CN111898669A - Machine learning-based direct-current submerged arc furnace abnormal event early warning system - Google Patents
Machine learning-based direct-current submerged arc furnace abnormal event early warning system Download PDFInfo
- Publication number
- CN111898669A CN111898669A CN202010725828.9A CN202010725828A CN111898669A CN 111898669 A CN111898669 A CN 111898669A CN 202010725828 A CN202010725828 A CN 202010725828A CN 111898669 A CN111898669 A CN 111898669A
- Authority
- CN
- China
- Prior art keywords
- data
- module
- machine learning
- current
- submerged arc
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Strategic Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Game Theory and Decision Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Fuzzy Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Probability & Statistics with Applications (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Vertical, Hearth, Or Arc Furnaces (AREA)
Abstract
The invention discloses a machine learning-based direct-current submerged arc furnace abnormal event early warning system which comprises a sensor module, a data acquisition module, a central processing module, a man-machine interaction module, a machine learning module, a numerical control output module and an execution mechanism module. The system is based on a machine learning technology which is located at the front of the current science and technology and is mature day by day, and mainly aims at predicting and early warning abnormal events such as material collapse, electrode soft and hard break and the like which often occur in the production process of the direct-current submerged arc furnace.
Description
Technical Field
The invention relates to the technical field of submerged arc furnace control, in particular to a machine learning-based direct-current submerged arc furnace abnormal event early warning system.
Background
At present, industrial ferroalloy direct-current smelting technologies with positive and negative electrodes arranged vertically to a hearth are few in industrial application, some related smelting technologies are in the process of searching, but the furnace type has already been brought into effect on the existing rare production line, the advantages of energy conservation and environmental protection are obvious, compared with an alternating-current furnace, the furnace type has the advantages of high power factor, small short-network reactance, no skin effect and the like, and the powder adaptability is good; compared with a direct current furnace adopting a bottom electrode technology, the direct current furnace has no bottom electrode consumption, so that the regular bottom electrode overhaul work is not needed, and the continuity of smelting production can be guaranteed.
However, in the production and smelting process of a vertical positive and negative electrode direct current electric furnace, according to the principle of electron bombardment, the heat productivity of the positive electrode anode region is larger than that of the negative electrode cathode region, so that the roasting characteristics of the two vertical electrodes are different, and the hard and soft break accidents of the electrodes often occur. Meanwhile, the submerged arc furnace is smelted in the body hearth to perform complex chemical reduction reaction, the process is hidden and difficult to measure, and large ambiguity, randomness and uncertainty exist, and if the operation is not proper, the accidents of material collapse and electrode soft and hard break often occur, the two abnormal events are the most frequently occurring accidents of the direct-current submerged arc furnace, and the normal production of the submerged arc furnace is seriously influenced. Once an accident occurs, a light person consumes a great deal of manpower, financial resources and time; serious people can damage some equipment of the submerged arc furnace body, even generate high-temperature smoke or hot melt to be sprayed out of the furnace, so that operating personnel in dangerous areas burn and scald, furnace body explosion and the like are caused, and the consequences are very serious.
The machine learning intelligence is put into use in industrial production, data is used as strategic resources, centralized management of equipment-associated data is completed with the help of the support of a big data processing system, expression and processing are performed through the data, an industrial data world is established, an algorithm model is applied, processing and analysis are performed, the core value of the industry is mined, the optimal production experience is precipitated, a production target is used as a guide, production process parameters are adjusted, the production flow is controlled reversely, and finally the preset production index is achieved.
The submerged arc furnace has the disadvantages of high equipment price, large maintenance work amount and huge economic loss after shutdown and production stoppage. At present, two abnormal events, namely material collapse and electrode soft and hard breakage, which often occur in the production process of the submerged arc furnace are difficult to predict, corresponding treatment is generally carried out after an accident occurs, the smooth production is difficult to guarantee, and the abnormal events can be avoided as much as possible only by means of artificial production experience, for example, a conservative capacity reduction mode is adopted, but the two abnormal events still occur at present and cannot be avoided at all, and meanwhile, different process personnel have different understanding and operation methods of the process of the DC submerged arc furnace, so that the possibility of the occurrence of the two abnormal events is increased. At present, a control mode of adopting a fixed threshold value of the electrode current is adopted to try to solve the problem, the effect of the control mode is improved compared with that of a simple manual experience, but the effect of the control mode cannot meet the production requirement due to the influence of complex furnace conditions, raw material components, environmental factors and the like in the production process, abnormal events still occur at times, and the problem is not solved fundamentally. Along with the evolution of scientific technology, equipment maintenance means are gradually developed from passive maintenance to intelligent predictive maintenance, so that the production loss and the maintenance cost of equipment caused by unplanned shutdown are reduced. Original manual maintenance and fixed threshold maintenance depend on experience and cannot meet increasingly complex working condition requirements, so that a machine learning-based system and method for early warning of abnormal events of the direct-current submerged arc furnace are urgently needed to be developed, abnormal information can be sensed before equipment fails by means of fault prediction combining big data analysis technology and mechanism analysis, and loss caused by unplanned shutdown is reduced.
