated result data reading module when the characteristic extraction condition meets a third condition, carrying out dynamic amplitude evaluation and derivative evaluation on the dynamic characteristic data, judging whether the dynamic characteristic data reach a steady state or not according to an evaluation result, and transmitting the dynamic characteristic data to one input of the steady state characteristic data processing module through one output of the dynamic characteristic data processing module if the dynamic characteristic data reach the steady state; if the steady state is not reached, the dynamic characteristic data is processed by the moving average and then is transmitted to one input of the extraction result processing module through the other output of the dynamic characteristic data processing module.
The steady-state characteristic data processing module is used for receiving steady-state characteristic data of the real-time data receiving module in simulation or the simulated result data reading module when the characteristic extraction condition meets a fourth condition, carrying out steady-state amplitude evaluation on the steady-state characteristic data and the dynamic characteristic data reaching a steady state, and outputting a result to the other input of the extraction result processing module;
the third condition is to perform dynamic feature extraction, and the fourth condition is not to perform dynamic feature extraction;
the extraction result processing module is used for generating a fault characteristic table from the acquired processing results of the dynamic characteristic data processing module and the steady-state characteristic data processing module and outputting the fault characteristic table to the fault characteristic library;
the fault feature library is used for storing the fault feature table generated by the feature extraction module to form a set of normalized fault feature library.
Referring to fig. 3, the present embodiment discloses a simulated fault feature extraction method, which includes the following steps:
step 1, loading a fault model library, and constructing a normal simulation model of an engine system through a model editing module based on the model library;
step 2, loading a fault mode library, using a fault mode management module to read information in the fault mode library, editing values and injection conditions of fault parameters, and using a parameter injection module to perform fault parameter injection on a normal model of the simulation engine system, wherein the fault parameter injection is single or batch injection;
step 3, generating a corresponding engine system fault simulation model by using a model editing module according to the parameters injected in the step 2, and respectively solving the engine system normal simulation model and the engine system fault simulation model by using a model simulation solving module;
step 4, setting feature extraction conditions according to an extraction condition setting module, wherein the feature extraction conditions comprise extraction feature variables, an extraction starting schedule, extraction frequencies, extraction numbers and extraction modes; the extracted feature variable condition is to extract dynamic feature data or steady-state feature data; the extraction mode condition is real-time data extraction in simulation or extraction of result data after simulation;
judging an extraction mode according to the feature extraction conditions, and if real-time data extraction is performed in the simulation, entering a step 5;
if the simulated result data is extracted, the simulated result data is transmitted to a simulated result processing module for screening treatment and is output to a simulated result data reading module in a characteristic extraction module for reading and storing, and the step 6 is entered;
step 5, real-time data in the simulation in the process of simulating and solving the normal simulation model of the engine system and the fault simulation model of the engine system are extracted in real time according to the characteristic extraction conditions in the step 4, and are transmitted to a real-time data receiving module in the simulation to be received and stored;
step 6, judging real-time data in simulation or result data after simulation to extract feature variables according to the feature extraction conditions in the step 4, if the dynamic feature data are extracted, entering the step 7, and if the steady-state feature data are extracted, entering the step 9;
step 7, extracting dynamic characteristic data, carrying out dynamic amplitude evaluation on the dynamic characteristic data by using a dynamic characteristic data processing module, and carrying out derivative evaluation on a dynamic amplitude evaluation result;
and 8, carrying out steady state judgment according to the amplitude evaluation result and the derivative evaluation result, wherein the specific judgment basis is as follows:
a. if the derivative evaluation result is continuously reduced in multiple steps, the amplitude evaluation result is monotonously changed, and then a steady state is achieved when the relative change of two continuous steps of derivatives is less than 1%;
b. if the derivative evaluation result shows positive-to-negative or negative-to-positive variation, reaching a steady state when the relative variation of the two continuous derivative evaluation results is less than 1%;
c. if the derivative estimate exhibits an oscillatory change, then steady state is reached when the continuous two-step value estimate changes by less than 0.01%.
