CN106959662A - A kind of electric melting magnesium furnace unusual service condition identification and control method - Google Patents
A kind of electric melting magnesium furnace unusual service condition identification and control method Download PDFInfo
- Publication number
- CN106959662A CN106959662A CN201710324622.3A CN201710324622A CN106959662A CN 106959662 A CN106959662 A CN 106959662A CN 201710324622 A CN201710324622 A CN 201710324622A CN 106959662 A CN106959662 A CN 106959662A
- Authority
- CN
- China
- Prior art keywords
- electric melting
- melting magnesium
- operating mode
- magnesium furnace
- identification result
- 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
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
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/04—Programme control other than numerical control, i.e. in sequence controllers or logic controllers
- G05B19/05—Programmable logic controllers, e.g. simulating logic interconnections of signals according to ladder diagrams or function charts
- G05B19/058—Safety, monitoring
-
- 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
- G05B2219/10—Plc systems
- G05B2219/14—Plc safety
- G05B2219/14006—Safety, monitoring in general
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Waste-Gas Treatment And Other Accessory Devices For Furnaces (AREA)
Abstract
The present invention provides a kind of electric melting magnesium furnace unusual service condition identification and control method.Methods described includes:Obtain the online data in the predetermined period in electric melting magnesium furnace operating mode;Using similarity mode strategy check in case library whether the case information matched with online data;If in the presence of, the case information according to matching provides the identification result of current online data, using identification result as current electric melting magnesium furnace operating mode anomalous identification result;Wherein, case library is the case information for the various unusual service conditions set up previously according to the historical data of electric melting magnesium furnace operating mode;If the case information of matching is not present in case library, the online data is analyzed using Bayesian Network Inference model, acquisition identification result, using identification result as current electric melting magnesium furnace operating mode anomalous identification result.The above method reduces energy consumption for the comprehensive utilization ratio of raising mineral resources, reduces environmental pollution, promotes safety in production, there is great meaning.
Description
Technical field
The invention belongs to computer technology, more particularly to a kind of electric melting magnesium furnace unusual service condition recognition methods and security control side
Method.
Background technology
Electric-melting magnesium is a kind of important refractory material, and it has advantages below:Strong insulating properties, high-melting-point, oxidation resistance
Strong and close structure, therefore be widely applied in a variety of industrial productions.The raw material of electric-melting magnesium are magnesite, most
In the case of number, raw material grade is low, and mineral composition complexity is various.
The smelting equipment that electric-melting magnesium is used is three-phase electric melting magnesium furnace, and the current control system of electric melting magnesium furnace tracks different operating modes
Under current setting value complete fusion process.In actual applications, when the granularity of raw material, composition, fusing point etc. change, electricity
The distance between pole bottom and molten bath can produce fluctuation.This unknown fluctuation will cause the fluctuation of electric current, and electric current before is set
Definite value will no longer be suitable for the operating mode of change, can now produce higher energy ezpenditure, and systematic function declines even with safety
Threaten.
By taking abnormal gas exhaust operating mode as an example, when the granularity of raw material changes, electrode action originally can not make to melt completely
The carbon dioxide discharge produced during refining, that is, the current setting value for being vented operating mode is no longer applicable, the pressure in electric melting magnesium furnace
Good general raises, and when pressure is raised to a certain extent, high temperature melting slurry will discharge out of the furnace with gas, and the high temperature melting slurry of splashing will
Badly damaged is caused to equipment, the personal safety to operating personnel causes great threat.Meanwhile, during generation of splashing,
Substantial amounts of energy loss improves single ton of energy consumption of electric melting magnesium furnace.Due to the close coupling between extraneous strong interference and variable,
It is nonlinear to exist, understand dependency relation between variable, and the model obtained under unusual service condition is very difficult.
At present, site operation personnel passes through the current information observed, image information and acoustic information and the experience of oneself
To recognize unusual service condition, according to the unusual service condition of identification, corresponding Adjusted Option is provided, it is most of real by manual mode
Apply, automatization level is relatively low.Manual identified and operation are sensitive to human error, and operating personnel's energy when handling multi-source information
Power is limited, often ignores the influence of some variables.Within the limited time, under huge stress, operator
Member is difficult often to make effective decision-making.The artificial method of adjustment of operating personnel places one's entire reliance upon the respective experience of operator,
It is difficult to ensure that the promptness and accuracy of decision-making.If any misoperation, it is negligent of detection or because irresistible natural cause causes
Equipment fault and cause the accidents such as production disruption, it will to production bring huge waste and loss.
The content of the invention
For existing technical problem, the present invention provides a kind of electric melting magnesium furnace unusual service condition identification and control method,
The above method is applied to improve the comprehensive utilization ratio of mineral resources in electric melting magnesium furnace operating mode, reduces energy consumption.
The present invention provides a kind of electric melting magnesium furnace unusual service condition identification and control method, including:
The online data in predetermined period in S1, acquisition electric melting magnesium furnace operating mode;
S2, checked in case library with the presence or absence of the case information that is matched with online data using similarity mode strategy;
If S3, in the presence of, the case information according to matching provides the identification result of current online data, using identification result as
The anomalous identification result of current electric melting magnesium furnace operating mode;
Wherein, case library is the case letter for the various unusual service conditions set up previously according to the historical data of electric melting magnesium furnace operating mode
Breath.
Alternatively, methods described includes:If the case information of matching is not present in S3a, case library, using Bayesian network
Network inference pattern is analyzed the online data, is obtained identification result, is regard identification result as current electric melting magnesium furnace operating mode
Anomalous identification result.
Alternatively, before step S2, methods described also includes:
S2a, the historical data according to electric melting magnesium furnace operating mode in preset time period, set up case library;
And according to the historical data and priori of electric melting magnesium furnace operating mode in preset time period, set up Bayesian network and push away
Manage model.