Disclosure of Invention
According to the problems in the prior art, the invention discloses a machine learning-based abnormal event early warning system for a direct-current submerged arc furnace, which comprises a sensor module, a signal processing module and a signal processing module, wherein the sensor module is used for sensing and detecting multiple physical data parameters of the direct-current submerged arc furnace in the production process and converting the sensed information into electric signals according to a certain rule;
the data acquisition module receives the electric signals transmitted by the sensor module according to a certain period, and performs data processing and output on the received electric signals by adopting anti-shake, filtering and passivation methods;
the central processing module is used for receiving the data information transmitted by the data acquisition module, and performing logic control on the received signals and adjusting the control strategy of the production process;
the human-computer interaction module receives the logic control information transmitted by the central processing module, displays the received data information in real time, sends a control instruction of a process operator to the central processing module for logic control, and realizes bidirectional data communication with the central processing module;
the machine learning module is used for analyzing and extracting characteristics of a mass of historical data measuring points generated in smelting production of the direct current furnace, estimating the output characteristics of the current electrode current in real time, acquiring the residual error of the output characteristics of the current electrode current according to the actual output characteristics of the current electrode current and the estimated output characteristics of the electrode current of the direct current submerged arc furnace model, setting a proper threshold value in the residual error of the output characteristics of the current electrode current as a judgment basis of material collapse and abnormal events of electrode soft and hard break, and outputting an early warning signal when the residual error of the characteristics is greater than the set threshold value; the machine learning module analyzes and models massive historical data measuring points produced in direct current furnace smelting by means of an equipment fault intelligent prediction engine, specifically establishes prediction algorithms of two abnormal events of material collapse and electrode soft and hard break in direct current furnace smelting, and performs reverse verification on the algorithms by using historical data so as to verify the prediction algorithms of the material collapse and electrode soft and hard break accidents and perform early warning analysis on the abnormity; the machine learning module is also used for carrying out big data processing, feature extraction, analysis learning and service application on the received logic control data and feeding back learning result information to the central processing module;
the numerical control output module receives a control instruction sent by the central processing module, and converts information into various physical signals to be output;
and the executing mechanism module is used for receiving the physical signals transmitted by the numerical control output module and implementing the executing actions corresponding to the physical signals on various devices in the production process of the direct-current submerged arc furnace.
Further, the machine learning module comprises a feature extraction module, an analysis learning module and a business application module;
the feature extraction module adopts an unsupervised machine learning method to carry out principal component analysis on data: firstly, receiving original data of the direct-current submerged arc furnace, carrying out zero-valued processing on the data, solving a covariance matrix of sampling characteristics, carrying out diagonalization processing on the covariance matrix, constructing a diagonal matrix with new characteristic covariance of 0 so as to calculate a characteristic set matrix consisting of new standard orthogonal basis vectors, and reserving new characteristics with high principal component contribution rate;
the analysis learning module adopts a supervised machine learning method to train and test the direct-current submerged arc furnace model: firstly, receiving characteristic data of the direct-current submerged arc furnace transmitted by a characteristic extraction module, carrying out normalization processing on the characteristic data, training an XGboost model by adopting a training data set, and carrying out parameter optimization on the XGboost model by adopting a test data set so as to finish characteristic classification;
and the business application module receives the abnormal data transmitted by the analysis and learning module to carry out state monitoring and fault diagnosis on the direct-current submerged arc furnace.
And the business application module receives the abnormal data transmitted by the analysis and learning module to carry out state monitoring and fault diagnosis on the direct-current submerged arc furnace.
Further, the machine learning module is used for acquiring historical data and accident recording data of each measuring point from an existing system on the basis of massive multidimensional data acquired and accumulated in the production process of the direct-current submerged arc furnace, an intelligent analysis engine is built in a server, machine learning and deep mining are carried out on the big data by using elastic data storage, computing and processing capacity and an artificial intelligence technology of the big data server, data exploration and characteristic modeling are carried out, and a model direct-current furnace algorithm result is generated and comprises an accident prediction model and an accident early warning model, wherein the accident prediction model comprises a material collapse prediction model and an electrode soft and hard failure prediction model, and the accident early warning model comprises a material collapse early warning module and an electrode soft and hard failure early warning model.
Further, the machine learning module constructs the measuring point data of the direct-current submerged arc furnace in a normal smelting state into a memory matrix, the row vector of the matrix represents the operation data of all measuring points at a certain moment, the column vector represents the operation data of the measuring points at different moments, and when the deviation between the real-time output and the expected output based on the current input is larger than a set threshold value, an abnormal state early warning is sent out.
Further, the machine learning module queries an abnormal mode deviating from a normal state by a big data simulation method and simultaneously queries abnormal data characteristics by adopting a mode of combining unsupervised learning and supervised learning according to a working principle that input and output state information has a certain mapping relation in a healthy running state of the direct current furnace.
Further, when the machine learning module performs reverse verification on the algorithm by using the historical data, the historical data is obtained in the following manner: and intercepting a historical data segment with a fixed time length before the occurrence time of the typical abnormal event for multiple times, and taking the historical data segment as input data of the direct current furnace.
Due to the adoption of the technical scheme, the direct-current submerged arc furnace abnormal event early warning system based on machine learning provided by the invention has the advantages that the positive and negative electrode pairs in the direct-current submerged arc furnace are arranged in pairs, and each pair of electrodes is not mutually restricted basically. Therefore, an abnormal event early warning system and an abnormal event early warning method are developed for each positive electrode pair and each negative electrode pair, early warning of the abnormal events of material collapse and electrode soft and hard break is achieved, the control strategy is communicated to a main control system to be adjusted, the overall intelligent level of the direct-current submerged arc furnace is improved, powerful guarantee is provided for continuous production, and huge economic losses brought to users due to the material collapse and the electrode soft and hard break can be avoided or reduced.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments described in the present application, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a schematic diagram of the system of the present invention;
FIG. 2 is a schematic diagram of a machine learning early warning method of the present invention;
FIG. 3 is a schematic diagram of the operation of the feature extraction module of the present invention;
FIG. 4 is a schematic diagram of the operation of the analysis learning module of the present invention;
FIG. 5 is a schematic diagram of a big data machine learning process of the DC submerged arc furnace according to the present invention.
