CN113325819B - Continuous annealing unit fault diagnosis method and system based on deep learning algorithm - Google Patents
Continuous annealing unit fault diagnosis method and system based on deep learning algorithm Download PDFInfo
- Publication number
- CN113325819B CN113325819B CN202110436732.5A CN202110436732A CN113325819B CN 113325819 B CN113325819 B CN 113325819B CN 202110436732 A CN202110436732 A CN 202110436732A CN 113325819 B CN113325819 B CN 113325819B
- Authority
- CN
- China
- Prior art keywords
- fault
- continuous annealing
- module
- annealing unit
- unit
- 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
Images
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
- G05B23/0259—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterized by the response to fault detection
- G05B23/0262—Confirmation of fault detection, e.g. extra checks to confirm that a failure has indeed occurred
-
- 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
- G05B2219/00—Program-control systems
- G05B2219/20—Pc systems
- G05B2219/24—Pc safety
- G05B2219/24065—Real time diagnostics
-
- 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
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Heat Treatment Of Strip Materials And Filament Materials (AREA)
- Control Of Heat Treatment Processes (AREA)
Abstract
The application discloses a continuous annealing unit fault diagnosis method and system based on a deep learning algorithm, the system comprises a data acquisition module, a fault mechanism mathematical model module, a deep learning module and a unit fault diagnosis control module, the system utilizes real-time online parameters, key offline data and typical case data stored in the past to cooperate with the fault mechanism mathematical model module, and utilizes a stack type self-coding deep learning algorithm to realize intelligent deep learning of continuous annealing unit fault diagnosis and intelligent training of key parameters, so that the contradiction coupling relation among all regulation and control parameters of the continuous annealing unit is reduced or weakened to the maximum extent, the online cooperation diagnosis control of fault information such as strip steel slipping, strip steel deflection, strip steel hot buckling and the like in a heating furnace of the continuous annealing unit is facilitated, the long-term stable and efficient operation of the continuous annealing unit is ensured, and the required fault information can be obtained, High grade strip products with good texture and mechanical properties.
Description
Technical Field
The invention relates to the field of intelligent measurement and control of strip steel continuous annealing faults, in particular to an intelligent fault detection diagnosis and intelligent processing process suitable for a continuous annealing (continuous annealing) unit process.
Background
With the rapid development of the modern industrial technology field and the application of important big data analysis means such as artificial intelligence and deep learning in the actual industrial production field, the requirements on the fault diagnosis technology of the cold rolling continuous annealing unit are more and more tending to automation, intellectualization and digitization. The main technological functions of the cold-rolled strip steel continuous annealing unit are as follows: the microstructure of the metal of the strip steel is changed by a high-temperature heating mode, so that the macroscopic properties of the metal, such as toughness, plasticity, tensile strength, strength and the like, are changed, and the high-quality ultrathin strip industrial raw material is obtained.
At present, a great deal of theoretical research and experimental tests are carried out on the defects of strip steel slipping, strip steel deviation, strip steel hot buckling and the like in the heating process mainly concentrated on the process faults of the continuous annealing unit by related researchers. Researches show that the defects of slipping, deviation, thermal buckling and the like of the steel strip in the heating furnace have strong coupling relationship, and the strong coupling relationship exists among various process parameters such as temperature, tension and friction coefficient in the annealing process. In the production process, the actual front and rear tension of the furnace roller can be changed by the speed of the motor which is too large or too small, and the change of the tension can influence the actual friction force of the strip material, so that the strip material has a slipping phenomenon and even deviates from the central line of the furnace roller. Meanwhile, the actual temperature of the heating furnace has extremely important influence on the transverse tensile stress distribution of the strip steel, overlarge transverse temperature difference can cause overlarge transverse load deviation, asymmetric load can cause traction force to the edge of the strip steel so as to cause deviation, and when the transverse load difference exceeds the yield strength of the strip steel, serious strip steel thermal buckling defect can be generated so as to produce unqualified strip steel with plate shape defect. Therefore, strictly speaking, the realization of real-time synchronous cooperative control by a continuous annealing unit fault system is the best way to solve a certain single fault problem or several complex fault defects, otherwise, the realization of single control of a certain fault is simply pursued and cooperative control among all faults is ignored, and long-term problems in the past can be caused, such as poor comprehensive control effect of unit faults, poor finished product production effect, low control efficiency and the like, and the long-term problems of failure in fundamental solution of important fault information can be generated, so that the production efficiency is seriously affected and the higher strip failure rate is caused.
At present, the automatic control processing system for the faults of the continuous annealing unit is applied to most factories, for example, the strip steel slipping problem in a heating furnace, the unit can be electrified by the power conversion of the motor to realize the function of judging whether the motor slips, and meanwhile, the slip fault can be prevented by adjusting the load power of the motor. For strip steel deviation faults occurring in a heating furnace of a continuous annealing unit, the conventional regulating and controlling means mostly adopt a CPC centering device to realize the automatic deviation rectifying function of the strip steel for the unit, the principle is to judge whether the strip steel deviates or not by using a special laser line arranged up and down of the strip steel, and when the strip steel deviates, the automatic deviation rectifying function of the strip steel can be realized by adjusting specific parameters. At present, no specific control means is provided for the hot buckling defect of the strip steel in the heating furnace of the continuous annealing unit, most of the control means is based on theoretical research and actual production experience, so that the buckling defect can be prevented and treated, and the prevention and treatment effect needs to be improved.
In addition, from the mesoscopic and microscopic perspectives, the continuous annealing unit is essentially characterized in that the metallographic structure of the metal is subjected to austenite phase transformation by using high temperature, and the metallographic structure of the metal is subjected to martensite phase transformation by means of high-temperature rapid quenching and the like to obtain a plurality of finished product strip steel with good quality such as high strength, high plasticity, high toughness and the like. In fact, to obtain a finished strip with good metallic properties, the temperature control of the various heating sections needs to be strictly controlled in order to achieve the optimum quenching temperature during the production process. However, in the actual production process, even if the heating furnace is provided with a ready-made temperature measurement and control point, the measured temperature is the actual temperature in the furnace chamber, and the actual temperature and the temperature distribution of the through plate under the high-speed operation cannot be accurately obtained. Therefore, much work needs to be perfected on the prediction of the actual temperature and the temperature distribution of the strip steel.