If the steady state is reached, the step 9 is entered;
if the steady state is not reached, carrying out moving average processing on the amplitude evaluation result and the derivative evaluation result corresponding to the feature data which does not reach the steady state in the step 7 to obtain a dynamic feature extraction result, and then entering the step 10;
step 9, using a steady-state characteristic data processing module to perform steady-state amplitude evaluation on the steady-state characteristic data extracted in the step 6 and the data of which the dynamic characteristic data extracted in the step 8 reach a steady state through steady-state judgment, and performing moving average processing on a steady-state amplitude evaluation result to obtain a steady-state characteristic extraction result;
and step 10, using an extraction result processing module, merging and processing according to the dynamic characteristic extraction result and the steady-state characteristic extraction result and the change rule of the obtained fault characteristic variable to generate a fault characteristic table, and storing the fault characteristic table in a fault characteristic library.
The processing modes of the dynamic characteristic extraction result and the steady-state characteristic extraction result are as follows:
the processing mode of the dynamic feature extraction result is that the positive and negative deviations of the amplitude evaluation results and the derivative evaluation results of all feature variables are recorded in sequence according to the time sequence aiming at each fault mode; the positive and negative deviation is that the difference between the amplitude evaluation result or the derivative evaluation result of the characteristic variable of the fault working condition and the normal working condition is a positive value or a negative value;
the steady-state feature extraction result processing mode is to compare the amplitudes of all feature variables in sequence from positive and negative deviation and influence sequencing aiming at each fault mode, so as to describe the change rule of each feature variable in different fault modes;
the positive and negative deviation is that the difference between the characteristic variable amplitude evaluation results of the fault working condition and the normal working condition is positive or negative;
the influence ordering is the absolute difference ordering of the characteristic variable amplitude evaluation results of the fault working condition and the normal working condition.
The embodiment of the present invention also provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the simulation-based fault feature extraction method described in the above embodiment; the storage medium may be a U-disk, a removable hard disk, a read-only memory (ROM), a random access memory (random access memory, RAM), or various media capable of storing program codes such as an optical disk.
The embodiment of the invention also provides a computer device, which comprises a processor, a memory connected with the processor, and a computer program capable of running on the processor, wherein the processor can realize the steps of the simulation-based fault feature extraction method described in the embodiment when executing the computer program; the computer device may be a notebook computer, a tablet computer, or a desktop computer, etc., and has the same beneficial effects as the computer device has a computer program that can implement the simulation-based fault feature extraction method in the above embodiment, and will not be described in detail in this embodiment.
Claims (10)
1. A simulation-based fault feature extraction system, characterized by:
the system comprises an engine system simulation model generation module, an extraction condition setting module, a model simulation solving module, a post-simulation result processing module, a feature extraction module and a fault feature library;
the engine system simulation model generation module is used for constructing an engine system normal simulation model and an engine system fault simulation model and outputting the engine system normal simulation model and the engine system fault simulation model to the model simulation solving module;
the extraction condition setting module is used for setting feature extraction conditions and is respectively connected with the model simulation solving module and the feature extraction module;
one output of the model simulation solving module is connected with one input of the feature extraction module, the other output of the model simulation solving module is connected with the input of the simulated result processing module, and the output of the simulated result processing module is connected with the other input of the feature extraction module; the model simulation solving module is used for compiling and solving the normal simulation model of the engine system and the fault simulation model of the engine system, and outputting real-time data in simulation or simulated result data processed by the simulated result processing module to the feature extraction module according to the feature extraction conditions;
the feature extraction module comprises a real-time data receiving module in simulation, a result data reading module after simulation, a dynamic feature data processing module, a steady-state feature data processing module and an extraction result processing module;
the input of the real-time data receiving module in the simulation is connected with one output of the model simulation solving module, and the real-time data receiving module in the simulation is used for receiving and storing the real-time data in the simulation;
the input of the simulated result data reading module is connected with the output of the simulated result processing module and is used for reading and storing the simulated result data;
the dynamic characteristic data processing module is used for receiving dynamic characteristic data in the real-time data receiving module in simulation or the result data reading module after simulation according to characteristic extraction conditions, carrying out dynamic amplitude evaluation and derivative evaluation on the dynamic characteristic data, and judging whether the dynamic characteristic data reach a steady state or not according to an evaluation result; one output of the dynamic characteristic data processing module is connected with one input of the extraction result processing module, and the other output of the dynamic characteristic data processing module is connected with one input of the steady-state characteristic data processing module;
one input of the steady-state characteristic data processing module receives steady-state dynamic characteristic data output by the dynamic characteristic data processing module, the other input receives steady-state characteristic data of the real-time data receiving module in simulation or the simulated result data reading module according to characteristic extraction conditions, steady-state amplitude evaluation is carried out on the steady-state characteristic data and the steady-state dynamic characteristic data, and the output of the steady-state characteristic data processing module is connected with the other input of the extraction result processing module;
the extraction result processing module is used for generating a fault feature table from the obtained dynamic feature extraction result and steady-state feature extraction result and outputting the fault feature table to the fault feature library;
the fault feature library is used for storing the fault feature table generated by the feature extraction module.