Detailed Description
In order to make the technical solutions and advantages of the present invention clearer, the following describes the technical solutions in the embodiments of the present invention clearly and completely with reference to the drawings in the embodiments of the present invention:
as shown in fig. 1, the system for early warning of abnormal events of a dc submerged arc furnace based on machine learning is based on a machine learning technology which is at the front of the current technology and is becoming mature day by day, and is mainly used for predicting and early warning of abnormal events such as material collapse, electrode soft and hard break and the like which often occur in the production process of the dc submerged arc furnace, and the basic principle thereof is as follows: various accidents of the submerged arc furnace in the smelting production process generally accompany with the change of a plurality of parameters, main change parameters of the various accidents and other parameter combinations with higher correlation with the parameters are found out, other parameter combinations can be established as input parameters, the main change parameters are used as input and output models of the output parameters, and the models are controlled to realize accurate control of the work of the DC submerged arc furnace, so that the early warning problem of the accidents of material collapse and electrode soft and hard break in the prior art is solved, and huge economic loss brought to users due to the occurrence of the accidents of the material collapse and the electrode soft and hard break is avoided or reduced. The system specifically comprises a sensor module M1, a data acquisition module M2, a central processing module M3, a man-machine interaction module M4, a machine learning module M5, a numerical control output module M6 and an execution mechanism module M7. The machine learning module M5 further comprises three sub-modules, namely a feature extraction module M5-1, an analysis learning module M5-2 and a business application module M5-3, wherein each module adopts the following working mode:
sensor module M1: the sensing detection is carried out on various physical data parameters in the production process, and the sensed information is converted into electric signals according to a certain rule to be output, so that the electric signals can be utilized by the data acquisition module M2.
The data acquisition module M2: the output signals of the sensor module M1 are collected according to a certain period, the collected signals are processed by anti-shake, filtering, passivation and other methods to improve the availability of data, and then the collected data are output to the central processing module M3.
Central processing module M3: the signals collected from the data acquisition module M2 are subjected to logic control processing, then the processed real-time information is transmitted to the man-machine interaction module M4 for process operators to check, and control instructions sent by the man-machine interaction module M4 are received to enter the logic control of the central processing module M3, so that the bidirectional data communication with the man-machine interaction module M4 is completed, and real man-machine friendly interaction is achieved. The central processing module M3 sends a large amount of data to the machine learning module M5, then to the machine learning module M5 for big data processing, feature extraction, analytical learning, and business application, and then the machine learning module M5 returns the learning outcome information to the central processing module M3; the central processing module M3 processes the received prediction information and adjusts the production process control strategy, and transmits a control signal to the execution mechanism module M7 through the numerical control output module M6 to complete the final real action of the field device, and then the data parameters of the production field process change along with the change of the data parameters and are sensed by the sensor module M1 to repeat in cycles, so that the reliable, normal and optimal operation of the whole production system is finally ensured.
Human-computer interaction module M4: friendly communication with process operators is realized, the information received from the central processing module M3 is displayed in real time, and the control instructions of the process operators are sent to the central processing module M3.
Machine learning module M5: a large amount of data is acquired in the central processing module M3, then the functions of big data processing, feature extraction, analysis learning and business application of the early warning system are carried out in sequence, and after a single machine learning task is completed, the information of the predicted result is reversely transmitted back to the central processing module M3.
Numerical control output module M6: the real-time receiving central processing module M3 sends out control command, and converts the information into various physical signals to the execution mechanism module M7.
The actuator module M7: and various physical signals sent by the numerical control output module M6 are received, and various equipment actions in the production process of the direct-current submerged arc furnace are completed.
Further, as shown in fig. 2, the machine learning module M5 has the following warning method: the data in the production process of the direct current submerged arc furnace read in the central processing module M3 are used as follows: the method comprises the steps of selecting electrode current as output characteristics by an early warning model according to parameters such as raw material ratio, electrode electrical parameters, furnace bottom temperature, furnace pressure, furnace gas components and electrode vibration, establishing a material collapse and electrode soft and hard break early warning model according to steps in an algorithm principle by splitting a training set and a testing set, and then predicting by using the model.
The historical data 201 is obtained as follows: based on massive multidimensional data collected and accumulated in the production process of the direct-current submerged arc furnace, historical data 201 of each measuring point is obtained from the existing system.
The acquisition mode of the input data 202 of the direct current furnace is as follows: in the dc furnace historical data 201, a historical data segment having a fixed time length before the occurrence time of a typical abnormal event is intercepted a plurality of times as dc furnace input data 202 to perform a process of feature calculation 203, thereby extracting corresponding features.
The electrode current output data 204 is obtained in the following manner: in the dc furnace historical data 201, a historical data segment having a fixed time length before the occurrence time of a typical abnormal event is captured a plurality of times, and the current data segment is in one-to-one correspondence with the dc furnace input data 202, and is used as electrode current output data 204 to perform feature calculation 205 to extract corresponding features.