In conclusion, the cold rolling continuous annealing unit has a huge structure, complex operation, complex and changeable actual operation conditions on site, and extremely strict requirements on the control means and technology and fault treatment. Even though a large number of control means and mechanism models are provided, the problems that the existing mechanism models have too many assumed conditions, large process parameter change regulation and control range, unrealistic model calculation results and the like and are not enough to meet the requirements of actual production are solved because of strong coupling, mutual influence and mutual containment among faults. In addition, data in the actual production process on site is huge and disordered, mining is difficult, effective samples are too few or pertinence is not strong, and the online fault diagnosis application effect of the mechanism mathematical model is influenced to a certain extent.
Based on the above problem analysis, it is necessary to develop an intelligent deep learning fault diagnosis and processing system suitable for a continuous annealing unit based on a traditional mechanism model and an automatic control means. By closely cooperating with regular characteristic analysis and intelligent deep learning big data algorithm characteristic analysis of a traditional mechanism model, simultaneously configuring a large number of bottom layer data sensors and instruments and meters according to the requirements of a continuous annealing unit fault measurement and control system, synchronously acquiring current online unit operation parameters and real-time regulation and control parameters, optimizing different control parameters to ensure stable operation of the unit, optimizing a fault diagnosis and control period, combining offline detection data, deeply excavating comprehensive influence coefficients and coupling relations of strip steel slipping, deviation and thermal buckling defects by using a big data deep learning algorithm, analyzing the coupling influence of each parameter on strip steel tissue morphology and mechanical properties, finally realizing intelligent diagnosis and processing of continuous annealing unit fault information, and efficiently producing high-quality qualified strips.
Disclosure of Invention
The invention aims to provide a continuous annealing unit fault diagnosis method and system based on a deep learning algorithm, so as to solve the problems in the technical background.
In order to realize the purpose, the invention adopts the following technical scheme:
the first aspect of the application provides a continuous annealing unit fault diagnosis method based on a deep learning algorithm, which comprises the following steps:
collecting production process data of a continuous annealing unit in real time, and acquiring operation parameters of the continuous annealing unit;
coupling characteristic relations among strip steel slipping, strip steel deviation and strip steel hot buckling by using a fault mechanism mathematical model module, and performing synchronous coupling calculation or same-parameter regulation and control characteristic analysis on operation parameters of a continuous annealing unit to obtain influence rules of single-parameter regulation and control of the continuous annealing unit on different fault information and influence rules of multiple-parameter regulation and control on the same fault information, obtain corresponding relations between the operation parameters of the continuous annealing unit and the different fault information, and obtain coupling relation curves and influence coefficients between each regulation and control parameter of the continuous annealing unit and the different fault information;
the deep learning module adopts a stack type self-coding algorithm to construct a stack type self-coding neural network structure to form a plurality of classification fault diagnosis models, and performs big data deep training on the regulation and control parameters and the influence coefficients thereof corresponding to different fault information acquired by the fault mechanism mathematical model module to obtain a regulation and control parameter optimization interval corresponding to each self-learned regulation and control parameter, and continuously provides the fault mechanism mathematical model module with a regulation parameter, a regulation curve or an influence coefficient matrix which is in accordance with the current working condition;
and the unit fault diagnosis control module compares the real-time acquired regulation and control parameters with the corresponding regulation and control parameter optimization intervals in real time, judges whether the real-time acquired regulation and control parameters are in the corresponding regulation and control parameter optimization intervals, judges that the continuous annealing unit operates abnormally if the real-time acquired regulation and control parameters are not in the corresponding regulation and control parameter optimization intervals, and judges that the continuous annealing unit operates normally if the real-time acquired regulation and control parameters are not in the corresponding regulation and control parameter optimization intervals.
In the above, the operation parameter is data of the whole continuous annealing unit acquired by the data acquisition module when the continuous annealing unit operates, for example, the operation parameter is data acquired by each sensor of the whole continuous annealing unit when the continuous annealing unit operates; the regulation and control parameters are parameters which have influence on the stable operation of the continuous annealing unit and are part or all of the operation parameters, and when early warning occurs, the faults of the continuous annealing unit can be improved by adjusting the regulation and control parameters, so that the stable operation of the continuous annealing unit is realized.
In the above content, the deep learning module continuously provides the fault mechanism mathematical model module with the adjustment parameters, the adjustment curves or the influence coefficient matrix according with the current working conditions, so as to adjust the parameters of the mechanism analysis model in the fault mechanism mathematical model module and continuously correct the parameters, thereby making the mechanism analysis model more accurate.
Preferably, the method for diagnosing the fault of the continuous annealing unit further comprises the following steps: and if the running abnormity of the continuous annealing unit is judged, the unit fault diagnosis control module sends real-time fault information and solution measures to the front-end display module, and the front-end display module performs real-time voice alarm and/or digital display.
Preferably, the method for diagnosing the fault of the continuous annealing unit further comprises the following steps: and if the running of the continuous annealing unit is judged to be abnormal, the unit fault diagnosis control module sends control signals to various field control systems to realize the online feedback control of the fault information of the continuous annealing unit.
Preferably, the method for diagnosing the fault of the continuous annealing unit further comprises the following steps:
and a real-time database is arranged, receives and stores the production process data of the continuous annealing unit collected by the data collection module in real time, and transmits the production process data to the failure mechanism mathematical model module and the deep learning module.
