CN117111589A - Fault diagnosis method for numerical control machine tool control system based on Petri network - Google Patents
Fault diagnosis method for numerical control machine tool control system based on Petri network Download PDFInfo
- Publication number
- CN117111589A CN117111589A CN202311371018.8A CN202311371018A CN117111589A CN 117111589 A CN117111589 A CN 117111589A CN 202311371018 A CN202311371018 A CN 202311371018A CN 117111589 A CN117111589 A CN 117111589A
- Authority
- CN
- China
- Prior art keywords
- data
- temperature
- model
- state data
- numerical control
- 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.)
- Granted
Links
- 238000003745 diagnosis Methods 0.000 title claims abstract description 68
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000001816 cooling Methods 0.000 claims abstract description 56
- 125000004122 cyclic group Chemical group 0.000 claims abstract description 31
- 238000013528 artificial neural network Methods 0.000 claims abstract description 22
- 238000012423 maintenance Methods 0.000 claims abstract description 22
- 230000007704 transition Effects 0.000 claims description 35
- 230000001629 suppression Effects 0.000 claims description 22
- 230000008859 change Effects 0.000 claims description 19
- 238000012544 monitoring process Methods 0.000 claims description 17
- 238000012549 training Methods 0.000 claims description 15
- 230000008569 process Effects 0.000 claims description 12
- 238000003062 neural network model Methods 0.000 claims description 11
- 238000005461 lubrication Methods 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 9
- 230000005540 biological transmission Effects 0.000 claims description 7
- 238000010586 diagram Methods 0.000 claims description 7
- 238000009499 grossing Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 4
- 230000001960 triggered effect Effects 0.000 claims description 4
- 238000009825 accumulation Methods 0.000 claims description 3
- 239000012809 cooling fluid Substances 0.000 claims description 3
- 239000002173 cutting fluid Substances 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 230000020169 heat generation Effects 0.000 claims description 3
- 230000002401 inhibitory effect Effects 0.000 claims description 3
- 230000007774 longterm Effects 0.000 claims description 3
- 230000000737 periodic effect Effects 0.000 claims description 3
- 238000004088 simulation Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 9
- 230000003993 interaction Effects 0.000 abstract description 4
- 238000004422 calculation algorithm Methods 0.000 description 12
- 230000007246 mechanism Effects 0.000 description 10
- 238000005457 optimization Methods 0.000 description 9
- 230000008901 benefit Effects 0.000 description 6
- 238000007637 random forest analysis Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 3
- 238000003754 machining Methods 0.000 description 3
- 206010063385 Intellectualisation Diseases 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000013526 transfer learning Methods 0.000 description 2
- 238000002054 transplantation Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 239000000110 cooling liquid Substances 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 239000010687 lubricating oil Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0243—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model
- G05B23/0254—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses a failure diagnosis method of a numerical control machine tool control system based on a Petri network, which is used in the field of numerical control machine tools and comprises the following steps: constructing an overall Petri net model, and performing fault diagnosis on the current state data; according to the historical state data, a cyclic neural network is adopted to take time sequence information into consideration, and a fault prediction model is constructed; when a potential fault is predicted, according to the fault type and the component condition; collecting continuous multi-batch operation state data of the machine tool after maintenance; encrypting the running state data of the numerical control machine tool, and constructing a temperature state library and a cooling state library; correlating the decrypted temperature state library with a cooling state library in the overall Petri network model; modeling of input and output signals of the numerical control system is increased. The invention constructs the integral Petri network model, considers the interaction influence among subsystems, more comprehensively reflects various states of the machine tool, and improves the diagnosis effect.
Description
Technical Field
The invention relates to the field of numerically-controlled machine tools, in particular to a failure diagnosis method of a numerically-controlled machine tool control system based on a Petri network.
Background
The numerical control machine tool is mechanical equipment capable of automatically completing machining according to technological requirements, and mainly comprises a machine tool main body, a transmission system, a numerical control system and the like. The numerical control system is the brain of the numerical control machine tool and mainly completes the functions of machine tool motion control, process management and the like. In the long-time high-speed running process of the numerical control system, various faults, such as main shaft sensor faults, servo driving faults, mechanical component fatigue and the like, are easy to occur. These faults directly affect the processing quality and the production efficiency.
Therefore, an effective fault diagnosis and prediction method of the numerical control machine tool is very necessary to realize state monitoring, fault early warning and diagnosis of the numerical control system. The Petri net is a graphical mathematical modeling tool, and can intuitively represent the state of the system and the state transition relation driven by the event. And a failure diagnosis model of the numerical control system is established based on the Petri network, and the advantage of graphical state transition modeling can be utilized to realize effective monitoring of the failure of the numerical control system.
However, the following disadvantages exist in the prior art:
1. the traditional fault diagnosis relies on manual experience to judge, the efficiency is low, automation cannot be realized, and traditional maintenance is carried out according to a fixed period, so that excessive or insufficient conditions are easy to occur.
2. The traditional method cannot deal with individual differences of the machine tool, and lacks analysis of fault evolution process.
3. The system state data lacks protection, has potential safety hazard, lacks the prejudgement and the initiative control to temperature variation, and simultaneously diagnostic model and rule are difficult to adjust and expand.
For the problems in the related art, no effective solution has been proposed at present.
Disclosure of Invention
In order to overcome the problems, the invention aims to provide a failure diagnosis method of a numerical control machine tool control system based on a Petri network, and aims to solve the problems that the traditional failure diagnosis depends on manual experience judgment, the efficiency is low, the automation cannot be realized, the traditional maintenance is carried out according to a fixed period, and the excessive or insufficient problem easily occurs.
For this purpose, the invention adopts the following specific technical scheme:
a failure diagnosis method of a numerical control machine tool control system based on a Petri network comprises the following steps:
s1, constructing an integral Petri net model, and collecting historical state data and current state data of a numerical control machine tool;
s2, performing fault diagnosis on the current state data by using the integral Petri network model;
s3, according to the historical state data, a cyclic neural network is adopted to consider time sequence information so as to capture a fault evolution rule of the numerical control machine tool and construct a fault prediction model;
S4, when the potential faults are predicted, checking related components, and performing maintenance operation according to the fault types and the component conditions to ensure the reliable operation of the machine tool;
s5, after maintenance operation, operating the machine tool, collecting continuous multi-batch operation state data of the machine tool after maintenance, and optimizing the overall Petri network model and the fault prediction model by using the continuous multi-batch operation state data;
s6, encrypting the running state data of the numerical control machine tool, constructing a temperature state library and a cooling state library at the same time, and storing historical temperature data and historical cooling system state data;
s7, associating the decrypted temperature state library and the decrypted cooling state library in the overall Petri network model, and starting active thermal error suppression when the temperature is too high, and starting passive thermal error suppression when the temperature is normal;
s8, on the basis of the integral Petri network model, modeling of input and output signals of a numerical control system is increased, coverage of the integral Petri network model is expanded, and meanwhile accuracy of the current integral Petri network model and the fault prediction model is evaluated regularly.
Optionally, the building the overall Petri net model and collecting the historical state data and the current state data of the numerical control machine tool comprises the following steps:
S11, determining a subsystem according to a control system structure diagram of the numerical control machine tool, wherein the subsystem at least comprises a lubrication system, a transmission system, a servo system, a cooling system, an electrical control system and a hydraulic system;
s12, building a Petri network model for each subsystem, and building a Petri network for describing the states and state transitions of the subsystem components;
s13, merging the Petri network models of the subsystems to construct an overall Petri network model;
s14, connecting all sensors of the numerical control machine tool to obtain the running state data of the machine tool;
s15, collecting historical state data, establishing a state database, and storing a time sequence of each sensor data;
s16, developing a data interface program, acquiring current state data, and inputting the current state data into the overall Petri network model;
s17, processing the historical state data, and training a whole Petri network model by using the processed historical state data;
s18, testing and verifying the overall Petri network model, evaluating the prediction accuracy of the overall Petri network model by using an independent test data set, and optimizing the overall Petri network model according to a test result so as to improve the accuracy of the overall Petri network model.
Optionally, the fault diagnosis on the current state data by using the overall Petri net model includes the following steps:
S21, filtering, smoothing and standardizing the acquired state data;
s22, inputting the preprocessed state data into the overall Petri network model;
s23, running a whole Petri net model, and carrying out simulation according to current state data to obtain the mark number change of each state;
s24, analyzing the change condition of the number of marks in each position, and judging whether the machine tool has faults and the fault type according to the set fault diagnosis rules;
s25, outputting a fault diagnosis result, collecting operation data after the fault diagnosis result, and improving the fault diagnosis performance of the overall Petri network model by using the operation data.