The training process of the direct current furnace model comprises the following steps: and (4) performing direct current furnace model training 206 in a machine learning mode according to the direct current furnace input data characteristics and the electrode current output data characteristics, and storing the model in a direct current furnace model prediction 209 unit. The real-time data 207 is from real-time data collected from the direct-current submerged arc furnace in the current actual production smelting process. The actual input 208 of the dc furnace is derived from real-time data 207 of the dc submerged arc furnace, and the specific parameters are the same as the input data 202 of the dc furnace in the historical data 201.
Further, the direct current furnace model prediction 209 process is: the unit receives model parameters generated by the direct current furnace model training 206, stores and utilizes the model parameters, calls the model to predict according to the actual input 208 characteristics of the direct current furnace, and estimates the current electrode current output characteristics in real time, as shown in fig. 2.
Further, the process of estimating the electrode current output 210 is as follows: the direct current furnace model prediction unit 209 calculates the predicted characteristics based on the direct current furnace actual input 208 data.
Further, the electrode current actual output 211 process: and (3) real-time data 207 from the direct-current submerged arc furnace, wherein specific parameters are the same as electrode current output data 204 in historical data, characteristic calculation is carried out, and corresponding characteristics are extracted.
Further, the electrode current residual calculation 213 process: and (4) aiming at the real-time calculated electrode current actual output 211 characteristic and the model estimated electrode current output characteristic, making an absolute value difference of the two characteristics, namely an electrode current output characteristic residual error.
Further, the abnormal event determination 214 process: and selecting a proper threshold value from the residual error of the electrode current output characteristic as an early warning signal of abnormal events such as material collapse, electrode soft and hard break and the like.
The explanation of the working principle of the dc submerged arc furnace abnormal event early warning system based on machine learning can be regarded as an explanation of two specific embodiments of the abnormal events of material collapse and electrode soft and hard break.
Predictability of smelting accidents of the direct-current submerged arc furnace is that certain abnormal conditions often exist in a period of time before the smelting accidents occur, and a machine learning algorithm aims to find the abnormal conditions and specifically adopts the following mode:
a local server deployment mode is adopted for machine learning analysis of data, a large data processing edge server is deployed locally, the computing capability and the speed are obviously superior to those of a common PC, and the method is used for research, modeling and development of a support algorithm and is easier to realize a control algorithm. Therefore, based on massive multidimensional data collected and accumulated in the production process of the direct-current submerged arc furnace, historical data and accident record data of each measuring point are obtained from the existing system, data communication is transmitted to a big data server, an intelligent analysis engine is built in the server, and machine learning and deep mining are carried out on the big data by utilizing the elastic data storage, the computing and processing capacity and the artificial intelligence technology of the big data server, so that data exploration and feature modeling are carried out. By means of an equipment fault intelligent prediction engine, mass historical data measuring points produced in direct current furnace smelting are analyzed and modeled, prediction algorithms of two abnormal events of material collapse and electrode soft and hard break in direct current furnace smelting are established in a targeted mode, the historical data are used for carrying out reverse verification on the algorithms, the prediction algorithms of the material collapse and electrode soft and hard break accidents are verified, and early warning analysis is carried out on the abnormal events.
Furthermore, the measuring point data of the direct-current submerged arc furnace in the normal smelting state are constructed into a memory matrix, the row vector of the matrix represents the operation data of all measuring points at a certain moment, and the column vector represents the operation data of the measuring points at different moments. And when the real-time output is greatly deviated from the expected output based on the current input, sending out an abnormal state early warning.
Under the healthy running state of the direct current furnace, a certain mapping relation exists between input and output state information. When the relation deviates, the expected mapping relation among the parameters is damaged, and the smelting process may be in an abnormal state. In the machine learning algorithm research of the design, the data characteristics of a sample set are combined, a mode of combining unsupervised learning and supervised learning is adopted, an abnormal mode deviating from a normal state is found out through big data simulation, and meanwhile, abnormal data characteristics are found out.
Unsupervised learning process example:
in the data feature extraction mentioned many times in the description of the machine learning process of the invention, we adopt an unsupervised learning mode, we use a characteristic value decomposition (EVD) method to perform Principal Component Analysis (PCA) on data to reduce the characteristic dimension of a research object, and at the same time, we need to reduce the inevitable information loss in the dimension reduction process as much as possible, the most effective way is to use a group of new features which are linearly independent and have fewer dimensions to replace the original sampling features, as shown in fig. 3, the specific calculation steps are as follows:
s201: firstly, acquiring original data of the direct-current submerged arc furnace;
s202: carrying out data zero-mean processing on the original data, namely data centralization;
s203: then, a covariance matrix of the sampling characteristics is obtained;
s204: then carrying out diagonalization processing on the covariance matrix;
s205: constructing a diagonal matrix with the covariance of the new characteristics being 0;
s206: calculating through the diagonal matrix to obtain a feature set matrix consisting of new orthonormal basis vectors;
s207: discarding unimportant features from the new feature set which is independent of each other, and reserving new features with higher principal component contribution rate;
s208: therefore, the main information is described for the research object, the calculation amount of subsequent data processing is greatly reduced, the characteristic calculation process is completed, and the dimension reduction is finished.