Alternatively, step S2a includes:
Abnormal information occurs for electric melting magnesium furnace operating mode in collected offline preset time period;
According to the information of collection, the relation between correlated variables in the feature and unusual service condition of unusual service condition is determined, is obtained
Historical data, case library and Bayesian Network Inference model are set up according to historical data;
Or,
Abnormal information occurs for electric melting magnesium furnace operating mode in collected offline preset time period;
According to the information of collection, the relation between correlated variables in the feature and unusual service condition of unusual service condition is determined, is obtained
Historical data, carries out filtration treatment to historical data, case library and Bayesian network is set up according to the historical data after filtration treatment
Network inference pattern.
Alternatively, the S2a includes:
Abnormal information occurs for electric melting magnesium furnace operating mode in collected offline preset time period;
According to the relation between correlated variables in priori and unusual service condition, the node and knot of Bayesian network are determined
Structure;
According to the historical data of collection and the structure of determination, the parameter of Bayesian network is obtained, so as to set up Bayes
Network reasoning model.
Alternatively, the step S1 also includes:
Filtration treatment is carried out to the online data of acquisition, to obtain the online data after filtration treatment;
Correspondingly, the online data after filtration treatment is used to be matched with the case information in case library in step S2.
Alternatively, methods described also includes:
S4, according to anomalous identification result, determine the remaining life time of equipment;
According to the relation between the remaining life time and adjustment amount pre-established, adjustment amount is calculated, security control is obtained
Decision information, so that control system adapts to adjust according to safety control strategy information;
Wherein, remaining life time is electric melting magnesium furnace operating mode equipment from state when obtaining online data to completely can not be just
The time often worked;
The adjustment amount is current setting value;What the relation between remaining life time and adjustment amount determined for offline mode
Relation.
Alternatively, the step S2 includes:
The similarity between the variable of each case in online data and case library is obtained, judges whether the similarity is more than
Predetermined threshold value, if, it is determined that online data case corresponding with the maximum similarity more than threshold value is matched.
Alternatively, methods described also includes:
If it is real that S5, the anomalous identification result obtained by Bayesian Network Inference model are defined as current electric melting magnesium furnace operating mode
The anomalous identification result, then be stored in the case library by the result that border occurs as case information.
Alternatively, methods described also includes:
After control system adapts to adjustment according to safety control strategy information, repeat to obtain next in electric melting magnesium furnace operating mode
Online data in predetermined period, according to the deterministic process of the online data, to determine the exception of the online data in step S1
Whether operating mode excludes, if so, then terminating, otherwise, repetition obtains the security control decision-making of the online data in next predetermined period
The process of information.
The electric melting magnesium furnace unusual service condition identification of the present invention and control method, effectively can go out according to electric melting magnesium furnace production process
Existing unusual service condition, by searching whether there is matching case in case library, in the case information in the absence of matching, is used
Bayesian network model carries out unusual service condition identification, and then according to anomalous identification result, formulate security control decision information and
The decision information is applied in control system, comprehensive utilization ratio, the reduction energy consumption of electric melting magnesium furnace operating mode Mineral Resource is improved.
Brief description of the drawings
The identification of electric melting magnesium furnace unusual service condition and the schematic flow sheet of control method that Figure 1A provides for one embodiment of the invention;
Unusual service condition identification and the process schematic of control method that Figure 1B provides for one embodiment of the invention;
Fig. 2 shows for the life cycle that control system unusual service condition in the embodiment of the present invention is changed into the order of severity from slight extent
It is intended to;
Fig. 3 is the process schematic of current electric melting magnesium furnace technique;
Fig. 4 is the reasoning by cases process schematic that proposes in the embodiment of the present invention;
Fig. 5 is the signal for the Bayesian Network Inference set up in the embodiment of the present invention for electric melting magnesium furnace abnormal gas exhaust operating mode
Figure;
Fig. 6 is the contrast schematic diagram of abnormal gas exhaust industry and mining city result;
Fig. 7 is the current diagram of electric melting magnesium furnace normal exhaust operating mode;
Fig. 8 is the current diagram of electric melting magnesium furnace exhaust operating mode extremely;
Fig. 9 is the abnormal gas exhaust operating conditions effect diagram of method proposed by the present invention;
Figure 10 is the abnormal gas exhaust operating conditions effect diagram of conventional method.
Embodiment
In order to preferably explain the present invention, in order to understand, below in conjunction with the accompanying drawings, by embodiment, to this hair
It is bright to be described in detail.
As shown in figure 3, smelting process of electro-fused magnesia furnace mainly includes charging, heating is melted, the operating mode such as exhaust.Produced by electric arc
Raw heat melts raw material, obtain final product.The different current setting value by setting, current control system is by adjusting
The distance in whole electrode and molten bath tracks the setting value of change, adjusts the size of electric arc, to meet the energy needed for different operating modes.
In order to obtain higher economic benefit, it is necessary to keep good operating characteristics and production safety in the production of electric-melting magnesium.
The device involved by method in following examples includes:Electric melting magnesium furnace unusual service condition is recognized and safety control system
(following abbreviation control systems), host computer, PLC, scene sensing pick-up structure.Wherein scene sensing pick-up structure is examined including electric current
Survey instrument, image detection instrument, sound detection instrument etc..Various instrumentations, detector are installed in electric melting magnesium furnace process-field
The signal of collection is sent to PLC by table by Profibus-DP buses, and collection signal is sent to by PLC by Ethernet timing
The data of receiving are transmitted to the identification of electric melting magnesium furnace unusual service condition and safety control system by position machine, host computer, and the system identification is abnormal
Operating mode, and security control decision-making is formulated for excluding unusual service condition.