2. The simulation-based fault signature extraction system of claim 1, wherein: the engine system simulation model generation module comprises a fault model library, a model editing module, a fault mode library, a fault mode management module and a parameter injection module;
the fault model library is used for storing fault models of engine system components, and the output of the fault model library is connected with one input of the model editing module;
the model editing module is used for constructing a normal simulation model of the engine system and a fault simulation model of the engine system, and one output of the model editing module is connected with one input of the fault mode management module;
the other input of the fault mode management module is connected with the fault mode library and is used for reading information in the fault mode library, editing values and injection conditions of fault parameters and outputting the values and the injection conditions to the parameter injection module;
the fault mode library is used for storing fault modes;
the parameter injection module is used for writing the fault parameter values in the fault mode library read by the fault mode management module into the engine system simulation model, and outputting the fault parameter values to the model editing module to generate the engine system fault simulation model.
3. The simulation-based fault signature extraction system of claim 2, wherein:
the fault model is divided into three layers of a subsystem, a component and a component according to the physical topological structure of the engine system, each component or component comprises parameters, variables and mathematical equations capable of describing the behavior of the component or component, and parameters related to the fault characteristics of the component or component are added to support subsequent fault parameter injection;
the fault mode comprises standard fault mode information and information of a mapping relation between the fault mode and the model;
the standard fault mode information comprises a fault mode name, a fault reason, a fault result, fault influence analysis, fault severity and fault occurrence probability;
the information of the mapping relation between the fault mode and the model comprises a fault parameter name and a fault mode triggering condition.
4. A simulation-based fault signature extraction system as in claim 3 wherein: the characteristic extraction conditions comprise an extraction characteristic variable, an extraction starting schedule, an extraction frequency, an extraction number and an extraction mode;
the extracted feature variable condition is to extract dynamic feature data or extract steady-state feature data;
the extraction mode condition is real-time data extraction in simulation or result data extraction after simulation.