Certainly, we can also obtain the feature set matrix composed of the target standard orthogonal basis vectors by directly obtaining all component achievement methods of singular value decomposition at one time through tools provided by Python or R and other languages, the two methods of the characteristic value decomposition (EVD) and the Singular Value Decomposition (SVD) are essentially real symmetric square matrixes based on the multiplication result of the sampling matrix and the transpose matrix thereof, and the feature set matrix composed of the new target standard orthogonal basis vectors is obtained by using the excellent characteristics of the square matrixes, which is not described herein again.
Supervised learning process example:
after data feature extraction and dimensionality reduction, a supervised learning mode is adopted for training and testing a direct-current submerged arc furnace model. And selecting a proper time length to split the data characteristics of the direct current furnace material collapse and electrode soft and hard break abnormal events, eliminating data of a fault elimination time period, marking a sample in the previous period of the accident occurrence time as an abnormal sample, and marking data in other times as a normal sample. After data splitting, two classification models can be respectively established for two accidents, and the XGBoost classification algorithm is used for establishing the two classification models, and the specific steps are shown in FIG. 4.
S211: firstly, acquiring characteristic data of the direct-current submerged arc furnace;
s212: then, normalization processing is carried out on the characteristic values to obtain dimensionless characteristic data, so that indexes of different units or magnitude levels can be conveniently compared and weighted, and the normalization is beneficial to better utilizing the data by the model;
s213: grouping the samples, wherein all the obtained sample data are divided into training samples and testing samples according to a ratio of 4:1 after being arranged and grouped;
s214: forming a training data set by training samples;
s215: the test samples form a test data set;
s216: training an XGboost model, wherein the idea of the XGboost algorithm is to divide and continuously add trees according to characteristics, one added tree is actually to learn a new function every time, the residual error predicted last time is fitted, leaf nodes of the trees correspond to a score, and finally the score corresponding to each tree is added to form the predicted value of the sample; training sample data is substituted into an XGboost model, loglos is selected as a loss function, a regularization term is added on the basis, a target function is subjected to secondary Taylor expansion, all values of all characteristics are traversed as splitting points on the basis of the regularization term, the point with the maximum gain is selected for splitting, a fault splitting tree is iteratively constructed by continuously splitting the characteristics until the regularization requirement of splitting is met, the splitting of the tree is not performed, and the model is trained;
s217: performing parameter tuning on the XGboost model by using test set data, and selecting a parameter corresponding to the highest accuracy of model classification under each group of parameters as an optimal parameter;
s218: and finishing the establishment of the feature classification model.
The XGboost algorithm can still keep a good training effect under the condition of large data volume, and finally, data to be tested in real time are substituted into a trained model to detect the abnormal operation state of the current direct-current submerged arc furnace, whether the accidents of material collapse and electrode soft and hard break are about to occur or not is predicted, and then the data are fed back to a basic automatic control system to adjust the smelting strategy of the direct-current submerged arc furnace in time, so that the abnormal events in the smelting process of the direct-current furnace are avoided.
In summary, a machine learning-based method for early warning of abnormal events of a direct-current submerged arc furnace includes the following specific complete steps:
s1, a sensor module M1 composed of a temperature sensor, a pressure sensor, a vibration sensor, a position transmitter, an electric energy detector, a limit switch and the like in the production field of the direct-current submerged arc furnace senses and detects various physical data parameters in the production process, converts sensed information into electric signals according to a certain rule, and transmits the electric signals to a data acquisition module M2 composed of an ET200SP and a frequency converter through a data cable or a communication cable;
s2, the data acquisition module M2 acquires the output signal of the sensor module M1 according to a certain period, and then transmits the acquired data to the central processing module M3 through an RJ45 Ethernet cable or an optical fiber via a factory bus switch;
s3, after the central processing module M3 carries out logic control processing on the signals collected from the data acquisition module M2, the signals are simultaneously transmitted to the man-machine interaction module M4 and the machine learning module M5 through an RJ45 Ethernet cable or an optical fiber via a terminal bus switch, the man-machine interaction module M4 executes the step 7, and the machine learning module M5 carries out big data machine learning of the direct-current submerged arc furnace according to the early warning method principle shown in the figures 3 and 5;
s4, in the machine learning process, the machine learning module M5 firstly extracts the features of each part of data according to the steps shown in the figure 2-1 by the feature extraction module M5-1 and adopts an unsupervised machine learning method;
s5, the extracted features are further transmitted to an analysis learning module M5-2, training and testing of the direct-current submerged arc furnace model are carried out according to the steps shown in the figure 4, and a supervised machine learning method is adopted;
s6, abnormal data appearing in the analysis learning module M5-2 are transmitted to the business application module M5-3, the business application module M5-3 realizes the functions of state monitoring, fault diagnosis and the like, and carries out data communication and seamless butt joint with the central processing module M3;
s7: the man-machine interaction module M4 displays important information in the central processing module M3 in real time, receives control instructions of process operators and reversely transmits the control instructions to the central processing module M3;
s8: when the early warning information of the abnormal event is received, the central processing module M3 automatically switches the production process control strategy of the direct-current submerged arc furnace, or receives the adjustment strategy instruction from the human-computer interaction module M4 process operator;
s9: the central processing module M3 is transmitted to the numerical control output module M6 through a factory bus switch by RJ45 ethernet cable or optical fiber;
s10: the control signal is transmitted to an execution mechanism module M7 consisting of an electromagnetic valve, a motor and the like through the numerical control output module M6, and the real action of the field equipment of the direct-current submerged arc furnace is completed.