The functions of said apparatus are briefly described as follows:
1. scene senses pick-up part:Including instrumentations such as current detecting, image detection, sound detections, by sensor
Composition, is responsible for the collection and transmission of process data;
②PLC:It is responsible for the signal A/D of collection to change, and host computer is transmitted signals to by Ethernet.
In the present embodiment, PLC controller can be had using CPU 414-2 serial Simens 400, the controller
DP mouthfuls of connection distributed I/Os of Profibus;Ethernet communication module is equipped with for PLC, plc data is accessed for host computer;PLC's
Controller and ethernet communication module are placed in the PLC rack in central control room.
3. host computer:Local plc data is collected, the identification of electric melting magnesium furnace unusual service condition and safety control system is sent to.
I7 thinking computers are can select in host computer the present embodiment, using WINDOW XP operating systems.
4. unusual service condition is identified for the identification of electric melting magnesium furnace unusual service condition and safety control system, and formulates security control
Decision-making is used to exclude unusual service condition.
Electric melting magnesium furnace unusual service condition is recognized and safety control system is on i7 thinking computers, is programmed using C#2008 soft
Part, the identification of electric melting magnesium furnace unusual service condition and security control algorithm use Matlab2010a programming softwares;
It is to use C#2008 programming softwares that PLC transmits software with the signal of exception control system.
Instrumentation is installed in electric melting magnesium furnace process-field, instrumentation passes the signal of collection by Profibus-DP
It is sent in PLC, PLC timings send collection signal to host computer by Ethernet, and the data of reception are transmitted to electric smelting by host computer
Magnesium stove unusual service condition is recognized and safety control system, the system identification unusual service condition, and formulates security control decision-making for excluding
It is abnormal.
With reference to shown in Figure 1A and Figure 1B, the electric melting magnesium furnace unusual service condition identification of the present embodiment and control method include:
The online data in predetermined period in S1, acquisition electric melting magnesium furnace operating mode.
In actual applications, filtration treatment can be carried out to the online data of acquisition, with obtain after filtration treatment in line number
According to.
For example, removing the noise of online data using filtering technique, the online data for removing noise is obtained.
S2, checked in case library with the presence or absence of the case information that is matched with online data using similarity mode strategy.
For example, if the online data in step S1 is the online data after filtration treatment, filtration treatment can be used
Online data afterwards is matched with the case information in case library.
In addition, step S2 may particularly include:
Obtain in online data and case library between the variable of each case similarity (for example, by similarity criteria,
Determine similarity), judge whether the similarity is more than predetermined threshold value, if, it is determined that online data and the maximum more than threshold value
The corresponding case matching of similarity;If similarity is no more than predetermined threshold value, it is believed that case and online data are equal in case library
Mismatch.
Currently, Euclidean distance is a kind of mode for calculating similarity, calculate in the present embodiment online data variable with it is offline
Euclidean distance between data variable is similarity.Obtain online data and the similarity of each case in case library.
Predetermined threshold value in the present embodiment can be the threshold value determined beforehand through expertise or those skilled in the art's knowledge.
Especially, the case library in the step is the various exceptions set up previously according to the historical data of electric melting magnesium furnace operating mode
The case information of operating mode.
If S3, in the presence of, the case information according to matching provides the identification result of current online data, using identification result as
The anomalous identification result of current electric melting magnesium furnace operating mode.
Alternatively, during implementing, this method may also include following step S4:
S4, according to anomalous identification result, determine the remaining life time of equipment;According to the remaining life time pre-established
Relation between adjustment amount, calculates adjustment amount, security control decision information is obtained, so that control system is according to security control plan
Slightly information adapts to adjustment;That is, security control decision information is applied to control system, if unusual service condition is excluded, terminated.It is no
Then, repeat the above steps, obtain new unusual service condition regulation measure.
Wherein, above-mentioned remaining life time is electric melting magnesium furnace operating mode equipment from state when obtaining online data to complete
Can not normal work time;
The adjustment amount is current setting value;What the relation between remaining life time and adjustment amount determined for offline mode
Relation.
Wherein, remaining life time is the time that electric melting magnesium furnace operating mode equipment can continue to run well, i.e., equipment is surplus
Remaining life time refer to since present state to equipment completely can not normal work time.
In the present embodiment, online data is token state, is changed over time;For electric melting magnesium furnace, adjustment amount is electric current
Setting value, it is an offline amount, and it is exactly to immobilize to decide.In the present embodiment, abnormal is because setting value
It is improper, so to change setting value.
Relation between above-mentioned remaining life time and adjustment amount is just to establish offline, intensity of anomaly and residue life
Computation rule between the life time is also to establish offline;According to ' anomalous identification result ' and ' intensity of anomaly and remaining time
Relation ', calculate remaining life time;According to ' remaining life time ' and ' pass between remaining life time and adjustment amount
System ', calculate adjustment amount.
In actual applications, after control system adapts to adjustment according to safety control strategy information, repeat to obtain electric smelting
Online data in magnesium furnaceman's condition in next predetermined period, according to the deterministic process of the online data, to determine in step S1
Whether the unusual service condition of online data excludes, if so, then terminating, otherwise, repeats to obtain the process of security control decision information.
Method is by gathering the online data of current working in the present embodiment, and then checked whether in case library phase
Like case, if there are similar cases, using the similar cases as identification result, and then according to identification result, safety control is obtained
Decision information processed, so that control system adapts to adjust according to the security control decision information, the decision information that will be obtained should
For there is the control system of unusual service condition (i.e. the identification of electric melting magnesium furnace unusual service condition and safety control system), if unusual service condition is arranged
Remove, then terminate, otherwise, re-recognize unusual service condition, obtain decision scheme.
Thus the production status of electric melting magnesium furnace can be given some on the spot guidance, improves efficiency, reduce energy consumption, it is to avoid economy is damaged
Lose.