Preferably, the method for diagnosing the fault of the continuous annealing unit further comprises the following steps:
setting a relational database, wherein the relational database is respectively in data transmission with the data acquisition module, the fault mechanism mathematical model module, the deep learning module and the unit fault diagnosis control module; the relational database receives and stores the production process data of the continuous annealing unit collected by the data collection module in real time, and transmits the production process data to the fault mechanism mathematical model module, the deep learning module and the unit fault diagnosis control module; meanwhile, the relational database stores the calculation result of the fault mechanism mathematical model module and the learning result of the deep learning module, and stores typical case data, wherein the typical case data are production process data of the continuous annealing unit in a set time period of automatic identification and screening of the unit fault diagnosis control module during operation, and the typical case data comprise a stable operation sample library and a typical fault sample library, the stable operation sample library is used for storing the production process data of the continuous annealing unit in a stable operation state, and the typical fault sample library is used for storing the production process data of the continuous annealing unit in various fault states.
More preferably, the method for diagnosing the fault of the continuous annealing unit further includes:
when the continuous annealing unit is abnormal in operation, the unit fault diagnosis control module classifies and compares real-time fault information with typical case data stored in the relational database to obtain specific parameters and solving measures causing the abnormal operation of the continuous annealing unit, voice broadcasting and/or digital display are carried out through the front-end display module, and closed-loop processing of the fault information is achieved through a parameter adjusting mode.
Preferably, the method for diagnosing the fault of the continuous annealing unit further comprises the following steps: the deep learning module adopts a stack type self-coding algorithm to construct a stack type self-coding neural network structure, a plurality of classification fault diagnosis models are formed by combining a softmax classifier, the regulation and control parameters corresponding to different fault information acquired by the fault mechanism mathematical model module are used as input, the network and the softmax classifier are sequentially trained by adopting a layer-by-layer greedy training method, and the influence coefficients of the regulation and control parameters under the same working condition or the variable working condition on band steel slipping, band steel deviation and band steel hot buckling are subjected to supervised fine tuning in the training process.
Preferably, the data acquisition module includes, but is not limited to, one or more of an underlying device sensor, a field PLC device, and a DCS device.
More preferably, the underlying equipment sensors include, but are not limited to, one or more of temperature sensors, tension sensors, rolling force sensors, speed sensors.
Preferably, the operation parameters of the continuous annealing unit acquired by the data acquisition module include, but are not limited to, one or more of motor rotation speed, tension before and after strip steel, strip steel speed, annealing temperature, annealing speed, furnace roller size, furnace roller tension, furnace roller load, unit set tension, unit center section acceleration, and strip shape before strip steel annealing.
Preferably, the fault information includes, but is not limited to, one or more of a slip fault, a deviation fault, and a thermal buckling fault.
The second aspect of the present application provides a continuous annealing unit fault diagnosis system based on a deep learning algorithm, including:
the data acquisition module is configured to acquire production process data of the continuous annealing unit in real time and acquire operation parameters of the continuous annealing unit;
the fault mechanism mathematical model module is connected with the data acquisition module, comprises a plurality of mechanism analysis models including at least a slip fault model, a deviation fault model and a thermal buckling fault model, and is configured to obtain the influence rule of the single parameter regulation and control of the continuous annealing unit on different fault information and the influence rule of the multiple parameter regulation and control on the same fault information by performing synchronous coupling calculation or same parameter regulation and control characteristic analysis on the operation parameters of the continuous annealing unit, obtain the corresponding relation between the operation parameters of the continuous annealing unit and different fault information and obtain the coupling relation curve and the influence coefficient between each regulation and control parameter of the continuous annealing unit and different fault information;
the deep learning module is connected with the fault mechanism mathematical model module, is configured to adopt a stacked self-coding algorithm, constructs a stacked self-coding neural network structure, forms a plurality of classified fault diagnosis models, performs big data deep training on the regulation parameters and the influence coefficients thereof corresponding to different fault information acquired by the fault mechanism mathematical model module, obtains the corresponding regulation parameter optimization sections after the self-learning of each regulation parameter, and continuously provides the fault mechanism mathematical model module with the regulation parameters, the regulation curves or the influence coefficient matrixes which accord with the current working conditions;
and the unit fault diagnosis control module is connected with the data acquisition module, the deep learning module and the fault mechanism mathematical model module and is configured to diagnose whether the continuous annealing unit normally operates by comparing the regulation parameters with the corresponding regulation parameter optimization areas in real time.
In the above, the operation parameter is data of the whole continuous annealing unit acquired by the data acquisition module when the continuous annealing unit operates, for example, the operation parameter is data acquired by each sensor of the whole continuous annealing unit when the continuous annealing unit operates; the regulation and control parameters are parameters influencing the stable operation of the continuous annealing unit and are part or all of the operation parameters, and when early warning occurs, the faults of the continuous annealing unit can be improved by regulating the regulation and control parameters, so that the stable operation of the continuous annealing unit is realized.
In the above content, the deep learning module continuously provides the fault mechanism mathematical model module with the adjustment parameters, the adjustment curves or the influence coefficient matrix according with the current working conditions, so as to adjust the parameters of the mechanism analysis model in the fault mechanism mathematical model module and continuously correct the parameters, thereby making the mechanism analysis model more accurate.
Preferably, the unit fault diagnosis control module is further interactive with the continuous annealing unit to realize a closed loop, and is configured to send control signals to various field control systems to realize online feedback control of fault information of the continuous annealing unit.
Preferably, the data acquisition module includes, but is not limited to, one or more of an underlying device sensor, a field PLC device, and a DCS device.
More preferably, the underlying equipment sensors include, but are not limited to, one or more of temperature sensors, tension sensors, rolling force sensors, speed sensors.
Preferably, the operation parameters of the continuous annealing unit acquired by the data acquisition module include, but are not limited to, one or more of motor rotation speed, tension before and after strip steel, strip steel speed, annealing temperature, annealing speed, furnace roller size, furnace roller tension, furnace roller load, unit set tension, unit center section acceleration, and strip shape before strip steel annealing.
Preferably, the fault information includes, but is not limited to, one or more of a slip fault, a tracking fault, and a thermal buckling fault.