S11: the data parameters of the production field process are changed along with the change of the production field process data, and are sensed by the sensor module M1, the process is continuously circulated, and the reliable, stable and optimal operation of the whole production system is always ensured.
As shown in fig. 3, the dc furnace algorithm effort 301 includes the generation of an accident prediction model 302 and an accident warning model 303.
The accident prediction model comprises a slump prediction model 304 and an electrode soft and hard break prediction model 305, and the recall rate on the verification effect is greater than 95%;
the accident early warning model comprises a material collapse early warning model 306 and an electrode soft and hard break early warning model 307, and the accident is early warned in advance through the existence of real-time residual errors in a period of time before the accident and dynamic threshold analysis. The material collapse accident is early warned 1 hour in advance; the electrode soft and hard break accident is early warned 4 hours in advance.
The machine learning-based direct-current submerged arc furnace early warning system and the machine learning-based direct-current submerged arc furnace early warning method are designed and developed, wherein part of related technologies can be partially applied and extended to an alternating-current submerged arc furnace control system, and the part of technologies are also within the protection range of the scheme and are all constrained by the technical scheme.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.
Claims (6)
1. The utility model provides a hot stove abnormal event early warning system in direct current ore deposit based on machine learning which characterized in that includes:
a sensor module (M1) which senses and detects a plurality of physical data parameters of the direct-current submerged arc furnace in the production process and converts the sensed information into electric signals according to a certain rule;
the data acquisition module (M2) receives the electric signals transmitted by the sensor module (M1) according to a certain period, and the data acquisition module (M2) performs data processing on the received electric signals by adopting an anti-shake, filtering and passivating method and outputs the processed electric signals;
a central processing module (M3) for receiving the data information transmitted by the data acquisition module (M2), wherein the central processing module (M3) performs logic control on the received signals and adjusts the control strategy of the production process;
the human-computer interaction module (M4) is used for receiving the logic control information transmitted by the central processing module (M3), the human-computer interaction module (M4) is used for displaying the received data information in real time, sending a control instruction of a process operator to the central processing module (M3) for logic control, and realizing bidirectional data communication with the central processing module (M3);
the machine learning module (M5) is used for receiving logic control data transmitted by the central processing module (M3), the machine learning module (M5) analyzes and extracts characteristics of a mass historical data measuring point produced by smelting of the direct current furnace, predicts the output characteristics of the current electrode current in real time, obtains an electrode current output characteristic residual error according to the actual output characteristics of the current electrode current and the electrode current output characteristics predicted by the direct current submerged arc furnace model, sets a proper threshold value in the electrode current output characteristic residual error as a judgment basis of material collapse and electrode soft and hard break abnormal events, and outputs an early warning signal when the characteristic residual error is greater than the set threshold value; the machine learning module (M5) analyzes and models massive historical data measuring points produced by direct current furnace smelting by means of an equipment fault intelligent prediction engine, specifically establishes prediction algorithms of two abnormal events of material collapse and electrode soft and hard break in the direct current furnace smelting, and reversely verifies the algorithms by using historical data so as to verify the prediction algorithms of the material collapse and electrode soft and hard break accidents and perform early warning analysis on the abnormity; the machine learning module (M5) also performs big data processing, feature extraction, analytical learning and business application on the received logic control data, and simultaneously feeds back learning result information to the central processing module (M3);
a numerical control output module (M6) for receiving control commands sent by the central processing module (M3), wherein the numerical control output module (M6) converts information into various physical signals to be output;
and the execution mechanism module (M7) receives the physical signal transmitted by the numerical control output module (M6), and the execution mechanism module (M7) implements the execution action corresponding to the physical signal on various devices in the production process of the direct-current submerged arc furnace.
2. The machine learning-based direct-current submerged arc furnace abnormal event early warning system as claimed in claim 1, further characterized in that: the machine learning module (M5) comprises a feature extraction module (M5-1), an analysis learning module (M5-2) and a business application module (M5-3);
the feature extraction module (M5-1) adopts an unsupervised machine learning method to carry out principal component analysis on data: firstly, receiving original data of the direct-current submerged arc furnace, carrying out zero-valued processing on the data, solving a covariance matrix of sampling characteristics, carrying out diagonalization processing on the covariance matrix, constructing a diagonal matrix with new characteristic covariance of 0 so as to calculate a characteristic set matrix consisting of new standard orthogonal basis vectors, and reserving new characteristics with high principal component contribution rate;
the analysis learning module (M5-2) adopts a supervised machine learning method to train and test the direct-current submerged arc furnace model: firstly, receiving characteristic data of the direct-current submerged arc furnace transmitted by a characteristic extraction module (M5-1), carrying out normalization processing on the characteristic data, training an XGboost model by adopting a training data set, and carrying out parameter tuning on the XGboost model by adopting a test data set so as to finish characteristic classification;
and the business application module (M5-3) receives the abnormal data transmitted by the analysis learning module (M5-2) to carry out state monitoring and fault diagnosis on the direct-current submerged arc furnace.
3. The machine learning-based direct-current submerged arc furnace abnormal event early warning system as claimed in claim 2, further characterized in that: the machine learning module (M5) acquires historical data and accident record data of each measuring point from the existing system on the basis of massive multidimensional data collected and accumulated in the production process of the direct-current submerged arc furnace, by establishing an intelligent analysis engine in a server, performing machine learning and deep mining on big data by utilizing the elastic data storage, the computing and processing capacity and the artificial intelligence technology of a big data server, performing data exploration and feature modeling, generating a direct current furnace algorithm result (301), the direct current furnace algorithm achievement (301) comprises an accident prediction model (302) and an accident early warning model (303), the accident prediction model (302) comprises a collapse prediction model (304) and an electrode soft and hard break prediction model (305), and the accident early warning model (303) comprises a collapse early warning module (306) and an electrode soft and hard break early warning model (307).