In a kind of optional implementation, as shown in Figure 1A, if the case of matching is not present in above-mentioned steps S2 in example storehouse
Example information, now, can perform following step S3a:
If the case information of matching is not present in S3a, case library, using Bayesian Network Inference model to described online
Data are analyzed, obtain identification result, using identification result as current electric melting magnesium furnace operating mode anomalous identification result;
And then can perform the steps such as above-mentioned steps S4.
That is, in step s 2, if the similarity of the case of matching is less than given threshold value, online data is drawn
It is divided into different intensity grades, and the online data after intensity grade will be divided as evidence and is input to Bayesian network unit,
Carry out in Bayesian Network Inference, the result that reasoning is obtained, the intensity of anomaly for possessing maximum a posteriori probability is the knowledge of unusual service condition
Other result.
Further, the identification result that Bayesian network is provided is stored in case library after confirmation as a new case
In, realize that case library is supplemented.
That is, during implementing, the above method may also include the step S5 not shown in following Figure 1A:
If it is real that S5, the anomalous identification result obtained by Bayesian Network Inference model are defined as current electric melting magnesium furnace operating mode
The anomalous identification result, then be stored in the case library by the result that border occurs as case information.
The unusual service condition that the present embodiment effectively can occur according to electric melting magnesium furnace production process, passes through reasoning by cases in case library
And the progress unusual service condition identification of Bayesian Network Inference model, the concept predicted by remaining life, according to anomalous identification result,
Formulate security control decision-making.
In a particular application, before foregoing step S2, the above method may also include following step S2a, such as Figure 1B
It is shown.
S2a, the historical data according to electric melting magnesium furnace operating mode in preset time period, set up case library;
And according to the historical data and priori of electric melting magnesium furnace operating mode in preset time period, set up Bayesian network and push away
Manage model.
For example the step of setting up case library may include:
Abnormal information occurs for electric melting magnesium furnace operating mode in S2a1, collected offline preset time period;
S2a2, the information according to collection, determine the relation between correlated variables in the feature and unusual service condition of unusual service condition,
Historical data is obtained, case library and Bayesian Network Inference model are set up according to historical data.
Or, in another optional implementation, the step of setting up case library may include:
Abnormal information occurs for electric melting magnesium furnace operating mode in S2a1 ', collected offline preset time period;
S2a2 ', the information according to collection, determine the pass between correlated variables in the feature and unusual service condition of unusual service condition
System, obtains historical data, carries out filtration treatment to historical data, case library and shellfish are set up according to the historical data after filtration treatment
This network reasoning model of leaf.
In the sub-step, to avoid influence of the noise to historical data, removed using filtering technique in historical data
Noise.
By the analysis of electric melting magnesium furnace unusual service condition in history, the pass between the feature and correlated variables of unusual service condition is determined
System.Based on the relation found, abnormal expression recognition rule, for example;If " _ premise, then _ conclusion ".Correlated variables is used as rule
Premise then, the degree of unusual service condition is as the conclusion of case, and finally, case library is established.
CBR in case library is used to carry out on-line data analysis in special instruction, the embodiment of the present invention.
Wherein, reasoning by cases is a kind of artificial intelligence approach, mainly uses the problem of similar cases occurred in history solve new.Mesh
Before, reasoning by cases process may include four basic processes:Retrieval, reuse, amendment and preservation.Elder generation is needed when retrieving similar cases
Similarity calculating method is determined, the most frequently used similar measurement mode is calculating Euclidean distance.
In addition, foregoing may include the step of setting up Bayesian Network Inference model:
Abnormal information occurs for electric melting magnesium furnace operating mode in the first step, collected offline preset time period;
Second step, according to the relation between correlated variables in priori and unusual service condition, determine Bayesian network
Node and structure;
3rd step, the historical data according to collection and the structure of determination, obtain the parameter of Bayesian network, so as to set up
Bayesian Network Inference model.
Especially, when building Bayesian network parameters, it is necessary to handle historical data, besides filtering, also
By correlated variables divided rank in all abnormal informations, the quantity and threshold value of divided rank are determined by expertise or trial-and-error method
It is fixed.According to the data calculating parameter for having divided grade.
Bayesian network in the present embodiment is a kind of expression of uncertain knowledge.The structure of Bayesian network can be with
The dependency relation between variable is expressed, parameter represents the degree of dependence between node variable.It is defeated as evidence when being collected into correlated variables
When entering to the Bayesian Network Inference model established, it can obtain paying close attention to the posterior probability of variable by inference mechanism.
Above-mentioned utilization historical data sets up case library, when new problem occurs, using similarity mode, is sought in case library
Similar cases are looked for, when the similarity degree of the case matched is more than given threshold value, abnormal knowledge is provided using the case retrieved
Other result;When the similarity degree of the case matched is less than given threshold value, unusual service condition identification is carried out using Bayesian network,
Provide identification result.
For security control decision part, the concept of remaining life time is introduced, system spare life time and tune is set up
Relation between whole amount.After identification result is obtained, the remaining life time of electric melting magnesium furnace unusual service condition, Jin Erli are calculated first
With the relation between remaining life time and adjustment amount, adjustment amount is obtained, to realize the security control of electric melting magnesium furnace.
The embodiment of the present invention carries out unusual service condition identification using reasoning by cases and Bayesian network, raw using the residue of equipment
There is provided security control scheme for the concept of life time.
The content for understanding the above method preferably, is carried out specifically below in conjunction with experimental result and technique/operating mode process
It is bright.
The technique of electric melting magnesium furnace is as shown in figure 3, capital equipment includes transformer, current loop, electrode lifting device, three-phase
Electrode and smelting furnace.Smelting process of electro-fused magnesia furnace mainly includes charging, and heating is melted, the operating mode such as exhaust.The heat produced by electric arc
Amount fusing raw material, obtain final product.The different current setting value by setting, current control system adjustment electrode is with melting
The distance in pond tracks the setting value of change, adjusts the size of electric arc, to meet the energy needed for different operating modes.Every 10-15
Minute, into stove, once, Operation mode cycle above is carried out filler, untill stove is filled.