Preferably, the continuous annealing unit fault diagnosis system further comprises a front-end display module, wherein the front-end display module is connected with the unit fault diagnosis control module and is configured to realize real-time display of continuous annealing unit real-time operation information and/or fault information, and/or intelligent voice alarm and/or manual fault adjustment.
In a preferred embodiment, the front-end display module includes, but is not limited to, one or more of a real-time voice alarm module, a production image monitoring module, an intelligent diagnosis and processing module, and a report query and statistics module.
Preferably, the continuous annealing unit fault diagnosis system further comprises a large field database, wherein the large field database comprises a real-time database and a relational database; wherein the content of the first and second substances,
the real-time database is configured to receive and store the production process data of the continuous annealing unit acquired by the data acquisition module in real time, and output the production process data to the input end of the failure mechanism mathematical model module and the input end of the deep learning module;
the relational database is respectively connected with the data acquisition module, the fault mechanism mathematical model module, the deep learning module and the unit fault diagnosis control module, and is configured to receive and store production process data of the continuous annealing unit acquired by the data acquisition module in real time and transmit the production process data to the fault texture data model module, the deep learning module and the unit fault diagnosis control module; meanwhile, the relational database stores the calculation result of the fault texture data model module and the learning result of the deep learning module, and stores the production process data of the continuous annealing unit in a stable operation state and in set time periods of various fault states.
More preferably, the failure diagnosis system for the continuous annealing unit further comprises a network transmission module, the data acquisition module stores the acquired production process data into the real-time database and the relational database through the network transmission module, and the network transmission module supports transmission protocols of a local area network, a LAN, a WiFi, a Zigbee, a Bluetooth, 5G, RFID and a GPS.
Preferably, the stacked self-coding neural network structure of the deep learning module is composed of multiple layers of sparse self-encoders, and each layer includes multiple sparse self-encoders.
Compared with the prior art, the technical scheme of the invention has the following beneficial effects:
the application provides a continuous annealing unit fault diagnosis method and system based on a deep learning algorithm, the system comprises a data acquisition module, a fault mechanism mathematical model module, a deep learning module and a unit fault diagnosis control module, utilizes real-time online parameters, key offline data and typical case data stored in the past to cooperate with the fault mechanism mathematical model module, utilizes a stacked self-coding deep learning algorithm to realize intelligent deep learning of continuous annealing unit fault diagnosis and intelligent training of key parameters, reduces or weakens contradiction coupling relations among all regulation and control parameters of the continuous annealing unit to the maximum extent, is favorable for realizing online cooperative diagnosis control of fault information such as strip steel slipping, strip steel deflection, strip steel thermal buckling and the like in a heating furnace of the continuous annealing unit, ensures long-term stable and efficient operation of the continuous annealing unit, so as to obtain a high-grade strip product with good structure form and mechanical property meeting the requirement. The technical scheme of the application breaks through various constraints that the problem of faults is difficult to solve when a traditional continuous annealing unit operates, provides a good processing mode for cooperative coupling diagnosis of faults of the continuous annealing unit, and can realize digital, intelligent and intelligent online fault diagnosis and detection of stable operation of the continuous annealing unit.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the application and, together with the description, serve to explain the application and are not intended to limit the application. In the drawings:
fig. 1 is a block diagram of an intelligent fault diagnosis system for a continuous annealing unit based on a deep learning algorithm according to a preferred embodiment of the present invention;
FIG. 2 is a schematic diagram of an intelligent fault diagnosis method for a continuous annealing unit based on a deep learning algorithm according to a preferred embodiment of the present invention;
fig. 3 is a diagram of a stacked self-coding fault diagnosis model according to a preferred embodiment of the present invention.
Illustration of the drawings:
1. a data acquisition module; 2. a real-time database; 3. a relational database; 4. a failure mechanism mathematical model module; 5. a deep learning module; 6. a unit fault diagnosis control module; 7. and a front end display module.
Detailed Description
In order to make the objects, technical solutions and effects of the present invention clearer and clearer, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
It should be noted that the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover non-exclusive inclusions, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
Fig. 1 shows a block diagram of a continuous annealing unit intelligent fault diagnosis system based on a deep learning algorithm. As shown in fig. 1, the intelligent fault diagnosis system for the continuous annealing unit comprises a data acquisition module 1, a real-time database 2, a relational database 3, a fault mechanism mathematical model module 4, a deep learning module 5, a unit fault diagnosis control module 6 and a front-end display module 7.
The data acquisition module 1 mainly comprises a bottom layer equipment sensor (a temperature sensor, a tension sensor, a rolling force sensor, a speed sensor and the like), field PLC equipment and DCS equipment, and is used for acquiring production process data of the continuous annealing unit in real time and acquiring operation parameters of the continuous annealing unit. The operation parameters of the continuous annealing unit acquired by the data acquisition module 1 include, but are not limited to, one or more of motor rotation speed, tension before and after strip steel, strip steel speed, annealing temperature, annealing speed, furnace roller size, furnace roller tension, furnace roller load, unit set tension, unit center section acceleration and strip shape before strip steel annealing.
The input end of the real-time database 2 is connected with the data acquisition module 1, the output end of the real-time database is connected with the fault mechanism mathematical model module 4 and the deep learning module 5, the production process data acquired by the data acquisition module 1 are collected and stored in real time mainly through equipment such as a field gateway and an Ethernet, and the production process data are transmitted to the fault mechanism mathematical model module 4 and the deep learning module 5.
The relational database 3 is respectively connected to the data acquisition module 1, the fault mechanism mathematical model module 4, the deep learning module 5, and the unit fault diagnosis control module 6, and is mainly used for storing a regulation parameter data result obtained by processing by the fault mechanism mathematical model module 4, storing typical case data, providing a data source for a deep learning big data algorithm of the deep learning module 5, and storing a data optimization processing result. Specifically, the relational database 3 receives and stores the production process data of the continuous annealing unit collected by the data collection module 1 in real time, and transmits the production process data to the fault mechanism mathematical model module 4, the deep learning module 5 and the unit fault diagnosis control module 6; meanwhile, the relational database 3 stores the calculation result of the fault mechanism mathematical model module 4 and the learning result of the deep learning module 5, and stores typical case data, wherein the typical case data is production process data of the continuous annealing unit in a set time period which is automatically identified and screened by the unit fault diagnosis control module 6 when the continuous annealing unit operates, and comprises a stable operation sample library and a typical fault sample library, the stable operation sample library is used for storing the production process data of the continuous annealing unit in a stable operation state, and the typical fault sample library is used for storing the production process data of the continuous annealing unit in various fault states.