5. The fault characteristic extraction method based on simulation is characterized by comprising the following steps of:
step 1, loading a fault model library, and constructing a normal simulation model of an engine system through a model editing module based on the model library;
step 2, loading a fault mode library, using a fault mode management module to read information in the fault mode library, editing values and injection conditions of fault parameters, and using a parameter injection module to perform fault parameter injection on a normal model of the simulation engine system;
step 3, generating a corresponding engine system fault simulation model by using a model editing module according to the parameters injected in the step 2, and respectively carrying out simulation solution on the engine system normal simulation model and the engine system fault simulation model by using a model simulation solution module;
step 4, setting feature extraction conditions according to an extraction condition setting module, wherein the feature extraction conditions comprise an extraction feature variable, an extraction starting schedule, an extraction frequency, an extraction number and an extraction mode; the extracted feature variable condition is to extract dynamic feature data or extract steady-state feature data; the extraction mode condition is real-time data extraction in simulation or result data extraction after simulation;
judging an extraction mode according to the feature extraction conditions, and if real-time data extraction is performed in the simulation, entering a step 5;
if the simulated result data is extracted, the simulated result data is transmitted to a simulated result processing module for screening processing, and is output to a simulated result data reading module in a characteristic extraction module for reading and storing, and the step 6 is entered;
step 5, real-time data in the simulation in the process of simulating and solving the normal simulation model and the fault simulation model of the engine system are extracted in real time according to the characteristic extraction conditions in the step 4, and are transmitted to a real-time data receiving module in the simulation to be received and stored;
step 6, judging feature variables extracted from real-time data in simulation or simulation result data according to the feature extraction conditions in the step 4, if the dynamic feature data are extracted, entering the step 7, and if the steady-state feature data are extracted, entering the step 9;
step 7, extracting dynamic characteristic data, and carrying out dynamic amplitude evaluation and derivative evaluation on the dynamic characteristic data by using a dynamic characteristic data processing module;
step 8, carrying out steady state judgment according to the amplitude evaluation result and the derivative evaluation result in the step 7,
if a steady state is reached, step 9 is entered,
if the steady state is not reached, carrying out moving average treatment on the amplitude value evaluation result and the derivative evaluation result in the step 7 to obtain a dynamic characteristic extraction result, and entering a step 10;
step 9, using a steady-state characteristic data processing module to evaluate steady-state amplitude values of the steady-state characteristic data extracted in the step 6 and the dynamic characteristic data which reach a steady state after steady-state judgment in the step 8, and performing the moving average processing on the steady-state amplitude value evaluation result to obtain a steady-state characteristic extraction result;
and step 10, using an extraction result processing module, merging and processing according to the dynamic characteristic extraction result and the steady-state characteristic extraction result and the change rule of the obtained fault characteristic variable to generate a fault characteristic table, and storing the fault characteristic table in a fault characteristic library.
6. The simulation-based fault signature extraction method as recited in claim 5, wherein the basis for the steady state determination in step 8 includes:
a. if the derivative evaluation result is continuously reduced in multiple steps, the amplitude evaluation result is monotonously changed, and then a steady state is achieved when the relative change of two continuous steps of derivatives is less than 1%;
b. if the derivative evaluation result shows positive-to-negative or negative-to-positive variation, reaching a steady state when the relative variation of the two continuous derivative evaluation results is less than 1%;
c. if the derivative estimate exhibits an oscillatory change, then steady state is reached when the continuous two-step value estimate changes by less than 0.01%.
7. The simulation-based fault signature extraction method as claimed in claim 5 or 6, wherein: the fault parameter injection is single or batch injection.
8. The simulation-based fault feature extraction method according to claim 7, wherein in step 10, the dynamic feature extraction result and the steady state feature extraction result are processed, specifically:
the processing mode of the dynamic feature extraction result is that the positive and negative deviations of the amplitude evaluation results and the derivative evaluation results of all feature variables are recorded in sequence according to the time sequence aiming at each fault mode; the positive and negative deviation is that the difference between the amplitude evaluation result or the derivative evaluation result of the characteristic variable of the fault working condition and the normal working condition is a positive value or a negative value;
the steady-state feature extraction result processing mode is that for each fault mode, the amplitudes of all feature variables are compared in sequence from the positive deviation and the negative deviation and the influence sequencing so as to describe the change rule of each feature variable in different fault modes;
the positive and negative deviation is that the difference between the characteristic variable amplitude evaluation results of the fault working condition and the normal working condition is a positive value or a negative value;
and the influence sorting is the sorting of absolute difference values of the characteristic variable amplitude evaluation results of the fault working condition and the normal working condition.
9. A computer-readable storage medium having stored thereon a computer program, characterized by: the computer program, when executed by a processor, implements the steps of the simulation-based fault feature extraction method of any one of claims 5-8.