Step 01:Set up case library and Bayesian Network Inference model
During electric-melting magnesium melting, substantial amounts of carbon dioxide can be produced, pressure is excessive to be caused to avoid in stove
Molten slurry splashes, it is necessary to adjust the setting value of electric current, and current control system follow current setting value adjusts the position of electrode, makes electricity
Pole is moved up and down, and makes to produce gap between electrode and raw material, carbon dioxide is smoothly discharged.But when the granularity of raw material becomes
During change, if current control system tracks original setting value, the gap between electrode and raw material can not make carbon dioxide
Pressure is excessive in smoothly discharge, stove, and molten slurry will splash with gas, produces abnormal gas exhaust operating mode.
When unusual service condition identification is exhausted, operating personnel can pay close attention to the information of three aspects:Current information, image letter
Breath and acoustic information.When abnormal gas exhaust operating mode occurs, current changing rate and current track error can change;Electrode and molten
Electric arc between pond is main sound-source signal, and when the state in stove changes, the amplitude and frequency of arcing sounds can be sent out
Changing;When splashing, operating personnel will be observed that high temperature melting slurry is sprayed outside stove, therefore image information can be used as auxiliary
Variable recognizes unusual service condition.
Abnormal gas exhaust operating mode is divided into three kinds of degree:Slightly, moderate and severe.Based on expertise, in misevolution
Different phase have different information and play a major role.In slight abnormal gas exhaust, acoustic information plays a major role;In moderate
When abnormal, acoustic information and current information play a major role;When severe is abnormal, current information and image information play main make
With.
Abnormal voice signal is divided into two kinds of degree:Slight and severe.Analyzed, splashed by Wigner-Ville distribution
Characteristic frequency can be extracted, be 200Hz.When abnormal gas exhaust operating mode is slight and moderate, the amplitude of voice signal can be carried
It is high;When abnormal gas exhaust operating mode is severe, due to the release of energy, the amplitude of voice signal can be reduced.Therefore for sound letter
Number, select following characteristic variable:The short-time energy of splashing characteristic frequency and the amplitude of splashing characteristic frequency.
Abnormal current signal is divided into three kinds of degree:Slightly, moderate and severe.Select current track error and electric current
Rate of change is its principal character variable.
Abnormal picture signal is divided into two kinds of degree:Slight and severe.When abnormal gas exhaust operating mode is serious, fire door model
Brightness of image in enclosing can be improved, and in image procossing, the change of brightness is embodied with gray scale.In three primary colors of image
In component, red component accounts for main function.Therefore for picture signal, following characteristic variable is selected:The change of average gray,
The change of average gray short-time energy, the change of gray variance, the change of the abundance of gray scale and red component, wherein gray scale is rich
Degree refers to the ratio more than normal picture average gray.All correlated characteristics are summarized in table 1 in abnormal gas exhaust operating mode.
The abnormal gas exhaust feature of table 1. and dependency relation
To avoid influence of the noise to historical data, the noise in historical data is removed using filtering technique.Pass through exhaust
The feature of the unusual service condition of analysis and the determination of unusual service condition, sets up case library, and form is shown in Table 2.Before feature A-I is as rule
Carry, the degree of abnormal gas exhaust operating mode as case result.
The structure of the case library of table 2.
To set up Bayesian Network Inference model, it is necessary to by all correlated variables divided ranks, the quantity of divided rank and
Threshold value is determined by expertise or trial-and-error method.By the analysis of expertise and abnormal gas exhaust operating mode, it can determine that exhaust is different
The node and structure of normal Bayesian network.Using data acquisition and pretreatment unit, the condition of Bayesian network can be obtained
The parameter of probability tables, i.e. Bayesian network, finally, Bayesian network is established, and sees Fig. 5.
Step 2:Online unusual service condition identification
Detailed process is as follows:
1) after online data is pretreated, similarity mode is carried out in case library.By similarity criteria, calculate similar
Degree, finds most like case.If the similarity of the case of matching is more than given threshold value, abnormal work is carried out using the case
Condition is recognized.Pass through the size of expertise threshold value.
2) if the similarity of the case of matching is respectively less than given threshold value, online data is divided into different degree etc.
Level, the online data for dividing intensity grade is input to Bayesian Network Inference model as evidence, carries out Bayesian Network Inference,
In the result that reasoning is obtained, the intensity of anomaly for possessing maximum a posteriori probability is the recognition result of unusual service condition.
3) after the anomalous identification result of new problem is obtained, serviceability the problem of new needs to be detected to judge it
Whether can be stored in as new case in case library.Similarly, the identification result obtained by Bayesian Network Inference model
It is also required to just be stored in case library as new case by checking.The detailed process of reasoning by cases is shown in Fig. 4.
The similarity that the case of matching is mainly explained below is less than the situation of given threshold value.By analyzing actual situation,
The event that exhaust operating mode may occur is summarized in table 3.
The event that the exhaust of table 3. operating mode may occur
Each event in table 3 is comprising 9 variables, and variables A-H is divided into three degree, respectively with digital 1-3 tables
Show, implication is respectively normal, slight abnormality and severely subnormal.Variable I is divided into four kinds of degree, is represented respectively with numeral 1-4,
Implication is respectively normal, slight abnormality, moderate exception and severely subnormal.By taking event 10 as an example, feature A-B state be it is normal,
Feature C-I state is severely subnormal.The implication of other events can be obtained with similar mode.
Event in table 3 will obtain anomalous identification result, as shown in table 4 as evidence by the reasoning of Bayesian network.