Optionally, the analyzing the change condition of the number of marks in each position, and judging whether the machine tool fails and the failure type according to the set failure diagnosis rule includes the following steps:
s241, counting the variation of the number of marks in each position, analyzing the value and the positive and negative directions of the variation, and judging the flowing direction of the marks;
s242, referring to a preset fault diagnosis rule, determining the corresponding relation between the mark flow direction and the fault type;
s243, taking a sensor state mark flow direction alarm state as an example, and judging that the sensor fails according to the corresponding relation;
S244, integrating the flow condition and the direction of the marks at each position, and judging whether the machine tool has faults and the fault type according to fault diagnosis rules;
s245, setting a threshold value of the variation quantity for the key position, and outputting a fault diagnosis result and a cause according to the determined fault type when the variation quantity exceeds the threshold value.
Optionally, according to the historical state data, the time sequence information is considered by adopting a cyclic neural network to capture a fault evolution rule of the numerical control machine tool and construct a fault prediction model, and the method comprises the following steps of:
s31, collecting a historical state data time sequence of a numerical control machine tool in a long time range;
s32, denoising, smoothing and standardizing the collected historical state data;
s33, constructing a cyclic neural network model, and receiving a preprocessed state data time sequence;
s34, training a cyclic neural network model by using the historical state data to obtain a relationship between the reflected state data and various faults;
s35, according to a model reflecting the relation between state data and various faults, adding a full-connection layer as an output layer to the model, adjusting model parameters at the same time, constructing a cyclic neural network fault prediction model, and predicting faults occurring in a period of time in the future of the numerical control machine tool;
The method for predicting the faults of the numerical control machine tool in a future period of time comprises the following steps:
acquiring current state monitoring data, wherein the state monitoring data at least comprises temperature and pressure;
inputting the acquired real-time state data into a trained cyclic neural network fault prediction model;
and outputting a fault probability prediction result in a future period of time by using the trained cyclic neural network fault prediction model.
Optionally, when the potential fault is predicted, checking the relevant components, and performing maintenance operation according to the fault type and the component condition, wherein the step of ensuring the reliable operation of the machine tool comprises the following steps:
s41, determining a failed component according to the failure type given by the failure prediction model;
s42, checking the failed component in detail;
s43, determining the type of maintenance operation to be performed according to the checking result, disassembling, replacing and adjusting parameters of the components to be replaced, and modifying software control parameters;
s44, adjusting the gap of the fault component to ensure that the gap is within a normal range;
s45, checking and adjusting a lubrication system, a transmission system, a servo system, a cooling system, an electrical control system and a hydraulic system to ensure that each related system works normally.
Optionally, the encrypting the operation state data of the numerically-controlled machine tool, simultaneously constructing a temperature state library and a cooling state library, and storing the historical temperature data and the historical cooling system state data comprises the following steps:
s61, generating a symmetric key or an asymmetric key for encryption;
s62, encrypting the running state data of the numerical control machine by using the generated secret key;
s63, storing the encrypted running state data into a state database;
s64, constructing a temperature state library and storing historical temperature monitoring data;
s65, constructing a cooling state library and storing historical cooling system state data;
s66, encrypting the data in the temperature state library and the cooling state library by using the same secret key;
s67, storing the encrypted temperature and cooling state data into a corresponding database;
s68, updating the secret key regularly to ensure long-term security of the data;
the formula for encrypting the running state data is as follows:
in the method, in the process of the invention,representing the running state data of the numerical control machine tool;
representing a public key;
mod represents a remainder operation;
representing a large prime number;
encrypted data representing operation state data of the numerical control machine tool.
Optionally, the decrypted temperature state library and the decrypted cooling state library are associated in the overall Petri net model, and when the temperature is too high, active thermal error suppression is started, and when the temperature is normal, passive thermal error suppression is started, including the following steps:
S71, decrypting the data in the encrypted temperature state library and the encrypted cooling state library by using the same secret key to obtain original temperature data;
s72, in the overall Petri net model, increasing the position of a temperature state library, and increasing the transition of overhigh temperature, wherein the condition is that the temperature is higher than the upper limit temperature;
s73, increasing the temperature to perform normal transition, wherein the condition is that the temperature is in a normal range;
s74, operating the integral Petri network model, and inputting decrypted real-time temperature data;
s75, when the temperature is excessively high and transition is triggered, the overall Petri network model marks the excessively high temperature, and the active suppression of thermal errors is executed;
s76, when the temperature normal transition triggers, the overall Petri net model marks that the temperature is normal, and the thermal error is passively restrained;
s77, collecting data after temperature control, and optimizing an overall Petri net model;
the performing active thermal error suppression includes the steps of:
when the temperature of the machine tool exceeds the set upper limit temperature, taking active measures to reduce the temperature, wherein the active measures at least comprise increasing the working frequency of a cooling system and adjusting the working parameters of the machine tool so as to reduce the working heat generation of the machine tool and achieve the purpose of actively inhibiting thermal errors;
the performing thermal error passive suppression includes the steps of:
When the temperature of the machine tool is in a normal range, a normal working state is maintained, and a current temperature state is maintained by monitoring temperature change and taking passive measures which at least comprise periodic cooling and cutting fluid supply so as to prevent overhigh temperature caused by working heat accumulation.
Optionally, the formula for decrypting the data in the encrypted temperature state library and the cooling state library is:
the formula for decrypting the data in the encrypted temperature state library and the encrypted cooling state library is as follows:
in the method, in the process of the invention,representing a key;
encryption data representing operation state data of the numerical control machine tool;
mod represents a remainder operation;
representing a large prime number.
Optionally, on the basis of the overall Petri net model, modeling of input and output signals of a numerical control system is increased, coverage range of the overall Petri net model is expanded, and meanwhile accuracy of the current overall Petri net model and the fault prediction model is periodically evaluated, and the method comprises the following steps of:
s81, checking a schematic diagram of the numerical control system, and determining an input signal and an output signal;
s82, adding the position of the input signal and the position of the output signal in the overall Petri net model;
s83, adding transition in the overall Petri network model according to the logic relation between the position of the input signal and the position of the output signal;
S84, collecting signal state data in actual operation of the numerical control system, training a whole Petri network model by using the collected data, and determining a transition threshold value and a direction parameter;
s85, running the optimized integral Petri net model, performing fault diagnosis by using signal state data collected in real time, collecting signal data after diagnosis, continuously optimizing the integral Petri net model, and simultaneously periodically evaluating the accuracy of the current integral Petri net model and the fault prediction model.
Compared with the prior art, the application has the following beneficial effects:
1. the application builds the integral Petri network model to consider the interaction influence among subsystems, can reflect various states of the machine tool more comprehensively, and improves the diagnosis effect; the collected state data is comprehensive and rich, and a basis is provided for constructing and optimizing a model; the Petri net model is optimized by utilizing the data after the fault, so that the diagnosis performance can be continuously improved, and the understanding of a fault mechanism is promoted; the cyclic neural network model can analyze the time evolution rule of the fault by storing time sequence information; a mixed prediction model is constructed, the advantages of each algorithm are integrated, and the accuracy and the robustness of prediction are improved; the state control is combined with the prediction, so that the active management of the state change is realized; the model is expanded in a modularized mode, so that the structure is clearer, and optimization and transplantation are facilitated; the model is optimized in an incremental learning mode, so that new data can be quickly adapted, and forgetting is prevented; the data driving model is continuously optimized, and self-learning and self-adaption are realized; the multi-sensor acquires rich data and comprehensively reflects the running state of the machine tool.
2. The invention feeds back the data training model to enable the model to be continuously adapted to the actual condition of the machine tool, thereby improving the accuracy of fault judgment; by comparing the data differences, the model can learn to extract fault features, not just fixed rules; the Petri net model can also be perfected by using data, so that the fault description is closer to the actual situation; the incremental learning mode enables the model to be quickly adapted to new data, and meanwhile, the existing knowledge is prevented from being forgotten; the incremental network structure keeps the original model characteristics, also absorbs new knowledge, and realizes good balance; the data drive continuously self-learns and optimizes, so that the system gradually achieves intellectualization and self-adaption; the model is changed from depending on fixed rules to self-learning and judging according to actual data.