4. The machine learning-based direct current submerged arc furnace abnormal event early warning system according to claim 3, further characterized in that: the machine learning module (M5) constructs the measuring point data of the direct current submerged arc furnace in the normal smelting state into a memory matrix, the row vector of the matrix represents the operation data of all measuring points at a certain moment, the column vector represents the operation data of the measuring points at different moments, and when the deviation between the real-time output and the expected output based on the current input is greater than a set threshold value, an abnormal state early warning is sent out.
5. The machine learning-based direct-current submerged arc furnace abnormal event early warning system as claimed in claim 2, further characterized in that: the machine learning module (M5) queries abnormal patterns deviating from the normal state by a big data simulation method and simultaneously queries abnormal data characteristics by adopting a mode of combining unsupervised learning and supervised learning according to the working principle that input and output state information of the direct current furnace has a certain mapping relation in the healthy running state.
6. The system for early warning of abnormal events of DC submerged arc furnace based on machine learning according to any of claims 1-5, wherein the machine learning module (M5) uses the historical data to perform reverse verification on the algorithm, and the historical data is obtained by the following method: and intercepting a historical data segment with a fixed time length before the occurrence time of the typical abnormal event for multiple times, and taking the historical data segment as input data of the direct current furnace.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010725828.9A CN111898669B (en) | 2020-07-24 | 2020-07-24 | Abnormal event early warning system of direct-current submerged arc furnace based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010725828.9A CN111898669B (en) | 2020-07-24 | 2020-07-24 | Abnormal event early warning system of direct-current submerged arc furnace based on machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111898669A true CN111898669A (en) | 2020-11-06 |
| CN111898669B CN111898669B (en) | 2023-07-14 |
Family
ID=73189945
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010725828.9A Active CN111898669B (en) | 2020-07-24 | 2020-07-24 | Abnormal event early warning system of direct-current submerged arc furnace based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111898669B (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112348101A (en) * | 2020-11-16 | 2021-02-09 | 中冶赛迪重庆信息技术有限公司 | Steel rolling fuel consumption early warning method and system based on abnormal data analysis |
| CN112926749A (en) * | 2020-12-30 | 2021-06-08 | 国网宁夏电力有限公司信息通信公司 | Intelligent power grid information equipment monitoring system and method |
| CN113281581A (en) * | 2021-04-14 | 2021-08-20 | 中国重型机械研究院股份公司 | LF furnace electrode testing system and testing method |
| CN113686922A (en) * | 2021-06-28 | 2021-11-23 | 深圳爱莫科技有限公司 | Online intelligent monitoring device and method for risks of reaction kettle |
| CN113962405A (en) * | 2021-09-02 | 2022-01-21 | 大连重工机电设备成套有限公司 | Intelligent submerged arc furnace remote operation and maintenance system |
| CN114202159A (en) * | 2021-11-12 | 2022-03-18 | 寰清能源科技(上海)有限公司 | Online monitoring method and system suitable for material collapse accident of closed submerged arc furnace |
| CN114518732A (en) * | 2022-02-10 | 2022-05-20 | 宁夏瑞资联实业有限公司 | Furnace control system of DCS submerged arc furnace |
| CN116168518A (en) * | 2022-12-30 | 2023-05-26 | 大连重工机电设备成套有限公司 | Lime kiln early warning system and method based on big data analysis |
| CN116379793A (en) * | 2023-06-02 | 2023-07-04 | 青岛智控菲特软件科技有限公司 | Submerged arc furnace short-net regulation and control data processing method |
| CN116647586A (en) * | 2023-07-24 | 2023-08-25 | 山东工泵电机有限公司 | Method for realizing remote control of water and fertilizer integrated intelligent pump house by cloud computing |
| CN117303510A (en) * | 2023-09-07 | 2023-12-29 | 兰州理工大学 | Self-adaptive control method for waveform characteristic analysis of wastewater treatment power supply |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100036529A1 (en) * | 2006-10-18 | 2010-02-11 | Bp Oil International Limited | Abnormal event detection using principal component analysis |
| CN104020672A (en) * | 2014-05-29 | 2014-09-03 | 西安交通大学 | Intelligent control method of DC hollow electrode calcium carbide furnace |
| CN108460144A (en) * | 2018-03-14 | 2018-08-28 | 西安华光信息技术有限责任公司 | A kind of coal equipment fault early-warning system and method based on machine learning |
| US20200150622A1 (en) * | 2018-11-13 | 2020-05-14 | Guangdong University Of Technology | Method for detecting abnormity in unsupervised industrial system based on deep transfer learning |
-
2020
- 2020-07-24 CN CN202010725828.9A patent/CN111898669B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100036529A1 (en) * | 2006-10-18 | 2010-02-11 | Bp Oil International Limited | Abnormal event detection using principal component analysis |
| CN104020672A (en) * | 2014-05-29 | 2014-09-03 | 西安交通大学 | Intelligent control method of DC hollow electrode calcium carbide furnace |