Table 4. is directed to the recognition result of the event in table 3
In table 4, identification result 1-4 represents 4 states of exhaust operating mode:Normally, mile abnormality, moderate exception and severe are different
Often, by taking event 10 as an example, its identification result is severely subnormal.The implication of the identification result of other events can use similar mode
Obtain.Event in table 3 is sorted by abnormal degree, understands that its identification result is to meet reality for the identification result of table 4
Border unusual service condition identification experience.
Method to better illustrate the embodiment of the present invention, the scheme of proposition is carried out with traditional current information that is used only
Unusual service condition is known method for distinguishing and contrasted, as a result as shown in Figure 6.In fig. 6 it can be seen that when the degree of abnormal gas exhaust operating mode
During for slight and moderate, the scheme of proposition results in identification result, and traditional method only when intensity of anomaly is serious
Identification result can be obtained.
Step 3:Formulate security control decision information.
For electric melting magnesium furnace abnormal gas exhaust, its exception control scheme is in exhaust operating mode current setting value y originallyj(t)(j
=1,2,3) on the basis of, offset Δ y is providedj(t) current control system, is allowed to track new current setting value yj' (t)=yj
(t)+Δyj(t) (j=1,2,3), wherein j represents three-phase electrode, unusual service condition is gradually recovered.The design of current offset values with
Abnormal degree is relevant with the remaining life time of system, when system remaining life time more in short-term, the abnormal order of severity
Higher, the offset of electric current is bigger.So, pass through the recognition result of unusual service condition, the remaining life time of computing system.Pass through
The remaining life time and the relation of adjustment amount pre-established, calculates the size of adjustment amount.When the abnormal identification result of system comes
When coming from reasoning by cases, remaining life time is calculated using following formula, the evolutionary process of intensity of anomaly is shown in Fig. 2.
In formula, R (t) represents remaining life time.τiRepresent the duration of i-th of abnormality.ki(t) it is that state is held
The coefficient of continuous time, to express duration of the moment t in i-th of abnormality.I (t) represents the abnormal shape in t
State, IiRepresent state i lower limit.
When the abnormal identification result of system derives from Bayesian Network Inference model, following formula (3)-(5) are used
Calculate remaining life time
(3) in formula, R (t) represent institute the weight of stateful remaining life time with.P (X=i) is represented after i-th of state
Test probability.The remaining time r of each statei(t) calculated with formula (4).For the coefficient k of different conditionsi(t) calculated with formula (5).
, it is necessary to set up the relation between remaining life time and adjustment amount after remaining life time is obtained.Will row
Adjustment amount when gas intensity of anomaly is slight, moderate and severe being set to roughly WithWherein j is represented
Three-phase electrode.The following point of selection carrys out fit correlation Δ y=f (R (t)),
(τ1+τ2+τ3, 0) andWhereinThe maximum adjustment amount allowed for technique.These points are connected by straight line, constitutes and divides
The form of section function.
So, by anomalous identification result, remaining life time is calculated.By between remaining life time and adjustment amount
Relation, calculates the size of adjustment amount, that is, provides security control decision-making.
Fig. 7 and Fig. 8 sets forth normal exhaust operating mode and abnormal exhaust operating mode.It can be seen from Fig. 8 due to
The setting value of electric current is not adjusted, carbon dioxide causes the fluctuation of electric current increasing without smoothly being discharged from stove.
Security control scheme is the offset for providing current setting value, current control system is tracked new current setting value different to exclude
Normal operating mode.
Step 4:Implementation decision excludes unusual service condition
The unusual service condition produced for Fig. 8, the security control decision-making that the present invention is formulated is applied to control system, and effect is such as
Shown in Fig. 9.It can be seen in figure 9 that in the 7th sampled point, current fluctuation exceeds normal range (NR), and this phenomenon continues to hold
It is continuous.After security control decision implement, in the 13rd sampled point, current fluctuation declines, about in the 19th sampled point, abnormal work
Condition is excluded.In order to embody the superiority of the present invention, of the invention and traditional is used only the side that current information carries out security decision
Method is compared, and the control effect of conventional method is as shown in Figure 10.From fig. 10 it can be seen that in the 7th sampled point current fluctuation
Beyond normal range (NR), and this phenomenon is continued for, until the 20th sampled point, when intensity of anomaly is changed into serious, tradition side
Method just provides security decision, and system gradually recovers normal afterwards.
By above example, unusual service condition can be recognized by indicating the method for the embodiment of the present invention, and according to identification
Unusual service condition formulate effective safety measure, measure can effectively be such that unusual service condition reverts to normally after implementing, and with
Tradition be used only current information carry out unusual service condition identification compared with the method for security control, method proposed by the present invention more added with
Imitate and with better performance.Further, the above method reduces energy consumption, subtracted for the comprehensive utilization ratio of raising mineral resources
Few environmental pollution, promotes safety in production, there is great meaning.
The know-why of the present invention is described above in association with specific embodiment, these descriptions are intended merely to explain the present invention's
Principle, it is impossible to be construed to limiting the scope of the invention in any way.Based on explaining herein, those skilled in the art
Would not require any inventive effort can associate other embodiments of the present invention, and these modes fall within this hair
Within bright protection domain.
Claims (10)
1. a kind of electric melting magnesium furnace unusual service condition identification and control method, it is characterised in that including:
The online data in predetermined period in S1, acquisition electric melting magnesium furnace operating mode;
S2, checked in case library with the presence or absence of the case information that is matched with online data using similarity mode strategy;
If S3, in the presence of the case information according to matching provides the identification result of current online data, using identification result as current
The anomalous identification result of electric melting magnesium furnace operating mode;
Wherein, case library is the case information for the various unusual service conditions set up previously according to the historical data of electric melting magnesium furnace operating mode.