3. The invention encrypts the machine tool state data, thereby ensuring confidentiality and integrity of the data; an independent temperature and cooling state library is constructed, so that the management and the use are convenient; the original data is obtained through decryption and then is used for a Petri network model, so that the balance between safety and usability is realized; the temperature prediction control realizes active management and prevents abnormal states; numerical control signal modeling is added, the model range is expanded, and the diagnosis coverage is improved; the accuracy of the model is evaluated regularly, and model optimization can be performed according to the result; by using an encryption algorithm and a management means, the data security is improved, and the temperature prediction mechanism realizes the pre-judgment and active control of temperature change.
Drawings
The above features, features and advantages of the present application, as well as the manner of attaining them and method of attaining them, will become more apparent and the application will be better understood by reference to the following description of embodiments, taken in conjunction with the accompanying drawings. Here shown in schematic diagram:
fig. 1 is a flowchart of a fault diagnosis method of a control system of a numerical control machine based on a Petri net according to an embodiment of the present application.
Detailed Description
In order to make the present application better understood by those skilled in the art, the following description will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
According to the embodiment of the application, a failure diagnosis method of a numerical control machine tool control system based on a Petri network is provided.
The application will be further described with reference to the accompanying drawings and the specific embodiments, as shown in fig. 1, a fault diagnosis method for a control system of a numerical control machine based on Petri net according to an embodiment of the application, the fault diagnosis method includes the following steps:
S1, constructing an integral Petri net model, and collecting historical state data and current state data of the numerical control machine tool.
Preferably, the construction of the overall Petri net model and the collection of the historical state data and the current state data of the numerical control machine tool comprise the following steps:
s11, determining a subsystem according to a control system structure diagram of the numerical control machine tool, wherein the subsystem at least comprises a lubrication system, a transmission system, a servo system, a cooling system, an electrical control system and a hydraulic system;
s12, building a Petri network model for each subsystem, and building a Petri network for describing the states and state transitions of the subsystem components;
wherein, the state of each subsystem component refers to the working state of the component, and may include: normal operating conditions, fault conditions, alarm conditions, etc. For example, the status of the spindle system may include: a normal rotation state of the main shaft, an overload state of the main shaft, an overheat state of the main shaft and the like; state transition refers to the process of a component transitioning from one operational state to another. For example, the spindle is changed from a normal rotation state to an overload state, which is a state transition process. And establishing a Petri network description state and state transition, and establishing Petri network positions of different states for each component. Transitions describing state transition conditions are set. Such as spindle position status. Setting transition T1, and triggering when the rotating speed is too high, so as to realize transition. By constructing the locations of the different states and setting the transition conditions, a Petri network can be established to describe the various states of the component and the transition relationships between the states.
S13, merging the Petri network models of the subsystems to construct an overall Petri network model;
s14, connecting all sensors of the numerical control machine tool to obtain the running state data of the machine tool;
the machine tool running state data comprise spindle system data, and the spindle system data comprise parameters such as spindle rotating speed, spindle load, spindle temperature and the like, and can be obtained through a rotating speed sensor, a force sensor, a temperature sensor and the like. Feed system data, including feed speed, machining load, vibration data, etc. of each shaft, may be acquired by encoders, force sensors, vibration sensors, etc. Servo system data, wherein the servo system data comprise parameters such as current, position and the like of a servo motor, and are acquired through a current sensor, an encoder and the like. And the cooling system data mainly comprise parameters such as temperature, flow, pressure and the like of the cooling liquid, and are obtained through equipment such as a temperature sensor, a flowmeter, a pressure sensor and the like. Lubrication system data, including temperature, pressure, flow rate, etc., of the lubricating oil, are collected by the sensors. Environmental data, and the environmental data includes parameters such as plant environmental temperature, humidity, etc.
S15, collecting historical state data, establishing a state database, and storing a time sequence of each sensor data;
s16, developing a data interface program, acquiring current state data, and inputting the current state data into the overall Petri network model;
s17, processing the historical state data, and training a whole Petri network model by using the processed historical state data;
s18, testing and verifying the overall Petri network model, evaluating the prediction accuracy of the overall Petri network model by using an independent test data set, and optimizing the overall Petri network model according to a test result so as to improve the accuracy of the overall Petri network model.
It should be explained that, when the overall Petri net model is constructed, in order to enable the model to comprehensively reflect various states of the numerical control machine tool, interaction influences of different subsystems need to be considered. For example, the rotational speed state of the spindle system may affect the operating state of the lubrication system; the machining load of the feed system can affect the effectiveness of the cooling system, etc. Therefore, during modeling, state association data among all subsystems are required to be collected, and an association model of the spindle rotation speed, the state of a lubrication system, the feeding load, the cooling effect and the like is established. And adding transition reflecting the influence relation among subsystems into the overall Petri network model. The integrated Petri net model constructed in this way can reflect the working state and the fault state of the numerical control machine more comprehensively and accurately. Meanwhile, the model also provides possibility for predicting the occurrence of faults and the evolution mode thereof, and is beneficial to further improving the effect of fault prediction.
S2, performing fault diagnosis on the current state data by using the integral Petri network model.
Preferably, the fault diagnosis of the current state data by using the overall Petri net model comprises the following steps:
s21, filtering, smoothing and standardizing the acquired state data;
s22, inputting the preprocessed state data into the overall Petri network model;
s23, running a whole Petri net model, and carrying out simulation according to current state data to obtain the mark number change of each state;
s24, analyzing the change condition of the number of marks in each position, and judging whether the machine tool has faults and the fault type according to the set fault diagnosis rules;
s25, outputting a fault diagnosis result, collecting operation data after the fault diagnosis result, and improving the fault diagnosis performance of the overall Petri network model by using the operation data.
Preferably, the analyzing the change condition of the number of marks in each position, and judging whether the machine tool fails and the failure type according to the set failure diagnosis rule includes the following steps:
s241, counting the variation of the number of marks in each position, analyzing the value and the positive and negative directions of the variation, and judging the flowing direction of the marks;
S242, referring to a preset fault diagnosis rule, determining the corresponding relation between the mark flow direction and the fault type;
s243, taking a sensor state mark flow direction alarm state as an example, and judging that the sensor fails according to the corresponding relation;
s244, integrating the flow condition and the direction of the marks at each position, and judging whether the machine tool has faults and the fault type according to fault diagnosis rules;
s245, setting a threshold value of the variation quantity for the key position, and outputting a fault diagnosis result and a cause according to the determined fault type when the variation quantity exceeds the threshold value.
It should be explained that, when the Petri net model is used for fault diagnosis, operation data after fault can be collected to optimize the model so as to improve the subsequent diagnosis effect. Such optimization may be by incremental learning or by transfer learning. The model is trained by using the data increment after the fault on the basis of the original model, so that the model is adapted to the new condition. The adaptation of the model to new conditions can also be improved by using fault data of other machine tools in a transfer learning mode. Therefore, the Petri net model can continuously accumulate experience, adapt to different machine tools and fault modes thereof, and accordingly accuracy of fault diagnosis is improved. Meanwhile, the difference between the Petri net model and the Petri net model of different machine tools can be analyzed by comparing the Petri net model and the Petri net model of different machine tools, so that new knowledge on the failure mechanism can be obtained. Therefore, the Petri net model is optimized by collecting and utilizing the data after the faults, so that the diagnosis effect can be continuously improved, and the understanding of the fault mechanism can be possibly promoted.
And S3, according to the historical state data, adopting a cyclic neural network to consider time sequence information so as to capture a fault evolution rule of the numerical control machine tool and construct a fault prediction model.
Preferably, the step of taking time sequence information into consideration by using a cyclic neural network according to the historical state data to capture a fault evolution rule of the numerical control machine tool and construct a fault prediction model comprises the following steps:
s31, collecting a historical state data time sequence of a numerical control machine tool in a long time range;
s32, denoising, smoothing and standardizing the collected historical state data;
s33, constructing a cyclic neural network model, and receiving a preprocessed state data time sequence;
in addition, the cyclic neural network in the construction of the cyclic neural network model consists of an input layer, a hidden layer and an output layer. Wherein the hidden layer comprises an LSTM (long short term memory) module which can store time series history information. When the model is constructed, the number of nodes of an input layer, the LSTM structure of a hidden layer, the number of nodes of an output layer and the like need to be determined. Receiving the preprocessed time series data; the time-series data herein refers to historical state data of the numerical control machine tool, including time-varying signals of temperature, pressure, and the like. These time series data need to be first preprocessed, including denoising, normalization, etc. And then the preprocessed time sequence is taken as a sample and is sent into an input layer of the constructed cyclic neural network model. By training, the model can learn the law contained in the state time sequence and establish the mapping relation between input data and output, thereby achieving the purpose of predicting faults. In summary, a cyclic neural network model of a received time sequence is constructed, a model structure is required to be determined, data is preprocessed, and then a training model is input. This is the basis for fault prediction using the neural network.