| CN108460144A (en) * | 2018-03-14 | 2018-08-28 | 西安华光信息技术有限责任公司 | A kind of coal equipment fault early-warning system and method based on machine learning |
| US20200150622A1 (en) * | 2018-11-13 | 2020-05-14 | Guangdong University Of Technology | Method for detecting abnormity in unsupervised industrial system based on deep transfer learning |
Non-Patent Citations (1)
| Title |
|---|
| 窦克勤;叶昊;张海峰;李宏玉;: "基于主元分析的高炉异常炉况检测", 上海交通大学学报, no. 12, pages 1862 - 1867 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112348101A (en) * | 2020-11-16 | 2021-02-09 | 中冶赛迪重庆信息技术有限公司 | Steel rolling fuel consumption early warning method and system based on abnormal data analysis |
| CN112348101B (en) * | 2020-11-16 | 2023-04-07 | 中冶赛迪信息技术(重庆)有限公司 | Steel rolling fuel consumption early warning method and system based on abnormal data analysis |
| CN112926749A (en) * | 2020-12-30 | 2021-06-08 | 国网宁夏电力有限公司信息通信公司 | Intelligent power grid information equipment monitoring system and method |
| CN113281581B (en) * | 2021-04-14 | 2022-04-01 | 中国重型机械研究院股份公司 | Test method of LF furnace electrode test system |
| CN113281581A (en) * | 2021-04-14 | 2021-08-20 | 中国重型机械研究院股份公司 | LF furnace electrode testing system and testing method |
| CN113686922A (en) * | 2021-06-28 | 2021-11-23 | 深圳爱莫科技有限公司 | Online intelligent monitoring device and method for risks of reaction kettle |
| CN113962405A (en) * | 2021-09-02 | 2022-01-21 | 大连重工机电设备成套有限公司 | Intelligent submerged arc furnace remote operation and maintenance system |
| CN114202159A (en) * | 2021-11-12 | 2022-03-18 | 寰清能源科技(上海)有限公司 | Online monitoring method and system suitable for material collapse accident of closed submerged arc furnace |
| CN114518732A (en) * | 2022-02-10 | 2022-05-20 | 宁夏瑞资联实业有限公司 | Furnace control system of DCS submerged arc furnace |
| CN116168518A (en) * | 2022-12-30 | 2023-05-26 | 大连重工机电设备成套有限公司 | Lime kiln early warning system and method based on big data analysis |
| CN116379793A (en) * | 2023-06-02 | 2023-07-04 | 青岛智控菲特软件科技有限公司 | Submerged arc furnace short-net regulation and control data processing method |
| CN116379793B (en) * | 2023-06-02 | 2023-08-15 | 青岛智控菲特软件科技有限公司 | Submerged arc furnace short-net regulation and control data processing method |
| CN116647586A (en) * | 2023-07-24 | 2023-08-25 | 山东工泵电机有限公司 | Method for realizing remote control of water and fertilizer integrated intelligent pump house by cloud computing |
| CN117303510A (en) * | 2023-09-07 | 2023-12-29 | 兰州理工大学 | Self-adaptive control method for waveform characteristic analysis of wastewater treatment power supply |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111898669B (en) | 2023-07-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111898669B (en) | Abnormal event early warning system of direct-current submerged arc furnace based on machine learning | |
| CN111596604B (en) | Intelligent fault diagnosis and self-healing control system and method for engineering equipment based on digital twinning | |
| CN107807860B (en) | Power failure analysis method and system based on matrix decomposition | |
| CN105932774A (en) | Device state early warning method in smart transformer substation based on ICA algorithm | |
| CN118213936A (en) | Voltage monitoring and voltage protecting system and method | |
| CN116614177B (en) | Optical fiber state multidimensional parameter monitoring system | |
| CN114386537A (en) | Lithium battery fault diagnosis method and device based on Catboost and electronic equipment | |
| CN116961215A (en) | Rapid fault response processing method for power system | |
| CN118094232B (en) | Dynamic monitoring production energy consumption control system and method | |
| CN106296440A (en) | Based on transformer station's warning information analysis and decision system integrated for ANN and ES and method | |
| CN117674128A (en) | Automatic fault removal method based on power dispatching system | |
| CN116824517A (en) | Substation operation and maintenance safety control system based on visualization | |
| CN117708550A (en) | Automatic data analysis and model construction method for electric power big data | |
| CN118409157A (en) | High-voltage leakage monitoring and fault positioning system for power network in coal mine area | |
| CN118193954A (en) | Power distribution network abnormal data detection method and system based on edge calculation | |
| CN116704729A (en) | Industrial kiln early warning system and method based on big data analysis | |
| CN111046633A (en) | LSTM-based power grid primary frequency modulation data prediction method and device | |
| CN113536607B (en) | Transformer substation signal transmission system evaluation method and system | |
| Liwen et al. | Fault Diagnosis of Distribution Network Based on Artificial Intelligence | |
| Lin et al. | Mechanical fault prediction based on principal component analysis | |
| Chen et al. | A transformer vibration amplitude prediction method via fusion of multi-signals | |
| CN117034157B (en) | Hydropower equipment fault identification method and system combining multimodal operation data | |
| Bai et al. | Abnormal Detection Scheme of Substation Equipment based on Intelligent Fusion Terminal | |
| CN118552186B (en) | Substation operation and maintenance management system and method based on digital twinning | |
| CN116168518A (en) | Lime kiln early warning system and method based on big data analysis |
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 |