2. according to the method described in claim 1, it is characterised in that methods described includes:
If the case information of matching is not present in S3a, case library, using Bayesian Network Inference model to the online data
Analyzed, obtain identification result, using identification result as current electric melting magnesium furnace operating mode anomalous identification result.
3. according to the method described in claim 1, it is characterised in that before step S2, methods described also includes:
S2a, the historical data according to electric melting magnesium furnace operating mode in preset time period, set up case library;
And according to the historical data and priori of electric melting magnesium furnace operating mode in preset time period, set up Bayesian Network Inference mould
Type.
4. method according to claim 3, it is characterised in that step S2a includes:
Abnormal information occurs for electric melting magnesium furnace operating mode in collected offline preset time period;
According to the information of collection, the relation between correlated variables in the feature and unusual service condition of unusual service condition is determined, history is obtained
Data, case library and Bayesian Network Inference model are set up according to historical data;
Or,
Abnormal information occurs for electric melting magnesium furnace operating mode in collected offline preset time period;
According to the information of collection, the relation between correlated variables in the feature and unusual service condition of unusual service condition is determined, history is obtained
Data, carry out filtration treatment to historical data, set up case library according to the historical data after filtration treatment and Bayesian network is pushed away
Manage model.
5. method according to claim 3, it is characterised in that the S2a includes:
Abnormal information occurs for electric melting magnesium furnace operating mode in collected offline preset time period;
According to the relation between correlated variables in priori and unusual service condition, the node and structure of Bayesian network are determined;
According to the historical data of collection and the structure of determination, the parameter of Bayesian network is obtained, so as to set up Bayesian network
Inference pattern.
6. according to the method described in claim 1, it is characterised in that the step S1 also includes:
Filtration treatment is carried out to the online data of acquisition, to obtain the online data after filtration treatment;
Correspondingly, the online data after filtration treatment is used to be matched with the case information in case library in step S2.
7. according to any described method of claim 2 to 6, it is characterised in that methods described also includes:
S4, according to anomalous identification result, determine the remaining life time of equipment;
According to the relation between the remaining life time and adjustment amount pre-established, adjustment amount is calculated, security control decision-making is obtained
Information, so that control system adapts to adjust according to safety control strategy information;
Wherein, remaining life time is electric melting magnesium furnace operating mode equipment from state when obtaining online data to completely can not normal work
The time of work;
The adjustment amount is current setting value;Relation between remaining life time and adjustment amount is the pass that offline mode is determined
System.
8. method according to claim 7, it is characterised in that the step S2 includes:
The similarity between the variable of each case in online data and case library is obtained, judges whether the similarity is more than default
Threshold value, if, it is determined that online data case corresponding with the maximum similarity more than threshold value is matched.
9. method according to claim 2, it is characterised in that methods described also includes:
If S5, the anomalous identification result obtained by Bayesian Network Inference model are defined as the actual hair of current electric melting magnesium furnace operating mode
The anomalous identification result, then be stored in the case library by raw result as case information.
10. method according to claim 7, it is characterised in that the step S4 includes:
If anomalous identification result is matched by case library lookup, remaining life time R is calculated using following formula one
(t);
Formula one:
Wherein,R (t) represents remaining life time, τiRepresent the duration of i-th of abnormality, ki(t)
It is the coefficient of state duration, to express duration of the moment t in i-th of abnormality, I (t) is represented in t
Abnormality, IiRepresent state i lower limit;
Or,
If anomalous identification result is determined by Bayesian Network Inference model, remaining life is calculated using following formula two
Time;
Formula two:
Wherein,
R (t) represent institute the weight of stateful remaining life time with.P (X=i) represents the posterior probability of i-th of state, ri(t)
Represent the remaining time of each state, ki(t) coefficient for different conditions is represented, I (t) represents the abnormality in t,
IiRepresent state i lower limit.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710324622.3A CN106959662B (en) | 2017-05-10 | 2017-05-10 | A kind of identification of electric melting magnesium furnace unusual service condition and control method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710324622.3A CN106959662B (en) | 2017-05-10 | 2017-05-10 | A kind of identification of electric melting magnesium furnace unusual service condition and control method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106959662A true CN106959662A (en) | 2017-07-18 |
CN106959662B CN106959662B (en) | 2019-03-08 |
Family
ID=59481878
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710324622.3A Active CN106959662B (en) | 2017-05-10 | 2017-05-10 | A kind of identification of electric melting magnesium furnace unusual service condition and control method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106959662B (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108960256A (en) * | 2018-06-28 | 2018-12-07 | 东软集团股份有限公司 | A kind of determination method, device and equipment of components damage degree |
CN110059359A (en) * | 2019-03-21 | 2019-07-26 | 江苏东方国信工业互联网有限公司 | A kind of system and method for the control furnace body technique based on big data analysis |
WO2020119610A1 (en) * | 2018-12-11 | 2020-06-18 | 华为技术有限公司 | Fault root cause identification method and apparatus and device |