S34, training a cyclic neural network model by using the historical state data to obtain a relationship between the reflected state data and various faults;
and S35, adding a full-connection layer as an output layer to the model according to the model reflecting the relation between the state data and various faults, adjusting model parameters, constructing a cyclic neural network fault prediction model, and predicting faults of the numerical control machine tool in a future period of time.
The method for predicting the faults of the numerical control machine tool in a future period of time comprises the following steps:
acquiring current state monitoring data, wherein the state monitoring data at least comprises temperature and pressure;
inputting the acquired real-time state data into a trained cyclic neural network fault prediction model;
and outputting a fault probability prediction result in a future period of time by using the trained cyclic neural network fault prediction model. It should be noted that, when the fault prediction model is constructed, other algorithms such as the cyclic neural network and the random forest may be considered to be combined to construct the hybrid model. For example, a cyclic neural network is used to analyze the time sequence of state data to obtain a prediction of probability of occurrence of a fault. And then taking the prediction probability as the input of a random forest model, and training the random forest to carry out multi-classification on different fault types. Therefore, the time sequence modeling capability of the cyclic neural network and the classification capability of the random forest are used in a mixed mode, so that the fault prediction model can integrate the advantages of the two algorithms, and the prediction accuracy is improved. Meanwhile, the angles of attention of each algorithm in the mixed model are different, one focuses on time evolution, one focuses on category distinction, and the two can mutually verify, so that the robustness of the model is enhanced. Therefore, in designing the prediction model, it is possible to consider that a plurality of algorithms are used in combination to improve the prediction performance.
And S4, checking related components when the potential faults are predicted, and performing maintenance operation according to the fault types and the component conditions to ensure the reliable operation of the machine tool.
Preferably, when a potential fault is predicted, the relevant components are checked, and maintenance operations are performed according to the fault type and the component conditions, and the following steps are included in ensuring reliable operation of the machine tool:
s41, determining a failed component according to the failure type given by the failure prediction model;
s42, checking the failed component in detail;
s43, determining the type of maintenance operation to be performed according to the checking result, disassembling, replacing and adjusting parameters of the components to be replaced, and modifying software control parameters;
s44, adjusting the gap of the fault component to ensure that the gap is within a normal range;
s45, checking and adjusting a lubrication system, a transmission system, a servo system, a cooling system, an electrical control system and a hydraulic system to ensure that each related system works normally.
It should be noted that, when the predicted failure is maintained, the component state data and the operation data before and after the maintenance may be stored. These data can be fed back to the fault prediction model to assist it in learning the normal state range of the component. The model may learn criteria for determining a rotational speed sensor failure, for example, by comparing differences in the spindle rotational speed sensor signals before and after maintenance. Meanwhile, the data can also provide basis for optimizing the Petri net model. The description of the Petri network model to the faults can be further perfected by comparing the state transition differences of the models before and after maintenance. Therefore, the state data before and after the maintenance of the component are recorded and fed back to the prediction model and the Petri network model, so that the follow-up fault prediction and diagnosis effect can be improved. The data driving model is continuously optimized, so that the system can gradually achieve self-learning and self-adaption of the fault mode.
And S5, after maintenance operation, operating the machine tool, collecting continuous multi-batch operation state data of the machine tool after maintenance, and optimizing the overall Petri network model and the fault prediction model by using the continuous multi-batch operation state data.
It should be noted that, the model optimization may be performed by incremental learning after the collection of the maintained operation data. I.e. each time a new batch of data is collected, the model is incrementally trained using the batch of data, gradually adjusting the model parameters. Incremental learning can adapt to new data faster and prevent the model from "forgetting" the original knowledge, compared to traditional approaches that re-train the entire model with all new data. In addition, the new data can also be used for training of new neural network branches and then fused with the original model. The incremental network structure constructing method can keep the original model characteristics and absorb new knowledge. In a word, the incremental model optimization is performed by using the maintained data, so that the incremental model optimization can be quickly adapted to a new state of a machine tool on the basis of keeping an original model, and the effects of the Petri network and the fault prediction model are improved.
S6, encrypting the running state data of the numerical control machine tool, simultaneously constructing a temperature state library and a cooling state library, and storing historical temperature monitoring data and historical cooling system state data.
Preferably, the encrypting the operation state data of the numerical control machine tool, simultaneously constructing a temperature state library and a cooling state library, and storing the historical temperature data and the historical cooling system state data comprises the following steps:
s61, generating a symmetric key or an asymmetric key for encryption;
s62, encrypting the running state data of the numerical control machine by using the generated secret key;
s63, storing the encrypted running state data into a state database;
s64, constructing a temperature state library and storing historical temperature monitoring data;
s65, constructing a cooling state library and storing historical cooling system state data;
s66, encrypting the data in the temperature state library and the cooling state library by using the same secret key;
s67, storing the encrypted temperature and cooling state data into a corresponding database;
s68, updating the secret key regularly to ensure long-term security of the data;
the formula for encrypting the running state data is as follows:
in the method, in the process of the invention,representing the running state data of the numerical control machine tool;
representing a public key;
mod represents a remainder operation;
representing a large prime number;
encrypted data representing operation state data of the numerical control machine tool.
It should be noted that, the security of the encrypted protection state data can be improved from both the aspects of algorithm and management. From the algorithm, a hybrid encryption algorithm can be used, namely symmetric encryption and asymmetric encryption are adopted at the same time, so that the encryption strength and the encryption efficiency can be achieved. In addition, it is also contemplated that using a blockchain-based encryption mechanism, the data may be made more difficult to tamper with the blockchain-based distributed ledger and consensus mechanism. From the management aspect, a key management system can be established to manage the key life cycle, and the security control of the processes of key generation, distribution, storage, update, revocation and the like is included. At the same time, access control and rights management are implemented on the database, severely limiting the subjects that can operate the database. In summary, the encryption protection of the state data is improved from the two aspects of algorithm and management, so that the sensitive data can be safer and more reliable.
S7, associating the decrypted temperature state library and the decrypted cooling state library in the overall Petri network model, and starting active thermal error suppression when the temperature is too high, and starting passive thermal error suppression when the temperature is normal.
Preferably, the decrypted temperature state library and the decrypted cooling state library are associated in the overall Petri net model, and when the temperature is too high, active thermal error suppression is started, and when the temperature is normal, passive thermal error suppression is started, which comprises the following steps:
s71, decrypting the data in the encrypted temperature state library and the encrypted cooling state library by using the same secret key to obtain original temperature data;
s72, in the overall Petri net model, increasing the position of a temperature state library, and increasing the transition of overhigh temperature, wherein the condition is that the temperature is higher than the upper limit temperature;
s73, increasing the temperature to perform normal transition, wherein the condition is that the temperature is in a normal range;
s74, operating the integral Petri network model, and inputting decrypted real-time temperature data;
s75, when the temperature is excessively high and transition is triggered, the overall Petri network model marks the excessively high temperature, and the active suppression of thermal errors is executed;
s76, when the temperature normal transition triggers, the overall Petri net model marks that the temperature is normal, and the thermal error is passively restrained;
S77, collecting data after temperature control, and optimizing the overall Petri net model
The performing active thermal error suppression includes the steps of:
when the temperature of the machine tool exceeds the set upper limit temperature, taking active measures to reduce the temperature, wherein the active measures at least comprise increasing the working frequency of a cooling system and adjusting the working parameters of the machine tool so as to reduce the working heat generation of the machine tool and achieve the purpose of actively inhibiting thermal errors;
the performing thermal error passive suppression includes the steps of:
when the temperature of the machine tool is in a normal range, a normal working state is maintained, and a current temperature state is maintained by monitoring temperature change and taking passive measures which at least comprise periodic cooling and cutting fluid supply so as to prevent overhigh temperature caused by working heat accumulation.
Preferably, the formula for decrypting the data in the encrypted temperature state library and the cooling state library is as follows:
in the method, in the process of the invention,representing a key;
encryption data representing operation state data of the numerical control machine tool;
mod represents a remainder operation;
representing a large prime number, which refers to a large random prime number,/for example>The value of (2) is typically generated by a key authority using a prime search algorithm.