The fault mechanism mathematical model module 4 comprises a plurality of mechanism analysis models including a slip fault model, a deviation fault model and a thermal buckling fault model. The method is characterized in that the method utilizes relevant theoretical research of defects of strip steel slipping, strip steel deviation, strip steel buckling and the like of the existing continuous annealing unit, and combines a large amount of simulation and experiments to carry out deep excavation and summary on various fault information to find out key influence parameters of different fault information (such as slipping fault, deviation fault and thermal buckling fault). Because different regulation and control parameters have strong coupling relation to different fault information, the fault mechanism mathematical model module 4 needs to perform synchronous coupling calculation or same parameter regulation and control characteristic analysis on operation parameters such as furnace roller load, strip steel front and back tension, strip steel speed, annealing temperature and the like to obtain an influence relation curve of single parameter regulation and control of the continuous annealing unit on different fault information and/or an influence relation curve of multiple parameter regulation and control on the same fault information, obtain basic corresponding relation characteristics between the operation parameters of the continuous annealing unit and different fault information, obtain a coupling relation curve and an influence coefficient between each regulation and control parameter of the continuous annealing unit and different fault information, and realize the online synchronous intelligent diagnosis function of multiple faults of the continuous annealing unit. And further, when the continuous annealing unit breaks down, synchronous coupling control of multiple faults can be realized by adjusting a certain regulation parameter or multiple regulation parameters.
The deep learning module 5 is connected with the real-time database 2, the relational database 3, the fault mechanism mathematical model module 4 and the unit fault diagnosis control module 6. The deep learning module 5 adopts a stack type self-coding algorithm of a big data deep learning algorithm to construct a deep neural network structure of a stack type encoder, form a plurality of classification fault diagnosis models, sequentially and deeply train the strip steel parameters, the unit parameters and other regulation and control parameters to obtain the optimized data (such as the optimal value of the regulation and control parameters or the regulation and control parameter optimization interval) of the corresponding regulation and control parameters after the self-learning of each regulation and control parameter, and continuously provide the fault mechanism mathematical model module 4 with the regulation parameters, the regulation curves or the influence coefficient matrix which are in line with the current working condition.
The unit fault diagnosis control module 6 is respectively connected with the data acquisition module 1, the fault mechanism mathematical model module 4 and the deep learning module 5, and is mainly used for diagnosing whether the continuous annealing unit normally operates by comparing the regulation parameters with the corresponding regulation parameter optimization intervals in real time, and realizing the online feedback control of the fault information of the continuous annealing unit by sending control signals to various field control systems when the continuous annealing unit abnormally operates.
The front-end display module 7 is connected with the unit fault diagnosis control module 6 and is used for realizing the functions of real-time running information and/or real-time fault information display of the continuous annealing unit, and/or intelligent voice alarm, and/or manual fault regulation.
Fig. 2 shows a schematic diagram of an intelligent fault diagnosis method for a continuous annealing unit based on a deep learning algorithm. Referring to fig. 2, the intelligent fault diagnosis method for the continuous annealing unit mainly includes the following steps:
step 1: the data acquisition module 1 is communicated with the real-time database 2 and the relational database 3, the production process data of the continuous annealing unit are acquired in real time, the operation parameters of the continuous annealing unit are acquired, and meanwhile, the relational database 3 carries out real-time online acquisition and offline storage on typical case data (a stable operation sample library and a typical fault sample library).
And 2, step: by utilizing a fault mechanism mathematical model module 4, coupling characteristic relations among strip steel slipping, strip steel deviation and strip steel hot buckling, and performing synchronous coupling calculation or same-parameter regulation characteristic analysis on operation parameters such as furnace roller load, motor rotating speed, strip steel front and back tension, strip steel speed, annealing temperature and the like to obtain an influence rule of single-parameter regulation and control of a continuous annealing unit on different fault information and an influence rule of multiple-parameter regulation and control on the same fault information, obtain corresponding relations between the operation parameters of the continuous annealing unit and different fault information, and obtain a coupling relation curve and an influence coefficient between each regulation and control parameter of the continuous annealing unit and different fault information.
And step 3: based on the analysis of the fault mechanism mathematical model module 4 in the step 2, the real-time online data and the offline measurement data are acquired, the past typical case data stored in the relational database 3 are cooperated, and the deep learning module 5 trains the topological structure among a plurality of coupling parameters by using a stack type self-coding algorithm of a big data deep learning algorithm, and finely adjusts the influence coefficients of the control parameters under the same working condition or under the variable working condition on the band steel slip, the band steel deviation and the band steel hot buckling. And (3) carrying out a series of big data deep training such as coding, layer-by-layer training, decoding and the like on the data set to obtain optimal values or optimal intervals corresponding to all the regulation and control parameters. Meanwhile, the proportion weight or the adjustment factor of the important parameters under different working conditions is optimized so as to improve the online diagnosis precision and the fault regulation and control precision of the unit fault diagnosis control module 6 and provide the adjustment parameters, the adjustment curve or the influence coefficient matrix which are in accordance with the current working conditions for the fault mechanism mathematical model module 4 as far as possible.
The method comprises the steps of firstly constructing a deep neural network structure of a stacked encoder formed by stacking a plurality of layers of sparse automatic encoders, then combining a softmax classifier to form a plurality of classification fault diagnosis models, combining parameters such as steel coil numbers, strip steel parameters, motor loads, tension and temperature of strips and the like as input, sequentially training the network and the classifier by adopting a layer-by-layer greedy training method, and carrying out fine adjustment on the whole neural network structure in a supervised manner in the training process to improve the training precision.