CN111598137A (en) * | 2020-04-24 | 2020-08-28 | 北京金山云网络技术有限公司 | Method and device for providing reasoning service and electronic equipment |
CN113064390A (en) * | 2021-03-17 | 2021-07-02 | 国网辽宁省电力有限公司辽阳供电公司 | Case reasoning-based active warning method for pollutant emission of cement production enterprise |
CN113237332A (en) * | 2021-04-14 | 2021-08-10 | 东北大学 | Method for identifying working condition of electro-fused magnesia furnace |
CN113705601A (en) * | 2021-05-21 | 2021-11-26 | 国网辽宁省电力有限公司鞍山供电公司 | Method for identifying running state of electric smelting magnesium furnace based on random forest |
CN115354144A (en) * | 2022-08-29 | 2022-11-18 | 湖南华菱涟源钢铁有限公司 | Roller type quenching process operation parameter setting method based on case matching |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102637019A (en) * | 2011-02-10 | 2012-08-15 | 武汉科技大学 | Intelligent integrated fault diagnosis method and device in industrial production process |
CN102765643A (en) * | 2012-05-31 | 2012-11-07 | 天津大学 | Elevator fault diagnosis and early-warning method based on data drive |
CN104298225A (en) * | 2014-09-25 | 2015-01-21 | 中国石油化工股份有限公司 | Modeling and graphical displaying method for causal relationship reasoning model of unusual working conditions of chemical process |
-
2017
- 2017-05-10 CN CN201710324622.3A patent/CN106959662B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102637019A (en) * | 2011-02-10 | 2012-08-15 | 武汉科技大学 | Intelligent integrated fault diagnosis method and device in industrial production process |
CN102765643A (en) * | 2012-05-31 | 2012-11-07 | 天津大学 | Elevator fault diagnosis and early-warning method based on data drive |
CN104298225A (en) * | 2014-09-25 | 2015-01-21 | 中国石油化工股份有限公司 | Modeling and graphical displaying method for causal relationship reasoning model of unusual working conditions of chemical process |
Non-Patent Citations (1)
Title |
---|
吴志伟: "《嵌入式电熔镁炉智能控制系统研究》", 28 February 2017 * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108960256A (en) * | 2018-06-28 | 2018-12-07 | 东软集团股份有限公司 | A kind of determination method, device and equipment of components damage degree |
WO2020119610A1 (en) * | 2018-12-11 | 2020-06-18 | 华为技术有限公司 | Fault root cause identification method and apparatus and device |
CN111385106A (en) * | 2018-12-11 | 2020-07-07 | 华为技术有限公司 | Method, device and equipment for identifying fault root cause |
US11956118B2 (en) | 2018-12-11 | 2024-04-09 | Huawei Technologies Co., Ltd. | Fault root cause identification method, apparatus, and device |
CN110059359A (en) * | 2019-03-21 | 2019-07-26 | 江苏东方国信工业互联网有限公司 | A kind of system and method for the control furnace body technique based on big data analysis |
CN111598137A (en) * | 2020-04-24 | 2020-08-28 | 北京金山云网络技术有限公司 | Method and device for providing reasoning service and electronic equipment |
CN113064390A (en) * | 2021-03-17 | 2021-07-02 | 国网辽宁省电力有限公司辽阳供电公司 | Case reasoning-based active warning method for pollutant emission of cement production enterprise |
CN113064390B (en) * | 2021-03-17 | 2022-03-01 | 国网辽宁省电力有限公司辽阳供电公司 | Case reasoning-based active warning method for pollutant emission of cement production enterprise |
CN113237332A (en) * | 2021-04-14 | 2021-08-10 | 东北大学 | Method for identifying working condition of electro-fused magnesia furnace |
CN113705601A (en) * | 2021-05-21 | 2021-11-26 | 国网辽宁省电力有限公司鞍山供电公司 | Method for identifying running state of electric smelting magnesium furnace based on random forest |
CN115354144A (en) * | 2022-08-29 | 2022-11-18 | 湖南华菱涟源钢铁有限公司 | Roller type quenching process operation parameter setting method based on case matching |
CN115354144B (en) * | 2022-08-29 | 2024-01-30 | 湖南华菱涟源钢铁有限公司 | Roller quenching process operation parameter setting method based on case matching |
Also Published As
Publication number | Publication date |
---|---|
CN106959662B (en) | 2019-03-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106959662B (en) | A kind of identification of electric melting magnesium furnace unusual service condition and control method | |
Wu et al. | Data-driven abnormal condition identification and self-healing control system for fused magnesium furnace | |
CN111949700B (en) | Intelligent safety guarantee real-time optimization method and system for petrochemical device | |
CN101251942B (en) | Underground space fire intelligent detection early alarming and forecasting method and apparatus | |
Li et al. | Abnormal condition identification and safe control scheme for the electro-fused magnesia smelting process | |
CN111858231A (en) | Single index abnormality detection method based on operation and maintenance monitoring | |
CN111651931A (en) | Blast furnace fault diagnosis rule derivation method based on deep neural network | |
Naso et al. | A fuzzy-logic based optical sensor for online weld defect-detection | |
CN109685283A (en) | A kind of mineral hot furnace operating condition prediction technique based on confidence rule base reasoning | |
CN108803528A (en) | Process industry system prediction model based on multivariate correlation and time lag | |
CN105171146A (en) | Self-adaptive control system and method for electric spark machining | |
CN106600076A (en) | Regenerative thermal oxidizer (RTO) waste gas treatment device monitoring data analysis and early warning method | |
CN115358626A (en) | Safety inspection early warning system for mineral processing production equipment | |
CN114355185A (en) | Operation management and control system for intelligent switch | |
Saravanan | Fuzzy controller design of lighting control system by using VI package | |
CN115423134B (en) | Heavy film inflation film manufacturing machine operation detecting system based on artificial intelligence | |
CN108932531A (en) | Equipment state on-line checking algorithm based on model feature value | |
CN108832614A (en) | Power distribution equipment fault processing method and device based on fuzzy control | |
CN112732817A (en) | Big data processing method of Internet of things system | |
CN112229206A (en) | Method for controlling coal feeding of kiln head of rotary cement kiln | |
CN116306272A (en) | Converter heat loss rate prediction method based on big data | |
Koprinkova-Hristova | Reinforcement learning for predictive maintenance of industrial plants | |
CN105624374B (en) | A kind of air-cushion type quenching system and automatic control system | |
CN114217592A (en) | Control system and method for copper top-blowing production process | |
CN207380871U (en) | A kind of fired power generating unit Realtime Alerts monitoring system based on system subdivision |
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 |