It should be noted that when implementing temperature state control in the Petri net model, a prediction mechanism of increasing temperature may be considered. For example, a temperature prediction model is constructed by using time series data of temperature and related states, so as to predict the temperature change trend of a future period of time. Then in the Petri net model, state transition is triggered in advance according to whether the predicted temperature is higher than a limit value. The prediction starting thermal error active suppression mode can control earlier than the simple current temperature judgment, and can realize the pre-judgment and active control of temperature change. Compared with the mode of passive waiting for temperature rising and then suppressing, the mode of predicting starting can better prevent temperature overshoot and protect the safe operation of the machine tool. Therefore, the temperature prediction mechanism is added in the Petri network temperature control, so that the management capability of temperature change can be improved.
S8, on the basis of the integral Petri network model, modeling of input and output signals of a numerical control system is increased, coverage of the integral Petri network model is expanded, and meanwhile accuracy of the current integral Petri network model and the fault prediction model is evaluated regularly.
Preferably, the modeling of input and output signals of the numerical control system is increased on the basis of the overall Petri net model, the coverage range of the overall Petri net model is expanded, and meanwhile, the accuracy of the current overall Petri net model and the accuracy of the fault prediction model are periodically evaluated, and the method comprises the following steps of:
S81, checking a schematic diagram of the numerical control system, and determining an input signal and an output signal;
s82, adding the position of the input signal and the position of the output signal in the overall Petri net model;
s83, adding transition in the overall Petri network model according to the logic relation between the position of the input signal and the position of the output signal;
s84, collecting signal state data in actual operation of the numerical control system, training a whole Petri network model by using the collected data, and determining a transition threshold value and a direction parameter;
s85, running the optimized integral Petri net model, performing fault diagnosis by using signal state data collected in real time, collecting signal data after diagnosis, continuously optimizing the integral Petri net model, and simultaneously periodically evaluating the accuracy of the current integral Petri net model and the fault prediction model.
It should be explained that when the Petri net model is extended, splitting the model can be considered, and sub-models can be respectively built according to the functional modules. For example, an input signal sub-model, an output signal sub-model, etc. of the numerical control system may be separately established. Interfaces and rules are then defined to implement connections and constraints between the sub-models. The modularized method can make the structure of the whole model clearer and is convenient for customizing and optimizing the single module. And after modularization, if the model needs to be continuously expanded, the model needs to be carried out on the corresponding module, other parts are not affected, and the risk of system modification is reduced. In addition, the modularization is also beneficial to the standardized definition of the model, so that the model can be transplanted and reused. In a word, the use of the modularized method to expand the Petri net model can obtain the advantages of clear structure, convenient optimization, easy expansion and the like.
In summary, by means of the technical scheme, the method for constructing the integral Petri network model considers the interaction influence among subsystems, can reflect various states of a machine tool more comprehensively, and improves the diagnosis effect; the collected state data is comprehensive and rich, and a basis is provided for constructing and optimizing a model; the Petri net model is optimized by utilizing the data after the fault, so that the diagnosis performance can be continuously improved, and the understanding of a fault mechanism is promoted; the cyclic neural network model can analyze the time evolution rule of the fault by storing time sequence information; a mixed prediction model is constructed, the advantages of each algorithm are integrated, and the accuracy and the robustness of prediction are improved; the state control is combined with the prediction, so that the active management of the state change is realized; the model is expanded in a modularized mode, so that the structure is clearer, and optimization and transplantation are facilitated; the model is optimized in an incremental learning mode, so that new data can be quickly adapted, and forgetting is prevented; the data driving model is continuously optimized, and self-learning and self-adaption are realized; the multi-sensor acquires rich data and comprehensively reflects the running state of the machine tool; the invention feeds back the data training model to enable the model to be continuously adapted to the actual condition of the machine tool, thereby improving the accuracy of fault judgment; by comparing the data differences, the model can learn to extract fault features, not just fixed rules; the Petri net model can also be perfected by using data, so that the fault description is closer to the actual situation; the incremental learning mode enables the model to be quickly adapted to new data, and meanwhile, the existing knowledge is prevented from being forgotten; the incremental network structure keeps the original model characteristics, also absorbs new knowledge, and realizes good balance; the data drive continuously self-learns and optimizes, so that the system gradually achieves intellectualization and self-adaption; the model is changed from depending on fixed rules to self-learning and judgment according to actual data; the invention encrypts the machine tool state data, thereby ensuring confidentiality and integrity of the data; an independent temperature and cooling state library is constructed, so that the management and the use are convenient; the original data is obtained through decryption and then is used for a Petri network model, so that the balance between safety and usability is realized; the temperature prediction control realizes active management and prevents abnormal states; numerical control signal modeling is added, the model range is expanded, and the diagnosis coverage is improved; the accuracy of the model is evaluated regularly, and model optimization can be performed according to the result; and the encryption algorithm and the management means are used, so that the data security is improved. The temperature prediction mechanism realizes the pre-judgment and active control of temperature change.
Although the invention has been described with respect to the preferred embodiments, the embodiments are for illustrative purposes only and are not intended to limit the invention, as those skilled in the art will appreciate that various modifications can be made without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (10)
1. The fault diagnosis method of the numerical control machine tool control system based on the Petri network is characterized by comprising the following steps of:
s1, constructing an integral Petri net model, and collecting historical state data and current state data of a numerical control machine tool;
s2, performing fault diagnosis on the current state data by using the integral Petri network model;
s3, according to the historical state data, a cyclic neural network is adopted to consider time sequence information so as to capture a fault evolution rule of the numerical control machine tool and construct a fault prediction model;
s4, when the potential faults are predicted, checking related components, and performing maintenance operation according to the fault types and the component conditions to ensure the reliable operation of the machine tool;
s5, after maintenance operation, operating the machine tool, collecting continuous multi-batch operation state data of the machine tool after maintenance, and optimizing the overall Petri network model and the fault prediction model by using the continuous multi-batch operation state data;
S6, encrypting the running state data of the numerical control machine tool, simultaneously constructing a temperature state library and a cooling state library, and storing historical temperature monitoring data and historical cooling system state data;
s7, associating the decrypted temperature state library and the decrypted cooling state library in the overall Petri network model, and starting active thermal error suppression when the temperature is too high, and starting passive thermal error suppression when the temperature is normal;
s8, on the basis of the integral Petri network model, modeling of input and output signals of a numerical control system is increased, coverage of the integral Petri network model is expanded, and meanwhile accuracy of the current integral Petri network model and the fault prediction model is evaluated regularly.
2. The method for diagnosing faults of the control system of the numerical control machine based on the Petri net according to claim 1, wherein the steps of constructing an overall Petri net model and collecting historical state data and current state data of the numerical control machine comprise the following steps:
s11, determining a subsystem according to a control system structure diagram of the numerical control machine tool, wherein the subsystem at least comprises a lubrication system, a transmission system, a servo system, a cooling system, an electrical control system and a hydraulic system;
s12, building a Petri network model for each subsystem, and building a Petri network for describing the states and state transitions of the subsystem components;
S13, merging the Petri network models of the subsystems to construct an overall Petri network model;
s14, connecting all sensors of the numerical control machine tool to obtain the running state data of the machine tool;
s15, collecting historical state data, establishing a state database, and storing a time sequence of each sensor data;
s16, developing a data interface program, acquiring current state data, and inputting the current state data into the overall Petri network model;
s17, processing the historical state data, and training a whole Petri network model by using the processed historical state data;
s18, testing and verifying the overall Petri network model, evaluating the prediction accuracy of the overall Petri network model by using an independent test data set, and optimizing the overall Petri network model according to a test result so as to improve the accuracy of the overall Petri network model.
3. The fault diagnosis method for the numerical control machine control system based on the Petri net according to claim 2, wherein the fault diagnosis for the current state data by using the overall Petri net model comprises the following steps:
s21, filtering, smoothing and standardizing the acquired state data;
s22, inputting the preprocessed state data into the overall Petri network model;
S23, running a whole Petri net model, and carrying out simulation according to current state data to obtain the mark number change of each state;
s24, analyzing the change condition of the number of marks in each position, and judging whether the machine tool has faults and the fault type according to the set fault diagnosis rules;
s25, outputting a fault diagnosis result, collecting operation data after the fault diagnosis result, and improving the fault diagnosis performance of the overall Petri network model by using the operation data.
4. The fault diagnosis method for the numerical control machine control system based on the Petri net according to claim 3, wherein the analyzing the change condition of the number of marks in each position, judging whether the machine has faults and the fault type according to the set fault diagnosis rules comprises the following steps:
s241, counting the variation of the number of marks in each position, analyzing the value and the positive and negative directions of the variation, and judging the flowing direction of the marks;
s242, referring to a preset fault diagnosis rule, determining the corresponding relation between the mark flow direction and the fault type;
s243, taking a sensor state mark flow direction alarm state as an example, and judging that the sensor fails according to the corresponding relation;
s244, integrating the flow condition and the direction of the marks at each position, and judging whether the machine tool has faults and the fault type according to fault diagnosis rules;
S245, setting a threshold value of the variation quantity for the key position, and outputting a fault diagnosis result and a cause according to the determined fault type when the variation quantity exceeds the threshold value.
5. The method for diagnosing faults of a control system of a numerical control machine based on a Petri net according to claim 4, wherein the step of taking time sequence information into consideration by using a cyclic neural network according to historical state data to capture a fault evolution rule of the numerical control machine and construct a fault prediction model comprises the following steps:
s31, collecting a historical state data time sequence of a numerical control machine tool in a long time range;
s32, denoising, smoothing and standardizing the collected historical state data;
s33, constructing a cyclic neural network model, and receiving a preprocessed state data time sequence;
s34, training a cyclic neural network model by using the historical state data to obtain a relationship between the reflected state data and various faults;
s35, according to a model reflecting the relation between state data and various faults, adding a full-connection layer as an output layer to the model, adjusting model parameters at the same time, constructing a cyclic neural network fault prediction model, and predicting faults occurring in a period of time in the future of the numerical control machine tool;
The method for predicting the faults of the numerical control machine tool in a future period of time comprises the following steps:
acquiring current state monitoring data, wherein the state monitoring data at least comprises temperature and pressure;
inputting the acquired real-time state data into a trained cyclic neural network fault prediction model;
and outputting a fault probability prediction result in a future period of time by using the trained cyclic neural network fault prediction model.
6. The method for diagnosing faults in a control system of a numerical control machine based on a Petri net as claimed in claim 5, wherein when a potential fault is predicted, the relevant components are checked, maintenance operation is performed according to the fault type and the component condition, and the reliable operation of the machine is ensured, comprising the steps of:
s41, determining a failed component according to the failure type given by the failure prediction model;
s42, checking the failed component in detail;
s43, determining the type of maintenance operation to be performed according to the checking result, disassembling, replacing and adjusting parameters of the components to be replaced, and modifying software control parameters;
s44, adjusting the gap of the fault component to ensure that the gap is within a normal range;
S45, checking and adjusting a lubrication system, a transmission system, a servo system, a cooling system, an electrical control system and a hydraulic system to ensure that each related system works normally.
7. The method for diagnosing faults of a control system of a numerical control machine based on a Petri net according to claim 1, wherein the step of encrypting the operation state data of the numerical control machine and simultaneously constructing a temperature state library and a cooling state library and storing historical temperature monitoring data and historical cooling system state data comprises the following steps:
s61, generating a symmetric key or an asymmetric key for encryption;
s62, encrypting the running state data of the numerical control machine by using the generated secret key;
s63, storing the encrypted running state data into a state database;
s64, constructing a temperature state library and storing historical temperature monitoring data;
s65, constructing a cooling state library and storing historical cooling system state data;
s66, encrypting the data in the temperature state library and the cooling state library by using the same secret key;
s67, storing the encrypted temperature and cooling state data into a corresponding database;
s68, updating the secret key regularly to ensure long-term security of the data;
the formula for encrypting the running state data is as follows:
；
In the method, in the process of the invention,representing the running state data of the numerical control machine tool;
representing a public key;
mod represents a remainder operation;
representing a large prime number;
encrypted data representing operation state data of the numerical control machine tool.
8. The failure diagnosis method for the numerical control machine control system based on the Petri net according to claim 7, wherein the decrypted temperature state library and cooling state library are associated in the overall Petri net model, and when the temperature is too high, active thermal error suppression is started, and when the temperature is normal, passive thermal error suppression is started, which comprises the following steps:
s71, decrypting the data in the encrypted temperature state library and the encrypted cooling state library by using the same secret key to obtain original temperature data;
s72, in the overall Petri net model, increasing the position of a temperature state library, and increasing the transition of overhigh temperature, wherein the condition is that the temperature is higher than the upper limit temperature;
s73, increasing the temperature to perform normal transition, wherein the condition is that the temperature is in a normal range;
s74, operating the integral Petri network model, and inputting decrypted real-time temperature data;
s75, when the temperature is excessively high and transition is triggered, the overall Petri network model marks the excessively high temperature, and the active suppression of thermal errors is executed;
s76, when the temperature normal transition triggers, the overall Petri net model marks that the temperature is normal, and the thermal error is passively restrained;
S77, collecting data after temperature control, and optimizing an overall Petri net model;
the performing active thermal error suppression includes the steps of:
when the temperature of the machine tool exceeds the set upper limit temperature, taking active measures to reduce the temperature, wherein the active measures at least comprise increasing the working frequency of a cooling system and adjusting the working parameters of the machine tool so as to reduce the working heat generation of the machine tool and achieve the purpose of actively inhibiting thermal errors;
the performing thermal error passive suppression includes the steps of:
when the temperature of the machine tool is in a normal range, a normal working state is maintained, and a current temperature state is maintained by monitoring temperature change and taking passive measures which at least comprise periodic cooling and cutting fluid supply so as to prevent overhigh temperature caused by working heat accumulation.
9. The method for diagnosing faults of the numerical control machine control system based on the Petri net according to claim 8, wherein the formula for decrypting the data in the encrypted temperature state library and the encrypted cooling state library is as follows:
；
in the method, in the process of the invention,representing a key;
encryption data representing operation state data of the numerical control machine tool;
mod represents a remainder operation;
Representing a large prime number.
10. The method for diagnosing faults of a numerical control machine tool control system based on a Petri net according to claim 1, wherein the steps of increasing modeling of input and output signals of a numerical control system based on an overall Petri net model, expanding coverage of the overall Petri net model, and simultaneously periodically evaluating accuracy of a current overall Petri net model and a fault prediction model include the following steps:
s81, checking a schematic diagram of the numerical control system, and determining an input signal and an output signal;
s82, adding the position of the input signal and the position of the output signal in the overall Petri net model;
s83, adding transition in the overall Petri network model according to the logic relation between the position of the input signal and the position of the output signal;
s84, collecting signal state data in actual operation of the numerical control system, training a whole Petri network model by using the collected data, and determining a transition threshold value and a direction parameter;
s85, running the optimized integral Petri net model, performing fault diagnosis by using signal state data collected in real time, collecting signal data after diagnosis, continuously optimizing the integral Petri net model, and simultaneously periodically evaluating the accuracy of the current integral Petri net model and the fault prediction model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311371018.8A CN117111589B (en) | 2023-10-23 | 2023-10-23 | Fault diagnosis method for numerical control machine tool control system based on Petri network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311371018.8A CN117111589B (en) | 2023-10-23 | 2023-10-23 | Fault diagnosis method for numerical control machine tool control system based on Petri network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117111589A true CN117111589A (en) | 2023-11-24 |
| CN117111589B CN117111589B (en) | 2024-01-02 |
Family
ID=88804222
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311371018.8A Active CN117111589B (en) | 2023-10-23 | 2023-10-23 | Fault diagnosis method for numerical control machine tool control system based on Petri network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117111589B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117555287A (en) * | 2024-01-12 | 2024-02-13 | 中国机械总院集团云南分院有限公司 | CAE-based numerical control machine tool machining dynamic performance monitoring method and system |
Citations (41)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS54141969A (en) * | 1978-04-26 | 1979-11-05 | Toshiba Corp | Diagnosis of numerical control device |
| JPH04130953A (en) * | 1990-09-21 | 1992-05-01 | Fuji Xerox Co Ltd | Simulation device based upon petri net graph |
| EP0856791A1 (en) * | 1997-01-29 | 1998-08-05 | Verein Deutscher Werkzeugmaschinenfabriken e.V. (VDW) | System with and method for fault recovery |
| CN101118437A (en) * | 2007-09-03 | 2008-02-06 | 石毅 | New style numerically controlled machine remote condition monitoring and failure diagnosis system realizing method |
| CN101424936A (en) * | 2006-09-29 | 2009-05-06 | 沈阳高精数控技术有限公司 | Control method for closed-loop virtual system for numerically controlled machine |
| CN102566505A (en) * | 2012-02-27 | 2012-07-11 | 温州大学 | Intelligent fault diagnosis method for numerical control machine |
| CN102663500A (en) * | 2012-03-27 | 2012-09-12 | 许继集团有限公司 | Fuzzy Petri net transformer station fault diagnosis method based on time membership analysis |
| CN102680817A (en) * | 2012-04-28 | 2012-09-19 | 辽宁省电力有限公司朝阳供电公司 | Transformer fault diagnosis method based on fuzzy Petri net |
| US20150106314A1 (en) * | 2013-10-16 | 2015-04-16 | University Of Tennessee Research Foundation | Method and apparatus for constructing a dynamic adaptive neural network array (danna) |
| CN105081879A (en) * | 2015-08-27 | 2015-11-25 | 成都飞机工业（集团）有限责任公司 | Fault diagnosis and prediction method for numerical control machine tool spindle |
| CN105549533A (en) * | 2016-02-15 | 2016-05-04 | 上海理工大学 | Numerical control machine tool temperature field online monitoring and intelligent pre-warning system |
| CN105678467A (en) * | 2016-01-15 | 2016-06-15 | 国家电网公司 | Regulation and control integrated data analysis and aid decision making system and method under ultrahigh-voltage alternating current and direct current networking |
| CN105676781A (en) * | 2016-03-08 | 2016-06-15 | 华侨大学 | Numerically-controlled machine tool control system fault diagnosis method based on Petri net |
| CN105731209A (en) * | 2016-03-17 | 2016-07-06 | 天津大学 | Intelligent prediction, diagnosis and maintenance method for elevator faults on basis of Internet of Things |
| CN106353690A (en) * | 2016-09-20 | 2017-01-25 | 上海海事大学 | Method for diagnosing lithium battery faults by Petri net |
| CN106523393A (en) * | 2016-11-18 | 2017-03-22 | 山东科技大学 | Fault diagnosis method used for downhole draining system |
| KR20180048218A (en) * | 2016-10-31 | 2018-05-10 | 현대위아 주식회사 | Machine tool breakdown diagnosis system based on Machine Learning, and method thereof |
| CN108544303A (en) * | 2018-03-30 | 2018-09-18 | 上海交通大学 | A kind of main shaft of numerical control machine tool fault diagnosis method and system |
| CN109086945A (en) * | 2018-08-31 | 2018-12-25 | 沈阳航空航天大学 | A kind of machine tool chief axis spare part prediction technique based on operation analysis of covariance |
| CN110119518A (en) * | 2018-02-06 | 2019-08-13 | 洛阳中科晶上智能装备科技有限公司 | A method of engine failure reason is diagnosed using neural network model |
| CN110261159A (en) * | 2019-07-31 | 2019-09-20 | 上海应用技术大学 | Flexible manufacturing cutter subsystem fault diagnostic method |
| CN110333694A (en) * | 2019-07-31 | 2019-10-15 | 上海应用技术大学 | Numerical control device method for diagnosing faults based on Fuzzy Petri Net |
| CN110543142A (en) * | 2019-08-22 | 2019-12-06 | 电子科技大学 | Fault diagnosis and early warning system of numerical control machine tool |
| CN110647133A (en) * | 2019-09-09 | 2020-01-03 | 深圳市永达电子信息股份有限公司 | Rail transit equipment state detection maintenance method and system |
| CZ2018518A3 (en) * | 2018-09-30 | 2020-05-06 | 4Dot Mechatronic Systems S.R.O. | Diagnostic machine tools system |
| CN111313354A (en) * | 2020-02-11 | 2020-06-19 | 北京交通大学 | Method for determining dynamic equivalent boundary of alternating current-direct current power system |
| CN112257988A (en) * | 2020-09-29 | 2021-01-22 | 中广核工程有限公司 | Complex accident feature identification and risk early warning system and method for nuclear power plant |
| CN112947305A (en) * | 2021-02-05 | 2021-06-11 | 吉林大学 | Numerically-controlled machine tool reliability assessment method considering fault correlation |
| CN113655774A (en) * | 2021-07-21 | 2021-11-16 | 大唐互联科技(武汉)有限公司 | Machine tool remote diagnosis method and system based on industrial internet big data |
| KR20210141087A (en) * | 2020-05-15 | 2021-11-23 | 인제대학교 산학협력단 | Apparatus and method for diagnosing trouble of machine tool |
| CN114139589A (en) * | 2021-03-02 | 2022-03-04 | 湖州师范学院 | Fault diagnosis method, device, equipment and computer readable storage medium |
| WO2022074823A1 (en) * | 2020-10-09 | 2022-04-14 | 日本電気株式会社 | Control device, control method, and storage medium |
| KR20220046274A (en) * | 2020-10-07 | 2022-04-14 | 한국전기연구원 | System and method for fault diagnosis |
| CN115187014A (en) * | 2022-06-24 | 2022-10-14 | 山东科技大学 | Business process prediction method based on Petri network reachability graph interpretation |
| CN115358533A (en) * | 2022-07-27 | 2022-11-18 | 韩兆井 | Predictive maintenance management system based on big data |
| CN115493734A (en) * | 2022-09-29 | 2022-12-20 | 中国机械总院集团云南分院有限公司 | Electric control system for reliability test of numerical control machine tool |
| JP7240691B1 (en) * | 2021-09-08 | 2023-03-16 | 山東大学 | Data drive active power distribution network abnormal state detection method and system |
| CN115983452A (en) * | 2022-12-19 | 2023-04-18 | 河南科技大学 | Bearing fault prediction method based on LSTM |
| CN116205265A (en) * | 2023-01-12 | 2023-06-02 | 国电南瑞科技股份有限公司 | Power grid fault diagnosis method and device based on deep neural network |
| CN116578040A (en) * | 2023-07-13 | 2023-08-11 | 北京机床研究所有限公司 | Digital twin model construction method, system and terminal for machine tool fault diagnosis |
| CN116628869A (en) * | 2022-12-20 | 2023-08-22 | 南京航空航天大学 | Analysis method for fault propagation mechanism of numerical control machine tool based on transfer entropy theory |
-
2023
- 2023-10-23 CN CN202311371018.8A patent/CN117111589B/en active Active
Patent Citations (41)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS54141969A (en) * | 1978-04-26 | 1979-11-05 | Toshiba Corp | Diagnosis of numerical control device |
| JPH04130953A (en) * | 1990-09-21 | 1992-05-01 | Fuji Xerox Co Ltd | Simulation device based upon petri net graph |
| EP0856791A1 (en) * | 1997-01-29 | 1998-08-05 | Verein Deutscher Werkzeugmaschinenfabriken e.V. (VDW) | System with and method for fault recovery |
| CN101424936A (en) * | 2006-09-29 | 2009-05-06 | 沈阳高精数控技术有限公司 | Control method for closed-loop virtual system for numerically controlled machine |
| CN101118437A (en) * | 2007-09-03 | 2008-02-06 | 石毅 | New style numerically controlled machine remote condition monitoring and failure diagnosis system realizing method |
| CN102566505A (en) * | 2012-02-27 | 2012-07-11 | 温州大学 | Intelligent fault diagnosis method for numerical control machine |
| CN102663500A (en) * | 2012-03-27 | 2012-09-12 | 许继集团有限公司 | Fuzzy Petri net transformer station fault diagnosis method based on time membership analysis |
| CN102680817A (en) * | 2012-04-28 | 2012-09-19 | 辽宁省电力有限公司朝阳供电公司 | Transformer fault diagnosis method based on fuzzy Petri net |
| US20150106314A1 (en) * | 2013-10-16 | 2015-04-16 | University Of Tennessee Research Foundation | Method and apparatus for constructing a dynamic adaptive neural network array (danna) |
| CN105081879A (en) * | 2015-08-27 | 2015-11-25 | 成都飞机工业（集团）有限责任公司 | Fault diagnosis and prediction method for numerical control machine tool spindle |
| CN105678467A (en) * | 2016-01-15 | 2016-06-15 | 国家电网公司 | Regulation and control integrated data analysis and aid decision making system and method under ultrahigh-voltage alternating current and direct current networking |
| CN105549533A (en) * | 2016-02-15 | 2016-05-04 | 上海理工大学 | Numerical control machine tool temperature field online monitoring and intelligent pre-warning system |
| CN105676781A (en) * | 2016-03-08 | 2016-06-15 | 华侨大学 | Numerically-controlled machine tool control system fault diagnosis method based on Petri net |
| CN105731209A (en) * | 2016-03-17 | 2016-07-06 | 天津大学 | Intelligent prediction, diagnosis and maintenance method for elevator faults on basis of Internet of Things |
| CN106353690A (en) * | 2016-09-20 | 2017-01-25 | 上海海事大学 | Method for diagnosing lithium battery faults by Petri net |
| KR20180048218A (en) * | 2016-10-31 | 2018-05-10 | 현대위아 주식회사 | Machine tool breakdown diagnosis system based on Machine Learning, and method thereof |
| CN106523393A (en) * | 2016-11-18 | 2017-03-22 | 山东科技大学 | Fault diagnosis method used for downhole draining system |
| CN110119518A (en) * | 2018-02-06 | 2019-08-13 | 洛阳中科晶上智能装备科技有限公司 | A method of engine failure reason is diagnosed using neural network model |
| CN108544303A (en) * | 2018-03-30 | 2018-09-18 | 上海交通大学 | A kind of main shaft of numerical control machine tool fault diagnosis method and system |
| CN109086945A (en) * | 2018-08-31 | 2018-12-25 | 沈阳航空航天大学 | A kind of machine tool chief axis spare part prediction technique based on operation analysis of covariance |
| CZ2018518A3 (en) * | 2018-09-30 | 2020-05-06 | 4Dot Mechatronic Systems S.R.O. | Diagnostic machine tools system |
| CN110261159A (en) * | 2019-07-31 | 2019-09-20 | 上海应用技术大学 | Flexible manufacturing cutter subsystem fault diagnostic method |
| CN110333694A (en) * | 2019-07-31 | 2019-10-15 | 上海应用技术大学 | Numerical control device method for diagnosing faults based on Fuzzy Petri Net |
| CN110543142A (en) * | 2019-08-22 | 2019-12-06 | 电子科技大学 | Fault diagnosis and early warning system of numerical control machine tool |
| CN110647133A (en) * | 2019-09-09 | 2020-01-03 | 深圳市永达电子信息股份有限公司 | Rail transit equipment state detection maintenance method and system |
| CN111313354A (en) * | 2020-02-11 | 2020-06-19 | 北京交通大学 | Method for determining dynamic equivalent boundary of alternating current-direct current power system |
| KR20210141087A (en) * | 2020-05-15 | 2021-11-23 | 인제대학교 산학협력단 | Apparatus and method for diagnosing trouble of machine tool |
| CN112257988A (en) * | 2020-09-29 | 2021-01-22 | 中广核工程有限公司 | Complex accident feature identification and risk early warning system and method for nuclear power plant |
| KR20220046274A (en) * | 2020-10-07 | 2022-04-14 | 한국전기연구원 | System and method for fault diagnosis |
| WO2022074823A1 (en) * | 2020-10-09 | 2022-04-14 | 日本電気株式会社 | Control device, control method, and storage medium |
| CN112947305A (en) * | 2021-02-05 | 2021-06-11 | 吉林大学 | Numerically-controlled machine tool reliability assessment method considering fault correlation |
| CN114139589A (en) * | 2021-03-02 | 2022-03-04 | 湖州师范学院 | Fault diagnosis method, device, equipment and computer readable storage medium |
| CN113655774A (en) * | 2021-07-21 | 2021-11-16 | 大唐互联科技(武汉)有限公司 | Machine tool remote diagnosis method and system based on industrial internet big data |
| JP7240691B1 (en) * | 2021-09-08 | 2023-03-16 | 山東大学 | Data drive active power distribution network abnormal state detection method and system |
| CN115187014A (en) * | 2022-06-24 | 2022-10-14 | 山东科技大学 | Business process prediction method based on Petri network reachability graph interpretation |
| CN115358533A (en) * | 2022-07-27 | 2022-11-18 | 韩兆井 | Predictive maintenance management system based on big data |
| CN115493734A (en) * | 2022-09-29 | 2022-12-20 | 中国机械总院集团云南分院有限公司 | Electric control system for reliability test of numerical control machine tool |
| CN115983452A (en) * | 2022-12-19 | 2023-04-18 | 河南科技大学 | Bearing fault prediction method based on LSTM |
| CN116628869A (en) * | 2022-12-20 | 2023-08-22 | 南京航空航天大学 | Analysis method for fault propagation mechanism of numerical control machine tool based on transfer entropy theory |
| CN116205265A (en) * | 2023-01-12 | 2023-06-02 | 国电南瑞科技股份有限公司 | Power grid fault diagnosis method and device based on deep neural network |
| CN116578040A (en) * | 2023-07-13 | 2023-08-11 | 北京机床研究所有限公司 | Digital twin model construction method, system and terminal for machine tool fault diagnosis |
Non-Patent Citations (7)
| Title |
|---|
| 李太松;贺泽宇;王冰;颜永红;唐向红;: "基于循环时间卷积网络的序列流推荐算法", 计算机科学, vol. 47, no. 03, pages 103 - 109 * |
| 李红斌;王嘉伟;: "基于多源信息的数控机床故障诊断系统的设计与实现", 电子设计工程, no. 18, pages 78 - 82 * |
| 池正;罗继亮;邵辉;: "数控机床控制系统的Petri网故障诊断方法", 制造业自动化, no. 04, pages 59 - 65 * |
| 汪惠芬;梁光夏;刘庭煜;钟维宇;柳林燕;: "基于改进模糊故障Petri网的复杂系统故障诊断与状态评价", 计算机集成制造系统, no. 12, pages 3049 - 3061 * |
| 潘超;岳建平;刘冰;于景龙;: "基于适应型Petri网的电网故障诊断方法", 电网技术, no. 01, pages 46 - 50 * |
| 田仁铁;杨建国;: "一种基于Petri网模型的数控系统的分析与在线实时故障诊断方法", 东华大学学报(自然科学版), no. 04, pages 503 - 508 * |
| 谷东伟;张学文;王志琼;申桂香;张英芝;: "Petri网逆网在主轴故障诊断中的应用", 重庆大学学报, no. 06, pages 111 - 117 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117555287A (en) * | 2024-01-12 | 2024-02-13 | 中国机械总院集团云南分院有限公司 | CAE-based numerical control machine tool machining dynamic performance monitoring method and system |
| CN117555287B (en) * | 2024-01-12 | 2024-04-09 | 中国机械总院集团云南分院有限公司 | CAE-based numerical control machine tool machining dynamic performance monitoring method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117111589B (en) | 2024-01-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bhinge et al. | Toward a generalized energy prediction model for machine tools | |
| CN117111589B (en) | Fault diagnosis method for numerical control machine tool control system based on Petri network | |
| Fernandes et al. | Data analysis and feature selection for predictive maintenance: A case-study in the metallurgic industry | |
| US7647131B1 (en) | Dynamic determination of sampling rates | |
| US7827122B1 (en) | Data mining of unfiltered controller data | |
| CN101989087B (en) | On-line real-time failure monitoring and diagnosing system device for industrial processing of residual oil | |
| Ferreiro et al. | Industry 4.0: predictive intelligent maintenance for production equipment | |
| CN112418455B (en) | Equipment failure prediction and health management system | |
| CN108241343A (en) | A kind of intelligent plant management platform system | |
| KR20180108446A (en) | System and method for management of ict infra | |
| Liu et al. | A switching hidden semi-Markov model for degradation process and its application to time-varying tool wear monitoring | |
| CN106383916B (en) | Data processing method based on predictive maintenance of industrial equipment | |
| CN102929210A (en) | Control and optimization system for feature-based numerical control machining process and control and optimization method therefor | |
| CN112254972B (en) | Excavator oil temperature early warning method and device, server and excavator | |
| CN112328703B (en) | Numerical control equipment health state diagnosis device and method based on incremental learning | |
| Hoffmann et al. | Proposal for requirements on industrial AI solutions | |
| CN114065601A (en) | Management method of operation information of thermal power generating unit, storage medium and electronic device | |
| Colasante et al. | A fuzzy knowledge-based system for diagnosing unpredictable failures in CNC machine tools | |
| Liao et al. | Machine tool feed axis health monitoring using plug-and-prognose technology | |
| KR102252887B1 (en) | System and method for abnormal detecting a hierarchical data in OT network | |
| Kumar et al. | Development of a condition based maintenance architecture for optimal maintainability of mine excavators | |
| Zeiser et al. | Requirements towards optimizing analytics in industrial processes | |
| Gui et al. | Industrial-oriented machine learning big data framework for temporal-spatial error prediction and control with DTSMGCN model | |
| Armendia et al. | Twin-control: a new concept towards machine tool health management | |
| Alex et al. | Machine monitoring based on cyberphysical systems for industry 4.0 |
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 |