And 4, step 4: the unit fault diagnosis control module 6 compares the real-time acquired regulation and control parameters with the corresponding regulation and control parameter optimization intervals in real time, judges whether the real-time acquired regulation and control parameters are in the corresponding regulation and control parameter optimization intervals, judges that the continuous annealing unit operates abnormally if the real-time acquired regulation and control parameters are not in the corresponding regulation and control parameter optimization intervals, and judges that the continuous annealing unit operates normally if the real-time acquired regulation and control parameters are not in the corresponding regulation and control parameter optimization intervals.
If the continuous annealing unit normally operates, the unit fault diagnosis control module 6 automatically identifies, screens and stores the production process data of the time period to the relational database 3 to form a stable operation sample library.
If the continuous annealing unit is abnormal in operation, the unit fault diagnosis control module 6 can execute the following operations:
1) the unit fault diagnosis control module 6 automatically identifies, screens and stores the production process data of the time period to the relational database 3 to form a typical fault sample library;
2) the unit fault diagnosis control module 6 classifies and compares the real-time fault information with typical case data stored in the relational database 3 to obtain specific parameters and solution measures causing the abnormal operation of the continuous annealing unit, so as to realize the intelligent diagnosis of the fault information of the continuous annealing unit;
3) the unit fault diagnosis control module 6 sends control signals to various field control systems to realize the online feedback control of the continuous annealing unit fault information.
And 5: when the continuous annealing unit is abnormal in operation, the unit fault diagnosis control module 6 sends real-time fault information and solution to the front end display module 7, the front end display module performs voice broadcasting and/or digital display, and closed-loop processing of the fault information is achieved through a parameter adjusting mode.
Fig. 3 shows a stacked self-coding fault diagnosis model diagram. The simple self-coding is a three-layer neural network model which comprises a data input layer, a hidden layer and an output reconstruction layer, and is an unsupervised learning model. In the method for diagnosing the faults of the continuous annealing unit, original unit data, such as parameters of temperature, tension, annealing speed and the like, are transmitted to a data input layer, the output is used as the input of a next layer of self-encoder by utilizing the training of a self-encoder, so that a multi-layer network structure is formed, and important parameter abstract characteristics of various units are extracted layer by layer. The encoding process of the stacked encoder is to perform the encoding steps of each layer of the automatic encoder in the order from the first layer of the automatic encoder to the last layer of the automatic encoder, and similarly, the decoding process of the stacked encoder is to perform each layer of the automatic encoder in the order from the back to the front. A better method for acquiring neural network parameters of a stacked encoder is to train by a layer-by-layer greedy training method. Firstly, training a first layer of a network by using original important regulation and control parameters and unit operation data as input to obtain parameters of the network; then, the first layer of the network converts the original input into a vector consisting of the activation values of the hidden units, and then the vector is used as the input of the second layer to continue training to obtain the parameters of the second layer; and finally, sequentially training the subsequent layers by adopting the same strategy. For the above training mode, when training the important parameters of each layer, the parameters of other layers are fixed and kept unchanged. Therefore, there is a limitation that if a better result is desired, after the pre-training process is completed, the parameters of all layers can be adjusted simultaneously by the back-propagation algorithm to improve the result, which process is generally referred to as fine-tuning. The fine tuning can greatly improve the performance of the deep learning network, is a common method for optimizing the parameters of the deep learning network, and can improve the deep learning training precision of the continuous annealing unit fault diagnosis method, improve the response rate and optimize the training parameters.
In conclusion, the invention provides a continuous annealing unit fault diagnosis system and method based on a deep learning algorithm. The method deeply combines typical mechanism analysis models such as a slip fault model, a deviation fault model, a thermal buckling fault model and a microstructure and material performance related model, utilizes a stack type self-coding algorithm in a big data deep learning algorithm, realizes accurate, sensitive and effective continuous annealing unit fault diagnosis and treatment, solves the problems of strong coupling effect and difficult fundamental treatment of continuous annealing unit faults, ensures the continuous annealing unit to run stably, smoothly and efficiently in real time, improves the mechanical property of processed finished strips, and can produce high-quality cold-rolled strips.
The embodiments of the present invention have been described in detail, but the embodiments are only examples, and the present invention is not limited to the embodiments described above. Any equivalent modifications and substitutions to those skilled in the art are also within the scope of the present invention. Accordingly, equivalent alterations and modifications are intended to be included within the scope of the present invention, without departing from the spirit and scope of the invention.
Claims (10)
1. A continuous annealing unit fault diagnosis method based on a deep learning algorithm is characterized by comprising the following steps:
collecting production process data of a continuous annealing unit in real time, and acquiring operation parameters of the continuous annealing unit;
coupling characteristic relations among strip steel slipping, strip steel deviation and strip steel hot buckling by using a fault mechanism mathematical model module, and performing synchronous coupling calculation or same-parameter regulation and control characteristic analysis on the operation parameters of the continuous annealing unit to obtain the influence rule of single-parameter regulation and control of the continuous annealing unit on different fault information and the influence rule of multiple-parameter regulation and control on the same fault information, obtain the corresponding relation between the operation parameters of the continuous annealing unit and the different fault information, and obtain a coupling relation curve and an influence coefficient between each regulation and control parameter of the continuous annealing unit and the different fault information;
the deep learning module adopts a stack type self-coding algorithm to construct a stack type self-coding neural network structure to form a plurality of classification fault diagnosis models, and performs big data deep training on the regulation and control parameters and the influence coefficients thereof corresponding to different fault information acquired by the fault mechanism mathematical model module to obtain a regulation and control parameter optimization interval corresponding to each self-learned regulation and control parameter, and continuously provides the fault mechanism mathematical model module with a regulation parameter, a regulation curve or an influence coefficient matrix which is in accordance with the current working condition;
the unit fault diagnosis control module compares the regulation and control parameters acquired in real time with the corresponding regulation and control parameter optimizing intervals in real time, judges whether the regulation and control parameters acquired in real time are in the corresponding regulation and control parameter optimizing intervals, judges that the continuous annealing unit operates abnormally if the regulation and control parameters acquired in real time are not in the corresponding regulation and control parameter optimizing intervals, and judges that the continuous annealing unit operates normally if the regulation and control parameters acquired in real time are not in the corresponding regulation and control parameter optimizing intervals;
the fault information comprises one or more of a slip fault, a deviation fault and a thermal buckling fault; the operation parameters comprise one or more of motor rotating speed, tension of the strip steel before and after, strip steel speed, annealing temperature, annealing speed, furnace roller size, furnace roller tension, furnace roller load, set tension of the unit, acceleration of the central section of the unit and strip shape of the strip steel before annealing.
2. The continuous annealing unit fault diagnosis method based on the deep learning algorithm as claimed in claim 1, wherein if the continuous annealing unit is judged to be abnormal in operation, the unit fault diagnosis control module sends real-time fault information and solution to the front-end display module, and the front-end display module performs real-time voice alarm and/or digital display.
3. The method for diagnosing the fault of the continuous annealing unit based on the deep learning algorithm as claimed in claim 1, wherein if the continuous annealing unit is judged to be abnormal in operation, the unit fault diagnosis control module sends control signals to various control systems on site to realize online feedback control of fault information of the continuous annealing unit.
4. The method for diagnosing the fault of the continuous annealing unit based on the deep learning algorithm as claimed in claim 1, further comprising:
a real-time database is arranged, receives and stores the production process data of the continuous annealing unit collected by the data collection module in real time, and transmits the production process data to the failure mechanism mathematical model module and the deep learning module; and
setting a relational database, wherein the relational database is respectively in data transmission with the data acquisition module, the fault mechanism mathematical model module, the deep learning module and the unit fault diagnosis control module; the relational database receives and stores the production process data of the continuous annealing unit acquired by the data acquisition module in real time, and transmits the production process data to the fault mechanism mathematical model module, the deep learning module and the unit fault diagnosis control module; meanwhile, the relational database stores the calculation result of the fault mechanism mathematical model module and the learning result of the deep learning module, and stores typical case data, wherein the typical case data are production process data of the continuous annealing unit in a set time period of automatic identification and screening of the unit fault diagnosis control module during operation, and the typical case data comprise a stable operation sample library and a typical fault sample library, the stable operation sample library is used for storing the production process data of the continuous annealing unit in a stable operation state, and the typical fault sample library is used for storing the production process data of the continuous annealing unit in various fault states.
5. The method for diagnosing the faults of the continuous annealing unit based on the deep learning algorithm as claimed in claim 4, wherein when the continuous annealing unit is abnormally operated, the unit fault diagnosis control module classifies and compares real-time fault information with typical case data stored in the relational database to obtain specific parameters and solutions causing the abnormal operation of the continuous annealing unit, voice broadcasting and/or digital display is performed through a front-end display module, and closed-loop processing of the fault information is realized through a parameter adjustment mode.
6. A continuous annealing unit fault diagnosis system based on a deep learning algorithm is characterized by comprising:
the data acquisition module is configured to acquire production process data of the continuous annealing unit in real time and acquire operation parameters of the continuous annealing unit;
the fault mechanism mathematical model module is connected with the data acquisition module, comprises a plurality of mechanism analysis models including at least a slip fault model, a deviation fault model and a thermal buckling fault model, and is configured to obtain the influence rule of the single-parameter regulation and control of the continuous annealing unit on different fault information and the influence rule of a plurality of parameter regulation and control on the same fault information by carrying out synchronous coupling calculation or same-parameter regulation and control characteristic analysis on the operation parameters of the continuous annealing unit, obtain the corresponding relation between the operation parameters of the continuous annealing unit and different fault information and obtain the coupling relation curve and the influence coefficient between each regulation and control parameter of the continuous annealing unit and different fault information;
the deep learning module is connected with the fault mechanism mathematical model module, is configured to adopt a stacked self-coding algorithm, constructs a stacked self-coding neural network structure, forms a plurality of classified fault diagnosis models, performs big data deep training on the regulation parameters and the influence coefficients thereof corresponding to different fault information acquired by the fault mechanism mathematical model module, obtains the corresponding regulation parameter optimization sections after the self-learning of each regulation parameter, and continuously provides the fault mechanism mathematical model module with the regulation parameters, the regulation curves or the influence coefficient matrixes which accord with the current working conditions;
the unit fault diagnosis control module is connected with the data acquisition module, the deep learning module and the fault mechanism mathematical model module and is configured to diagnose whether the continuous annealing unit normally operates by comparing the regulation parameters with the corresponding regulation parameter optimizing intervals in real time;
the fault information comprises one or more of a slip fault, a deviation fault and a thermal buckling fault; the operation parameters comprise one or more of motor rotating speed, tension of the strip steel before and after, strip steel speed, annealing temperature, annealing speed, furnace roller size, furnace roller tension, furnace roller load, set tension of the unit, acceleration of the central section of the unit and strip shape of the strip steel before annealing.
7. The continuous annealing unit fault diagnosis system based on the deep learning algorithm, according to claim 6, wherein the data collection module comprises one or more of a bottom layer device sensor, a field PLC device and a DCS device, wherein the bottom layer device sensor comprises one or more of a temperature sensor, a tension sensor, a rolling force sensor and a speed sensor.
8. The system of claim 6, wherein the unit fault diagnosis control module is further configured to interact with the continuous annealing unit to realize a closed loop, and is configured to send control signals to various control systems on site to realize online feedback control of fault information of the continuous annealing unit.
9. The continuous annealing unit fault diagnosis system based on the deep learning algorithm is characterized by further comprising a large field database, wherein the large field database comprises a real-time database and a relational database; wherein, the first and the second end of the pipe are connected with each other,
the real-time database is configured to receive and store the production process data of the continuous annealing unit, which is acquired by the data acquisition module in real time, and output the production process data to the input end of the fault mechanism mathematical model module and the input end of the deep learning module;
the relational database is respectively connected with the data acquisition module, the fault mechanism mathematical model module, the deep learning module and the unit fault diagnosis control module, and is configured to receive and store production process data of the continuous annealing unit acquired by the data acquisition module in real time and transmit the production process data to the fault mechanism data model module, the deep learning module and the unit fault diagnosis control module; meanwhile, the relational database stores the calculation result of the failure mechanism data model module and the learning result of the deep learning module, and stores the production process data of the continuous annealing unit in a stable operation state and in a set time period under various failure states.
10. The continuous annealing unit fault diagnosis system based on the deep learning algorithm according to claim 6, wherein the continuous annealing unit fault diagnosis system further comprises a front end display module, the front end display module is connected with the unit fault diagnosis control module and is configured to realize real-time display of continuous annealing unit real-time operation information and/or fault information, and/or intelligent voice alarm, and/or manual fault adjustment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110436732.5A CN113325819B (en) | 2021-04-22 | 2021-04-22 | Continuous annealing unit fault diagnosis method and system based on deep learning algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110436732.5A CN113325819B (en) | 2021-04-22 | 2021-04-22 | Continuous annealing unit fault diagnosis method and system based on deep learning algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113325819A CN113325819A (en) | 2021-08-31 |
| CN113325819B true CN113325819B (en) | 2022-08-19 |
Family
ID=77414967
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110436732.5A Active CN113325819B (en) | 2021-04-22 | 2021-04-22 | Continuous annealing unit fault diagnosis method and system based on deep learning algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113325819B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113920664A (en) * | 2021-09-24 | 2022-01-11 | 上海泰讯数据科技有限公司 | Digital electronic fence control method, module and control system |
| CN117666546B (en) * | 2024-01-31 | 2024-05-03 | 中核武汉核电运行技术股份有限公司 | Distributed control system fault diagnosis method and device |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS6289820A (en) * | 1985-10-14 | 1987-04-24 | Kawasaki Steel Corp | Method for diagnosing apparatus in continuous annealing furnace |
| EP1482699A3 (en) * | 1998-06-08 | 2005-02-09 | Telefonaktiebolaget LM Ericsson (publ) | Burst structure for multicarrier transmission, and synchronisation of bursts, symbols and frequency |
| US7353431B2 (en) * | 2004-02-11 | 2008-04-01 | Sun Microsystems, Inc. | Method and apparatus for proactive fault monitoring in interconnects |
| CN102033523B (en) * | 2009-09-25 | 2014-01-01 | 上海宝钢工业检测公司 | Strip steel quality forecasting, furnace condition early-warning and fault diagnosis method based on partial least square |
| CN107541597B (en) * | 2016-06-29 | 2019-03-05 | 宝山钢铁股份有限公司 | Strip running deviation monitoring and diagnostic method and the system of continuous annealing unit soaking pit |
| CN112029990B (en) * | 2020-07-16 | 2022-04-08 | 宝钢湛江钢铁有限公司 | Automatic deviation rectifying control method for cold-rolled strip steel continuous annealing heating furnace |
| CN112430727B (en) * | 2020-10-15 | 2022-05-20 | 宝钢日铁汽车板有限公司 | Furnace temperature early warning method and system for continuous annealing furnace |
| CN112287592A (en) * | 2020-10-27 | 2021-01-29 | 北京理工大学 | Industrial equipment fault diagnosis method and system based on deep confidence network |
-
2021
- 2021-04-22 CN CN202110436732.5A patent/CN113325819B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113325819A (en) | 2021-08-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113325819B (en) | Continuous annealing unit fault diagnosis method and system based on deep learning algorithm | |
| CN102033523B (en) | Strip steel quality forecasting, furnace condition early-warning and fault diagnosis method based on partial least square | |
| US11755976B2 (en) | Method for predicting oxygen load in iron and steel enterprises based on production plan | |
| CN107042234A (en) | The intelligent production line and production method gathered based on bar whole process big data | |
| CN105259827B (en) | A kind of real-time monitoring system and monitoring method of solid ferment process situation | |
| CN112430727B (en) | Furnace temperature early warning method and system for continuous annealing furnace | |
| US6546310B1 (en) | Process and device for controlling a metallurgical plant | |
| CN102151704B (en) | Stelmor line cooling method of high-speed wire by taking temperature as direct-control parameter | |
| CN113033974B (en) | Digital steel coil characteristic quality prediction and process parameter dynamic adjustment method and system based on improved LSTM network | |
| CN110322014A (en) | A kind of finished cement specific surface area prediction technique based on BP neural network | |
| CN112418538A (en) | Continuous casting billet inclusion prediction method based on random forest classification | |
| CN109884892A (en) | Process industry system prediction model based on crosscorrelation time lag grey correlation analysis | |
| CN111310830B (en) | Blocking fault diagnosis system and method for combine harvester | |
| DE102021201296A1 (en) | Prediction model for predicting product quality parameter values | |
| CN114002597A (en) | Motor fault diagnosis method and system based on GRU network stator current analysis | |
| CN113373295A (en) | Annealing furnace digital twin intelligent alarm system and method based on multilayer perceptron | |
| CN114418177B (en) | New product material distribution prediction method based on digital twin workshops for generating countermeasure network | |
| CN113705897B (en) | Product quality prediction method and system for industrial copper foil production | |
| CN106569455A (en) | PLC integrated control system-based precision machining method | |
| CN111612212A (en) | On-line optimization model updating method for coal powder fineness of coal mill | |
| CN115640980A (en) | Power grid engineering cost dynamic management system based on target control | |
| CN113435794B (en) | Intelligent monitoring method for post-processing of nozzle casting based on image processing | |
| CN115481815A (en) | Thermal power plant load distribution system and method based on neural network | |
| CN102486632A (en) | On-line analyzing method of terephthalic acid crystal particle diameter in P-xylene oxidation process | |
| CN115392104A (en) | Method for predicting mechanical property of cold-rolled continuous annealing strip steel based on annealing process |
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 |