10. A computer device comprising a processor, a memory coupled to the processor, and a computer program executable on the processor, characterized by: the processor, when executing the computer program, implements the steps of the simulation-based fault feature extraction method of any one of claims 5-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111014924.3A CN113806860B (en) | 2021-08-31 | 2021-08-31 | Fault feature extraction system, method, storage medium and equipment based on simulation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111014924.3A CN113806860B (en) | 2021-08-31 | 2021-08-31 | Fault feature extraction system, method, storage medium and equipment based on simulation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113806860A CN113806860A (en) | 2021-12-17 |
| CN113806860B true CN113806860B (en) | 2023-08-08 |
Family
ID=78894436
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111014924.3A Active CN113806860B (en) | 2021-08-31 | 2021-08-31 | Fault feature extraction system, method, storage medium and equipment based on simulation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113806860B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114493635A (en) * | 2022-01-26 | 2022-05-13 | 山东浪潮工业互联网产业股份有限公司 | Tracing method, device and storage medium for engine parts |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106650149A (en) * | 2016-12-30 | 2017-05-10 | 苏州同元软控信息技术有限公司 | Modelica model-based runtime fault injection system and method |
| AU2020102819A4 (en) * | 2020-10-16 | 2020-12-10 | Beihang University | Digital-twin-driven method and system for fault diagnosis of subsea production system of offshore oil |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7369932B2 (en) * | 2006-05-04 | 2008-05-06 | Honeywell International, Inc. | System and method for turbine engine fault detection using discrete event system modeling |
| GB201501205D0 (en) * | 2015-01-26 | 2015-03-11 | Rolls Royce Plc | Open switch fault detection and identification in a two-level voltage source power converter |
-
2021
- 2021-08-31 CN CN202111014924.3A patent/CN113806860B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106650149A (en) * | 2016-12-30 | 2017-05-10 | 苏州同元软控信息技术有限公司 | Modelica model-based runtime fault injection system and method |
| AU2020102819A4 (en) * | 2020-10-16 | 2020-12-10 | Beihang University | Digital-twin-driven method and system for fault diagnosis of subsea production system of offshore oil |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113806860A (en) | 2021-12-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111459700B (en) | Equipment fault diagnosis method, diagnosis device, diagnosis equipment and storage medium | |
| CN112149316A (en) | Aero-engine residual life prediction method based on improved CNN model | |
| CN109859029A (en) | Abnormal application detection method, device, computer equipment and storage medium | |
| CN110688288A (en) | Automatic testing method, device, equipment and storage medium based on artificial intelligence | |
| CN108279885B (en) | Method and device for integrating software of multiple model codes | |
| CN111125052B (en) | Big data intelligent modeling system and method based on dynamic metadata | |
| CN104268064A (en) | Abnormity diagnosis method and device of product logs | |
| CN113806860B (en) | Fault feature extraction system, method, storage medium and equipment based on simulation | |
| CN112612664B (en) | Electronic equipment testing method and device, electronic equipment and storage medium | |
| CN112147967A (en) | Automatic generation method and device for scene test case of automobile controller | |
| CN107272443B (en) | Control method and device for controlled system | |
| CN114459760B (en) | Rolling bearing fault diagnosis method and system in strong noise environment | |
| CN112882899B (en) | Log abnormality detection method and device | |
| CN112256576B (en) | Man-machine dialogue corpus testing method, device, equipment and storage medium | |
| CN116226676B (en) | Machine tool fault prediction model generation method suitable for extreme environment and related equipment | |
| CN117596191A (en) | Power Internet of things abnormality detection method, device and storage medium | |
| CN111967003A (en) | Automatic wind control rule generation system and method based on black box model and decision tree | |
| CN110347579B (en) | Deep learning test case selection method based on neuron output behavior pattern | |
| CN116108749A (en) | Wind turbine flow field prediction method based on deep learning | |
| CN111474905B (en) | Parameter drift fault diagnosis method in manufacturing process of electromechanical product | |
| WO2021147298A1 (en) | Sentiment lexicon construction method and system, sentiment recognition method and system, and storage medium | |
| CN114252725A (en) | HHT and ResNet 18-based single-phase earth fault type comprehensive identification method | |
| CN115344951A (en) | Cutter wear amount prediction method based on time convolution network and auxiliary learning | |
| CN115146445A (en) | Method and related device for researching sensitivity of engine | |
| CN113722818B (en) | Digital test system for complex system of ship |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |