CN109858709A - A kind of method, apparatus and equipment optimizing coke production - Google Patents
A kind of method, apparatus and equipment optimizing coke production Download PDFInfo
- Publication number
- CN109858709A CN109858709A CN201910156054.XA CN201910156054A CN109858709A CN 109858709 A CN109858709 A CN 109858709A CN 201910156054 A CN201910156054 A CN 201910156054A CN 109858709 A CN109858709 A CN 109858709A
- Authority
- CN
- China
- Prior art keywords
- data
- coke
- coke quality
- optimized
- quality
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Coke Industry (AREA)
Abstract
The invention discloses a kind of method, apparatus and equipment for optimizing coke production.This method comprises: constructing Coke Quality Prediction Models according to historical data, and Coke Quality prediction model is trained, and obtains the first coke quality data, wherein historical data includes coke production technique supplemental characteristic, mixed coal data and coke quality data；According to the first coke quality data and business it needs to be determined that achievement data to be optimized in the first coke quality data；Optimizing index data are treated using optimization algorithm to optimize, and obtain the recommended value of achievement data to be optimized according to optimum results；Coke production technique supplemental characteristic or mixed coal data are adjusted according to recommended value, to improve coke quality.The present invention has fully considered the influence of coke production technique parameter and mixed coal Coke Quality in actual production, therefore, optimizes coke production.
Description
Technical field
The present invention relates to Coking Coal Blending Technology field, in particular to a kind of method, apparatus and equipment for optimizing coke production.
Background technique
The process of coke production is mixed coal to be formed after being mixed in a certain ratio raw coal, and make mixed coal in coke oven
High-temperature retorting is carried out, coke and raw coke oven gas are obtained.Currently, the process due to coke production is various, complex process, therefore, focusing
It is very difficult that charcoal production, which is optimized to improve coke quality and reduce production cost,.
In the prior art, there are mainly two types of solutions to the problems described above, one is artificial coal blending, another kind is to be based on
Single mathematical model coal blending.Although both methods optimizes coke production to a certain extent, it has ignored practical life
In production therefore the influence of coke production technique parameter and mixed coal Coke Quality is not inconsistent with actual production.
Summary of the invention
The embodiment of the present invention is designed to provide a kind of method, apparatus and equipment for optimizing coke production, can be abundant
Consider the influence of coke production technique parameter and mixed coal Coke Quality in actual production, and it is raw to have advanced optimized coke
It produces.
In order to solve the above-mentioned technical problem, the embodiment of the invention provides the following technical solutions:
The present invention provides a kind of methods for optimizing coke production, comprising:
Coke Quality Prediction Models are constructed according to historical data, and Coke Quality prediction model is trained, and obtains the
One coke quality data, wherein historical data includes coke production technique supplemental characteristic, mixed coal data and coke quality number
According to；
According to the first coke quality data and business it needs to be determined that achievement data to be optimized in the first coke quality data；
Optimizing index data are treated using optimization algorithm to optimize, and obtain achievement data to be optimized according to optimum results
Recommended value；
Coke production technique supplemental characteristic or mixed coal data are adjusted according to recommended value, to improve coke quality.
Optionally, before constructing Coke Quality Prediction Models according to historical data, this method further include:
Obtain historical data；
Data prediction is carried out to historical data, obtains high quality historical data；
High quality historical data is screened by correlation calculations, after obtaining screening relevant to coke quality data
Coke production technique supplemental characteristic and mixed coal data.
Further, data prediction includes that Missing Data Filling processing, data format abnormality processing, data value range are different
At least one of often processing and data reprocessing.
Optionally, Coke Quality prediction model is trained, and obtains the first coke quality data, comprising:
To after screening coke production technique supplemental characteristic and mixed coal data split, obtain prediction of coke quality instruction
Practice collection；
It is predicted the sample in prediction of coke quality training set as the input Coke Quality of Coke Quality Prediction Models
Model is trained, and obtains the first coke quality data, wherein sample is that prediction of coke quality training is concentrated through screening and obtains
Characteristic.
Optionally, optimizing index data are treated using optimization algorithm to optimize, and is obtained according to optimum results to be optimized
The recommended value of achievement data, comprising:
Optimizing index data are treated using optimization algorithm to optimize, and obtain the second coke quality data；
Second coke quality data are compared with the first coke quality data, and are obtained according to comparison result to be optimized
The recommended value of achievement data.
Optionally, the second coke quality data are compared with the first coke quality data, and are obtained according to comparison result
To the recommended value of achievement data to be optimized, comprising:
Second coke quality data are compared with the first coke quality data；
If the second coke quality data are better than the first coke quality data, using the second coke quality data as to excellent
Change the recommended value of achievement data.
Optionally, achievement data to be optimized includes coke quality sulphur content, reactivity indexes, post reaction strength, crushing strength
At least one of with wear-resistant strength.
Optionally, optimization algorithm includes genetic algorithm, particle swarm algorithm, ant group algorithm, simulated annealing, crowd's search
One of algorithm and artificial bee colony algorithm.
The present invention provides a kind of devices for optimizing coke production, comprising:
Module is constructed, is configured to construct Coke Quality Prediction Models, and Coke Quality prediction model according to historical data
It is trained, obtains the first coke quality data, wherein historical data includes coke production technique supplemental characteristic, mixed coal number
According to coke quality data；
Determining module is configured to according to the first coke quality data and business it needs to be determined that in the first coke quality data
Achievement data to be optimized；
Optimization module is configured to be treated optimizing index data using optimization algorithm and optimized, and is obtained according to optimum results
To the recommended value of achievement data to be optimized；
Module is adjusted, is configured to adjust coke production technique supplemental characteristic or mixed coal data according to recommended value, to improve
Coke quality.
Optionally, the device further include:
Module is obtained, is configured to obtain historical data；
Preprocessing module is configured to carry out data prediction to historical data, obtains high quality historical data；
Screening module is configured to screen high quality historical data by correlation calculations, obtains and coke quality
Coke production technique supplemental characteristic and mixed coal data after the relevant screening of data.
Optionally, data prediction includes Missing Data Filling processing, data format abnormality processing, data value range exception
At least one of processing and data reprocessing.
Optionally, building module to after screening coke production technique supplemental characteristic and mixed coal data split, obtain
To prediction of coke quality training set, and using the sample in prediction of coke quality training set as the input of Coke Quality Prediction Models
Coke Quality prediction model is trained, and obtains the first coke quality data, wherein sample is prediction of coke quality training set
In pass through the obtained characteristic of screening.
Optionally, optimization module is treated optimizing index data using optimization algorithm and is optimized, and obtains the second coke quality
Second coke quality data are compared with the first coke quality data, and obtain finger to be optimized according to comparison result by data
Mark the recommended value of data.
Optionally, the second coke quality data are compared by optimization module with the first coke quality data, if second
Coke quality data are better than the first coke quality data, then using the second coke quality data as the suggestion of achievement data to be optimized
Value.
Optionally, achievement data to be optimized includes coke quality sulphur content, reactivity indexes, post reaction strength, crushing strength
At least one of with wear-resistant strength.
Optionally, optimization algorithm includes genetic algorithm, particle swarm algorithm, ant group algorithm, simulated annealing, crowd's search
One of algorithm and artificial bee colony algorithm.
The present invention provides a kind of equipment for optimizing coke production, comprising: processor and memory, processor are loaded and held
Instruction and data in line storage, for realizing method as described above.
Disclosure based on the above embodiment can know that the beneficial effect of the embodiment of the present invention is:
By constructing Coke Quality Prediction Models according to historical data, and Coke Quality prediction model is trained, and is obtained
To the first coke quality data, wherein historical data includes coke production technique supplemental characteristic, mixed coal data and coke quality
Data；According to the first coke quality data and business it needs to be determined that achievement data to be optimized in the first coke quality data；It adopts
Optimizing index data are treated with optimization algorithm to optimize, and obtain the recommended value of achievement data to be optimized according to optimum results；
Reality can be fully considered to improve coke quality by adjusting coke production technique supplemental characteristic or mixed coal data according to recommended value
In the production of border therefore the influence of coke production technique parameter and mixed coal Coke Quality optimizes coke production.
Detailed description of the invention
Fig. 1 is a kind of flow chart of the method for optimization coke production shown in an exemplary embodiment according to the present invention.
Fig. 2 is a kind of process of the method for optimization coke production shown in another exemplary embodiment according to the present invention
Figure.
Fig. 3 is a kind of block diagram of the device of optimization coke production shown in an exemplary embodiment according to the present invention.
Fig. 4 is a kind of block diagram of the equipment of optimization coke production shown in an exemplary embodiment according to the present invention.
Specific embodiment
Various schemes and feature of the invention are described herein with reference to attached drawing.
It should be understood that various modifications can be made to the embodiment invented herein.Therefore, description above should not regard
To limit, and only as the example of embodiment.Those skilled in the art will expect within the scope and spirit of this invention
Other modifications.
The attached drawing being included in the description and forms part of the description shows the embodiment of the present invention, and with it is upper
What face provided is used to explain the present invention substantially description and the detailed description given below to embodiment of the invention together
Principle.
It is of the invention by the description of the preferred form with reference to the accompanying drawings to the embodiment for being given as non-limiting example
These and other characteristic will become apparent.
Although being also understood that invention has been described referring to some specific examples, those skilled in the art
Member realizes many other equivalents of the invention in which can determine, they have feature as claimed in claim and therefore all
In the protection scope defined by whereby.
When read in conjunction with the accompanying drawings, in view of following detailed description, above and other aspect of the invention, feature and advantage will become
It is more readily apparent.
Specific embodiments of the present invention are described hereinafter with reference to attached drawing；It will be appreciated, however, that the embodiment invented is only
Various ways implementation can be used in example of the invention.Known and/or duplicate function and structure and be not described in detail to avoid
Unnecessary or extra details makes the present invention smudgy.Therefore, the specific structural and functionality invented herein is thin
Section is not intended to restrictions, but as just the basis of claim and representative basis be used to instructing those skilled in the art with
Substantially any appropriate detailed construction diversely uses the present invention.
This specification can be used phrase " in one embodiment ", " in another embodiment ", " in another embodiment
In " or " in other embodiments ", it can be referred to one or more of identical or different embodiment according to the present invention.
Fig. 1 is a kind of flow chart of the method for optimization coke production shown in an exemplary embodiment according to the present invention.Such as
Shown in Fig. 1, this method comprises:
110: Coke Quality Prediction Models being constructed according to historical data, and Coke Quality prediction model is trained, and is obtained
To the first coke quality data, wherein historical data includes coke production technique supplemental characteristic, mixed coal data and coke quality
Data.
In embodiments of the present invention, according to the historical data got from the database of Jiao Chang enterprise, machine learning is used
Algorithm constructs Coke Quality Prediction Models；Further, Coke Quality prediction model is trained, and obtains the first coke matter
Measure data.
Specifically, historical data is discrete data, and including coke production technique supplemental characteristic, mixed coal data and coke
Charcoal qualitative data.Here, coke production technique supplemental characteristic may include maximum coke pushing current, average coke pushing current, coke pushing system
One of number, maximum coking time, average coking time, collecting main pressure and blast-furnace gas pressure are a variety of；Mixed coal number
According to may include ash content, sulphur content, caking property, volatile matter, moisture, fineness, heap density, phosphorus content, degree of coalification and lithofacies composition in
It is one or more；Coke quality data may include the sulphur content of coke quality, phosphorus point, ash content, volatile matter, mechanical strength (i.e.
Usually said cold strength includes crushing strength M40 and wear-resistant strength M10), reactivity indexes (Coke Reactivity
Index, CRI), post reaction strength (Coke Strength after Reaction, CSR), in moisture content and breeze content
It is one or more.Machine learning algorithm can include but is not limited to deep neural network (Deep Neural Network,
DNN), random forest (Random forest), gradient boosted tree (Gradient Boosting Decison Tree, GBDT),
Support vector machines (Support Vector Machine, SVM), logistic regression, support vector machines, naive Bayesian, k nearest neighbor are calculated
Method etc.；Preferably, the present invention uses deep neural network, random forest scheduling algorithm.
It should be noted that suitable machine learning algorithm can be selected to construct coke according to different coke quality data
Charcoal quality prediction model.In addition it is also necessary to explanation, machine learning algorithm of the invention can be single machine learning and calculate
Method is also possible to the algorithm after improving according to feature and training result to original algorithm, or can also be polyalgorithm
Combination, the invention is not limited in this regard.
In addition, leading to the coke quality of different Jiao Chang enterprise productions since the production procedure of coke is various and complex process
It is different, therefore, before constructing Coke Quality Prediction Models, needs to select Jiao Chang enterprise, and from the number of the Jiao Chang enterprise
According to the historical data for obtaining its acquisition in library.In addition, correctness and validity in order to guarantee data, it is also necessary to historical data
Data prediction is carried out, to obtain the data available of high quality.Here, data prediction can include but is not limited to data cleansing
Processing, data integration processing, data regularization processing and data conversion process etc..
Further, pretreated historical data is changed into the training data of model by Feature Engineering.Here, special
Sign engineering refers to the process of the training data that pretreated data are changed into model, including feature construction, feature extraction and
Feature selecting.Specifically, feature construction refers to that artificial some features with physical significance of finding out from initial data (count
According to)；Primitive character is converted to one group of feature with obvious physical significance or statistical significance or core by feature extraction；Feature choosing
Selecting is the feature that one group of most statistical significance is selected from characteristic set, reaches dimensionality reduction.Further, the method for feature extraction can
To include but is not limited to principal component analysis (Principal Component Analysis, PCA), linear discriminant analysis
(Linear Discriminant Analysis, LDA) and singular value decomposition (Singular Value Decomposition,
SVD) etc.；The method of feature selecting can include but is not limited to filtering type (Filter) selection, packaging type (Wrapper) selection and
Embedded (Embedding) selection etc..
Further, it is also possible to derive new feature using such as One-Hot coding based on the feature filtered out.Here,
One-Hot coding is also known as an efficient coding, is mainly encoded using N bit status register to N number of state, each
State all by his independent register-bit, and only have when any one effectively.
Finally, Coke Quality prediction model is trained, the first coke quality data are obtained.Specifically, to pretreatment
Historical data afterwards is split, obtain prediction of coke quality training set Train_Set=[Sample1, Sample2 ...,
Samplen] and test set；It is concentrated through the sample Sample1 that features described above engineering obtains using prediction of coke quality training,
The input of Sample2 ..., Samplen (that is, the characteristic obtained by screening) as Coke Quality Prediction Models, it is corresponding
Output of the single coke quality data as Coke Quality Prediction Models, Coke Quality prediction model are trained.Into one
With walking, according to constructed Coke Quality Prediction Models the characteristics of, such as K- is used to roll over cross validation (K-fold Cross
The methods of) Validation optimizing is scanned for, obtains preferably model parameter data, i.e. the first coke quality data.Here,
K- folding cross validation for example can be 5 folding cross validations or 10 folding cross validations.
120: according to the first coke quality data and business it needs to be determined that index to be optimized in the first coke quality data
Data.
In embodiments of the present invention, the first coke quality data and business demand obtained according to training, which determine, to be wanted from which
A aspect STRENGTH ON COKE production optimizes.
Specifically, can from mixed coal data and coke production technique supplemental characteristic in terms of the two STRENGTH ON COKE production carry out
Optimization.If user wants to be optimized according to the production of mixed coal data STRENGTH ON COKE, optimizing index is that mixed coal index (is matched
Close coal data) a part, and specifically select which kind of index, need to be judged according to previous experience, i.e., according to different
The influence degree of mixed coal index Coke Quality selects suitable optimizing index.Similarly, if user wants according to coke
The production of technological production norm data STRENGTH ON COKE optimizes, then optimizing index is that (i.e. coke is raw for coke production technique parameter index
Produce technical parameter data) a part, and specifically select which kind of index, need to be judged according to previous experience, i.e. basis
The influence degree of different mixed coal index Coke Qualities selects suitable optimizing index.
It should be noted that index to be optimized can be judged and selected by artificial mode, modeling can also be passed through
Mode finds index to be optimized, the invention is not limited in this regard.
130: optimizing index data being treated using optimization algorithm and are optimized, and obtain index to be optimized according to optimum results
The recommended value of data.
In embodiments of the present invention, it treats optimizing index data using optimization algorithm to optimize, here, optimization algorithm can
To include frequently-used data Processing Algorithm, neural network algorithm and intelligent algorithm；Further, it is obtained according to optimum results to be optimized
The recommended value of achievement data, here, achievement data to be optimized may include strong after coke quality sulphur content, reactivity indexes, reaction
One of degree, crushing strength and wear-resistant strength are a variety of.
Specifically, frequently-used data Processing Algorithm can include but is not limited to grey correlation analysis (Grey Relational
Analysis, GRA), Partial Least Squares Regression (Partial Least Squares Regression, PLSR), time series
(Time Series, TS), Markov Chain (Markov Chain, MC), Bayes (Bayes) etc..Neural network algorithm can be with
Including but not limited to reverse transmittance nerve network (Back Propagation, BP), self-organizing feature map neural network
(Self-Organizing Feature Mapping, SOFM), great Mansfield moral (Hopfield) neural network, radial base letter
Number (Radial Basis Function, RBF) neural network etc..Intelligent algorithm can include but is not limited to genetic algorithm
(Genetic Algorithm, GA), particle swarm algorithm (Particle Swarm Optimization, PSO), ant group algorithm
(Ant Colony Optimization, ACO), simulated annealing (Simulated Annealing, SA), crowd, which search for, to be calculated
Method (Seeker Optimization Algorithm, SOA), artificial bee colony algorithm (Artificial Bee Colony
Algorithm, ABC), gravitation searching algorithm (Gravitational Search Algorithm, GSA), bacterium look for food
Algorithm (Bacteria Foraging Optimization Algorithm, BFOA), Hungary Algorithm, fish-swarm algorithm etc..It is excellent
Selection of land, the present invention use the optimization algorithms such as genetic algorithm, particle swarm algorithm.
140: coke production technique supplemental characteristic or mixed coal data being adjusted according to recommended value, to improve coke quality.
It in embodiments of the present invention, can be according to the recommended value tune after the recommended value for determining achievement data to be optimized
Whole coke production technique supplemental characteristic or mixed coal data, and then improve coke quality.
The technical solution provided according to embodiments of the present invention, by constructing Coke Quality Prediction Models according to historical data,
And Coke Quality prediction model is trained, and obtains the first coke quality data, wherein historical data includes coke production skill
Art supplemental characteristic, mixed coal data and coke quality data；According to the first coke quality data and business it needs to be determined that first is burnt
Achievement data to be optimized in charcoal qualitative data；It treats optimizing index data using optimization algorithm to optimize, and according to optimization
As a result the recommended value of achievement data to be optimized is obtained；Coke production technique supplemental characteristic or mixed coal number are adjusted according to recommended value
According to can fully consider coke production technique parameter and mixed coal Coke Quality in actual production to improve coke quality
It influences, therefore, optimizes coke production.
In another embodiment of the present invention, before constructing Coke Quality Prediction Models according to historical data, the party
Method further include: obtain historical data；Data prediction is carried out to historical data, obtains high quality historical data；Pass through correlation
High quality historical data is screened in calculating, the coke production technique parameter after obtaining screening relevant to coke quality data
Data and mixed coal data.
Specifically, since the production procedure of coke is various and complex process, lead to the coke matter of different Jiao Chang enterprise productions
Amount is different, therefore, before constructing Coke Quality Prediction Models according to historical data, it is necessary first to Jiao Chang enterprise is selected,
And the historical data of its acquisition is obtained from the database of the Jiao Chang enterprise.Here, historical data may include coke production skill
Art supplemental characteristic (for example, collecting main pressure etc.), mixed coal data (for example, mixed coal sulphur content etc.) and coke quality data (example
Such as, coke quality sulphur content etc.).Table 1 shows showing for coke production technique supplemental characteristic, mixed coal data and coke quality data
Example.
Table 1
| X1 | X2 | … | Xn | Y |
| X1_1 | X2_1 | … | Xn_1 | Y_1 |
| X1_2 | X2_2 | … | Xn_2 | Y_2 |
| X1_3 | X2_3 | … | Xn_3 | Y_3 |
| X1_4 | X2_4 | … | Xn_4 | Y_4 |
| … | … | … | … | … |
| X1_m | X2_m | … | Xn_m | Y_m |
Wherein, X1, X2 ..., Xn indicate coke production technique supplemental characteristic and mixed coal data, Y indicates coke quality number
According to.
It should be noted that coke quality data Y can be obtained directly by the meter on device (if had on device
Corresponding measurement point), it can also be calculated indirectly by the value of other measurement points, the invention is not limited in this regard.
Then, it in order to guarantee the correctness and validity of data, needs to carry out data to the historical data got to locate in advance
Reason.Here, data prediction may include at data cleansing processing, data integration processing, data regularization processing and data transformation
One of reason is a variety of.Data cleansing processing is corrected inconsistent for the noise in clearing data；Data integration processing is used for
Data are merged into a consistent data by multiple data sources to store, such as data warehouse；Data regularization processing is for by such as
Aggregation deletes redundancy feature or cluster to reduce the scale of data；Data conversion process is used for data compression to lesser area
Between, such as 0.0 to 1.0.In this embodiment, data prediction is data cleansing processing, and its object is to lack to existing part
The defects of mistake, repetition, noise, exception, is handled, to obtain the data available of high quality.Data cleansing processing may include lacking
One of mistake value filling processing, data format abnormality processing, data value range abnormality processing and data reprocessing are more
Kind.
Further, the high quality historical data obtained by data prediction is changed by coke matter by Feature Engineering
Measure the training data of prediction model.Here, Feature Engineering refers to the training data that pretreated data are changed into model
Process.Feature Engineering may include feature construction, feature extraction and feature selecting, wherein feature construction is from initial data
Artificial finds out some features (i.e. data) with physical significance；Feature extraction is that primitive character is converted to one group with bright
The feature of aobvious physical significance or statistical significance or core；Feature selecting is that one group of most statistical significance is selected from characteristic set
Feature reaches dimensionality reduction.Specifically, the method for feature extraction may include principal component analysis (Principal Component
Analysis, PCA), linear discriminant analysis (Linear Discriminant Analysis, LDA) and singular value decomposition
One of (Singular Value Decomposition, SVD)；The method of feature selecting may include filtering type
(Filter) one of selection, packaging type (Wrapper) selection and embedded (Embedding) selection.
In this embodiment, the purpose of Feature Engineering is to delete the abundant or invalid data of part meaning；According to industry
Experience screens feature, obtains factor that Coke Quality has a major impact as characteristic index；And use correlation
The correlation of algorithm calculating feature and coke quality.
Specifically, the part X1-Xn in table 1 contains all coke production technique supplemental characteristic and mixed coal data,
Some data are that height is relevant to high-value product yield in these data, some data are weak related to high-value product yield
Or it is incoherent, therefore, it is necessary to be screened by correlation analysis to X1-Xn, retaining Coke Quality data Y has weight
Influence data (i.e. effectively).Here, correlation analysis (Correlation analysis) refers to two or more tools
The variable element of standby correlation is analyzed, to measure the related intimate degree of two Variable Factors.Correlation analysis
It can include but is not limited to chart correlation analysis (line chart and scatter plot), covariance and covariance matrix, related coefficient
(Correlation coefficient), regression analysis (Regression analysis) and comentropy and mutual information
(Mutual information) etc..
Further, chart correlation analysis is that data are carried out visualization processing, is briefly exactly to draw a diagram.Association side
Difference is used to measure the global error of two variables, if the variation tendency of two variables is consistent, covariance is exactly positive value, illustrates two
A variable is positively correlated；If the variation tendency of two variables on the contrary, covariance is exactly negative value, illustrates two variable negative correlation；Such as
Two variables of fruit are mutually indepedent, then covariance is exactly 0, illustrate that two variables are uncorrelated.Related coefficient is between response variable
The statistical indicator of degree in close relations, the value interval of related coefficient is between 1 to -1, wherein 1 indicates two complete lines of variable
Property it is related, -1 indicates two variable perfect negative correlations, and 0 indicates that two variables are uncorrelated；There are three types of the calculating of related coefficient, i.e. skin
Er Xun (Pearson) related coefficient, Ken Deer (Kendall) related coefficient and Spearman (Spearman) rank correlation coefficient.
Regression analysis is the statistical method of two groups or more determining relationship between variables, and regression analysis is divided into one according to the quantity of variable
Member returns and multiple regression.Comentropy is used for the unordered degree of metric, and mutual information is actually wider relative entropy
Special case.
Further, it is also possible to derive new feature using such as One-Hot coding based on the feature filtered out.Here,
One-Hot coding is also known as an efficient coding, is mainly encoded using N bit status register to N number of state, each
State all by his independent register-bit, and only have when any one effectively.Table 2 is shown to be obtained by Feature Engineering
The example of feature and coke quality data.
Table 2
| XF1 | XF2 | … | XFn | Y |
| X1_1 | X2_1 | … | Xn_1 | Y_1 |
| X1_2 | X2_2 | … | Xn_2 | Y_2 |
| X1_3 | X2_3 | … | Xn_3 | Y_3 |
| X1_4 | X2_4 | … | Xn_4 | Y_4 |
| … | … | … | … | … |
| X1_m | X2_m | … | Xn_m | Y_m |
Wherein, XF1, XF2 ..., XFn are X1, the subset of X2 ..., Xn and the new feature derived.
It should be noted that one sample of each behavior in table 2, XF1, XF2 ..., XFn is corresponding all to be classified as sample
Feature, Y is corresponding to be classified as coke quality data.
In another embodiment of the present invention, Coke Quality prediction model is trained, and obtains the first coke quality
Data, comprising: to after screening coke production technique supplemental characteristic and mixed coal data split, obtain prediction of coke quality
Training set；Mould is predicted using the sample in prediction of coke quality training set as the input Coke Quality of Coke Quality Prediction Models
Type is trained, and obtains the first coke quality data, wherein sample be prediction of coke quality training be concentrated through screening obtain
Characteristic.
Specifically, to after screening coke production technique supplemental characteristic and mixed coal data split, obtain coke matter
Amount prediction training set；Further, prediction of coke quality training is used to be concentrated through the obtained characteristic of screening as coke
The input of quality prediction model, output of the corresponding single coke quality index as Coke Quality Prediction Models, training coke
Quality prediction model is to obtain the first coke quality data.
It should be noted that generally require sample being divided into independent three in the fields such as machine learning and pattern-recognition
Part, i.e. training set (Train set), verifying collection (Validation set) and test set (Test set), wherein training set
For estimating model, verifying collection is used to determine the parameter of network structure or Controlling model complexity, and test set is then examined
How is the performance of the optimal model of final choice.
In another embodiment of the present invention, optimizing index data are treated using optimization algorithm to optimize, and according to
Optimum results obtain the recommended value of achievement data to be optimized, comprising: and optimizing index data are treated using optimization algorithm and are optimized,
Obtain the second coke quality data；Second coke quality data are compared with the first coke quality data, and according to comparing
As a result the recommended value of achievement data to be optimized is obtained.
Specifically, optimizing index data are treated using optimization algorithms such as genetic algorithm, particle swarm algorithms to optimize,
Obtain the second coke quality data；Further, the second coke quality data are compared with the first coke quality data, such as
Fruit the second coke quality data are better than the first coke quality data, then using the second coke quality data as achievement data to be optimized
Recommended value.
In the following, using mixed coal volatile matter and mixed coal sulphur content as index to be optimized, to how obtaining index number to be optimized
According to recommended value be illustrated.
Table 3 shows the example of optimization front and back mixed coal volatile matter, mixed coal sulphur content and coke quality scoring.As shown in table 3,
The value of mixed coal volatile matter and mixed coal sulphur content is respectively 20 and 10 before optimization, after optimization algorithm, is searched out new
The value of mixed coal volatile matter and mixed coal sulphur content is respectively 15 and 15, at this point, coke quality is increased to 80 by 40 before, it can
See, coke quality is obviously improved, it is therefore possible to use the recommended value of optimizing index, that is, subsequent in production mixed coal
When, so that mixed coal volatilization is divided into 15, mixed coal sulphur content is 15.
Table 3
| Mixed coal volatile matter | Mixed coal sulphur content | Coke quality scoring | |
| Before optimization | 20 | 10 | 40 |
| After optimization | 15 | 15 | 80 |
All the above alternatives can form alternative embodiment of the invention using any combination, herein no longer
It repeats one by one.
Fig. 2 is a kind of process of the method for optimization coke production shown in another exemplary embodiment according to the present invention
Figure.As shown in Fig. 2, this method comprises:
202: obtaining historical data, wherein historical data includes coke production technique supplemental characteristic, mixed coal data and coke
Charcoal qualitative data；
204: data prediction being carried out to historical data, obtains high quality historical data；
206: high quality historical data being screened by correlation calculations, obtains sieve relevant to coke quality data
Coke production technique supplemental characteristic and mixed coal data after choosing；
208: to after screening coke production technique supplemental characteristic and mixed coal data split, it is pre- to obtain coke quality
Survey training set；
210: using the sample in prediction of coke quality training set as the input Coke Quality of Coke Quality Prediction Models
Prediction model is trained, and obtains the first coke quality data, wherein sample is that prediction of coke quality training is concentrated through screening
Obtained characteristic；
212: according to the first coke quality data and business it needs to be determined that index to be optimized in the first coke quality data
Data；
214: optimizing index data being treated using optimization algorithm and are optimized, the second coke quality data are obtained；
216: the second coke quality data are compared with the first coke quality data；
218: if the second coke quality data be better than the first coke quality data, using the second coke quality data as
The recommended value of achievement data to be optimized.
220: coke production technique supplemental characteristic or mixed coal data being adjusted according to recommended value, to improve coke quality.
The technical solution provided according to embodiments of the present invention can take into account coke when constructing Coke Quality Prediction Models
Therefore technological production norm data and mixed coal data more tally with the actual situation and prediction of coke quality result are more credible；Make
Optimizing is carried out to data with optimization algorithm, therefore, more precise and high efficiency；It is not limited to certain algorithm model, but is more focused on
Optimization algorithm is modeled and selected according to data characteristic, and therefore, flexibility is stronger.
Following is apparatus of the present invention embodiment, can be used for executing embodiment of the present invention method.For apparatus of the present invention reality
Undisclosed details in example is applied, embodiment of the present invention method is please referred to.
Fig. 3 is a kind of block diagram of the device of optimization coke production shown in an exemplary embodiment according to the present invention.Such as Fig. 3
Shown, which includes:
Module 310 is constructed, is configured to construct Coke Quality Prediction Models according to historical data, and Coke Quality predicts mould
Type is trained, and obtains the first coke quality data, wherein historical data includes coke production technique supplemental characteristic, mixed coal
Data and coke quality data；
Determining module 320 is configured to according to the first coke quality data and business it needs to be determined that the first coke quality data
In achievement data to be optimized；
Optimization module 330 is configured to be treated optimizing index data using optimization algorithm and optimized, and according to optimum results
Obtain the recommended value of achievement data to be optimized；
Module 340 is adjusted, is configured to adjust coke production technique supplemental characteristic or mixed coal data according to recommended value, to mention
High coke quality.
The technical solution provided according to embodiments of the present invention, by constructing Coke Quality Prediction Models according to historical data,
And Coke Quality prediction model is trained, and obtains the first coke quality data, wherein historical data includes coke production skill
Art supplemental characteristic, mixed coal data and coke quality data；According to the first coke quality data and business it needs to be determined that first is burnt
Achievement data to be optimized in charcoal qualitative data；It treats optimizing index data using optimization algorithm to optimize, and according to optimization
As a result the recommended value of achievement data to be optimized is obtained；Coke production technique supplemental characteristic or mixed coal number are adjusted according to recommended value
According to can fully consider coke production technique parameter and mixed coal Coke Quality in actual production to improve coke quality
It influences, therefore, optimizes coke production.
In another embodiment of the present invention, the device of Fig. 3 further include:
Module 350 is obtained, is configured to obtain historical data；
Preprocessing module 360 is configured to carry out data prediction to historical data, obtains high quality historical data；
Screening module 370 is configured to screen high quality historical data by correlation calculations, obtain and coke matter
Coke production technique supplemental characteristic and mixed coal data after measuring the relevant screening of data.
In another embodiment of the present invention, data prediction includes Missing Data Filling processing, data format exception
At least one of reason, data value range abnormality processing and data reprocessing.
In another embodiment of the present invention, construct 310 pairs of module screening after coke production technique supplemental characteristic and
Mixed coal data are split, and obtain prediction of coke quality training set, and using the sample in prediction of coke quality training set as
The input Coke Quality prediction model of Coke Quality Prediction Models is trained, and obtains the first coke quality data, wherein sample
It originally is that prediction of coke quality training is concentrated through the characteristic that screening obtains.
In another embodiment of the present invention, optimization module 330 treats the progress of optimizing index data using optimization algorithm
Optimization, obtains the second coke quality data, the second coke quality data is compared with the first coke quality data, and according to
Comparison result obtains the recommended value of achievement data to be optimized.
In another embodiment of the present invention, optimization module 330 is by the second coke quality data and the first coke quality
Data are compared, if the second coke quality data are better than the first coke quality data, the second coke quality data are made
For the recommended value of achievement data to be optimized.
In another embodiment of the present invention, achievement data to be optimized include coke quality sulphur content, it is reactivity indexes, anti-
At least one of intensity, crushing strength and wear-resistant strength after answering.
In another embodiment of the present invention, optimization algorithm includes genetic algorithm, particle swarm algorithm, ant group algorithm, mould
One of quasi- annealing algorithm, crowd's searching algorithm and artificial bee colony algorithm.
The function of modules and the realization process of effect are specifically detailed in the above method and correspond to step in above-mentioned apparatus
Realization process, details are not described herein.
Fig. 4 is a kind of block diagram of the equipment of optimization coke production shown in an exemplary embodiment according to the present invention.Such as Fig. 4
Shown, which includes at least memory 410 and processor 420, is stored with computer program, processor on memory 410
420 realize the method that embodiment as described above provides when executing the computer program on memory 410, comprising:
Coke Quality Prediction Models are constructed according to historical data, and Coke Quality prediction model is trained, and obtains the
One coke quality data, wherein historical data includes coke production technique supplemental characteristic, mixed coal data and coke quality number
According to；
According to the first coke quality data and business it needs to be determined that achievement data to be optimized in the first coke quality data；
Optimizing index data are treated using optimization algorithm to optimize, and obtain achievement data to be optimized according to optimum results
Recommended value；
Coke production technique supplemental characteristic or mixed coal data are adjusted according to recommended value, to improve coke quality.
The technical solution provided according to embodiments of the present invention, by constructing Coke Quality Prediction Models according to historical data,
And Coke Quality prediction model is trained, and obtains the first coke quality data, wherein historical data includes coke production skill
Art supplemental characteristic, mixed coal data and coke quality data；According to the first coke quality data and business it needs to be determined that first is burnt
Achievement data to be optimized in charcoal qualitative data；It treats optimizing index data using optimization algorithm to optimize, and according to optimization
As a result the recommended value of achievement data to be optimized is obtained；Coke production technique supplemental characteristic or mixed coal number are adjusted according to recommended value
According to can fully consider coke production technique parameter and mixed coal Coke Quality in actual production to improve coke quality
It influences, therefore, optimizes coke production.
Following computer program can also be performed in processor 420: obtaining historical data；Data are carried out to historical data to locate in advance
Reason, obtains high quality historical data；High quality historical data is screened by correlation calculations, is obtained and coke quality number
According to the coke production technique supplemental characteristic and mixed coal data after relevant screening.
Further, data prediction includes that Missing Data Filling processing, data format abnormality processing, data value range are different
At least one of often processing and data reprocessing.
Following computer program can also be performed in processor 420: to the coke production technique supplemental characteristic after screening and matching
It closes coal data to be split, obtains prediction of coke quality training set；Using the sample in prediction of coke quality training set as coke
The input Coke Quality prediction model of quality prediction model is trained, and obtains the first coke quality data, wherein sample is
Prediction of coke quality training is concentrated through the characteristic that screening obtains.
Following computer program can also be performed in processor 420: it is excellent to treat the progress of optimizing index data using optimization algorithm
Change, obtains the second coke quality data；Second coke quality data are compared with the first coke quality data, and according to than
Relatively result obtains the recommended value of achievement data to be optimized.
Following computer program can also be performed in processor 420: by the second coke quality data and the first coke quality number
According to being compared；If the second coke quality data be better than the first coke quality data, using the second coke quality data as
The recommended value of achievement data to be optimized.
Further, achievement data to be optimized include coke quality sulphur content, it is reactivity indexes, post reaction strength, anti-crushing strong
At least one of degree and wear-resistant strength.
Further, optimization algorithm includes that genetic algorithm, particle swarm algorithm, ant group algorithm, simulated annealing, crowd search
One of rope algorithm and artificial bee colony algorithm.
Above embodiments are only exemplary embodiment of the present invention, are not used in the limitation present invention, protection scope of the present invention
It is defined by the claims.Those skilled in the art can within the spirit and scope of the present invention make respectively the present invention
Kind modification or equivalent replacement, this modification or equivalent replacement also should be regarded as being within the scope of the present invention.
Claims (10)
1. a kind of method for optimizing coke production, comprising:
Coke Quality Prediction Models are constructed according to historical data, and the Coke Quality Prediction Models are trained, obtain the
One coke quality data, wherein the historical data includes coke production technique supplemental characteristic, mixed coal data and coke quality
Data；
According to the first coke quality data and business it needs to be determined that index to be optimized in the first coke quality data
Data；
The achievement data to be optimized is optimized using optimization algorithm, and obtains the index to be optimized according to optimum results
The recommended value of data；
The coke production technique supplemental characteristic or the mixed coal data are adjusted according to the recommended value, to improve coke matter
Amount.
2. the method according to claim 1, wherein constructing prediction of coke quality mould according to historical data described
Before type, the method also includes:
Obtain the historical data；
Data prediction is carried out to the historical data, obtains high quality historical data；
The high quality historical data is screened by correlation calculations, obtains sieve relevant to the coke quality data
Coke production technique supplemental characteristic and mixed coal data after choosing.
3. according to the method described in claim 2, it is characterized in that, the data prediction includes Missing Data Filling processing, number
According at least one of format abnormality processing, data value range abnormality processing and data reprocessing.
4. according to the method described in claim 2, it is characterized in that, described be trained the Coke Quality Prediction Models,
Obtain the first coke quality data, comprising:
To after the screening coke production technique supplemental characteristic and mixed coal data split, obtain prediction of coke quality instruction
Practice collection；
Using the sample in the prediction of coke quality training set as the input of the Coke Quality Prediction Models to the coke
Quality prediction model is trained, and obtains the first coke quality data, wherein the sample is the prediction of coke quality
Training is concentrated through the characteristic screened and obtained.
5. the method according to claim 1, wherein described use optimization algorithm to the achievement data to be optimized
It optimizes, and obtains the recommended value of the achievement data to be optimized according to optimum results, comprising:
The achievement data to be optimized is optimized using the optimization algorithm, obtains the second coke quality data；
The second coke quality data are compared with the first coke quality data, and institute is obtained according to comparison result
State the recommended value of achievement data to be optimized.
6. according to the method described in claim 5, it is characterized in that, described by the second coke quality data and described first
Coke quality data are compared, and obtain the recommended value of the achievement data to be optimized according to comparison result, comprising:
The second coke quality data are compared with the first coke quality data；
If the second coke quality data are better than the first coke quality data, by the second coke quality data
Recommended value as the achievement data to be optimized.
7. the method according to any one of claims 1 to 6, which is characterized in that the achievement data to be optimized includes
At least one of coke quality sulphur content, reactivity indexes, post reaction strength, crushing strength and wear-resistant strength.
8. the method according to any one of claims 1 to 6, which is characterized in that the optimization algorithm includes that heredity is calculated
One of method, particle swarm algorithm, ant group algorithm, simulated annealing, crowd's searching algorithm and artificial bee colony algorithm.
9. a kind of device for optimizing coke production, comprising:
Module is constructed, is configured to construct Coke Quality Prediction Models according to historical data, and to the Coke Quality Prediction Models
It is trained, obtains the first coke quality data, wherein the historical data includes coke production technique supplemental characteristic, cooperation
Coal data and coke quality data；
Determining module is configured to according to the first coke quality data and business it needs to be determined that the first coke quality data
In achievement data to be optimized；
Optimization module is configured to optimize the achievement data to be optimized using optimization algorithm, and is obtained according to optimum results
To the recommended value of the achievement data to be optimized；
Module is adjusted, is configured to adjust the coke production technique supplemental characteristic or the mixed coal number according to the recommended value
According to improve coke quality.
10. a kind of equipment for optimizing coke production, comprising: processor and memory, the processor load and execute described deposit
Instruction and data in reservoir, for realizing method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910156054.XA CN109858709B (en) | 2019-03-01 | 2019-03-01 | Method, device and equipment for optimizing coke production |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910156054.XA CN109858709B (en) | 2019-03-01 | 2019-03-01 | Method, device and equipment for optimizing coke production |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109858709A true CN109858709A (en) | 2019-06-07 |
| CN109858709B CN109858709B (en) | 2022-08-19 |
Family
ID=66899532
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910156054.XA Active CN109858709B (en) | 2019-03-01 | 2019-03-01 | Method, device and equipment for optimizing coke production |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109858709B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110276564A (en) * | 2019-07-10 | 2019-09-24 | 成渝钒钛科技有限公司 | For vanadium titanium blast furnace second level coke sulphur analysis control method |
| CN111240282A (en) * | 2019-12-31 | 2020-06-05 | 联想(北京)有限公司 | Process optimization method, device, equipment and computer readable storage medium |
| CN111461481A (en) * | 2020-02-25 | 2020-07-28 | 国网河南省电力公司电力科学研究院 | Power transmission cable quality analysis method based on neural network |
| CN111950854A (en) * | 2020-07-14 | 2020-11-17 | 包头钢铁（集团）有限责任公司 | Coke quality index prediction method based on multilayer neural network |
| CN112098263A (en) * | 2020-09-14 | 2020-12-18 | 山西亚鑫新能科技有限公司 | Method for parameter comprehensive prediction of coke thermal strength model |
| CN113110306A (en) * | 2020-01-10 | 2021-07-13 | 台达电子工业股份有限公司 | Machine table and operation method thereof |
| CN114266412A (en) * | 2021-12-29 | 2022-04-01 | 浙江中控技术股份有限公司 | Optimization method and device for coking production, electronic equipment and storage medium |
| CN114565150A (en) * | 2022-02-25 | 2022-05-31 | 青岛海尔工业智能研究院有限公司 | Method, device and equipment for optimizing foaming process and storage medium |
| CN114662763A (en) * | 2022-03-24 | 2022-06-24 | 包头钢铁（集团）有限责任公司 | Method and system for evaluating cost performance of single coal for coking coal blending |
| CN115034465A (en) * | 2022-06-01 | 2022-09-09 | 黄河科技集团创新有限公司 | Coke quality prediction method based on artificial intelligence |
| JPWO2023276523A1 (en) * | 2021-06-28 | 2023-01-05 | ||
| CN118331210A (en) * | 2024-06-12 | 2024-07-12 | 合肥城市云数据中心股份有限公司 | Intelligent optimization method for coking production process based on knowledge distillation |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101541030A (en) * | 2009-05-06 | 2009-09-23 | 华为技术有限公司 | Method for predicting data based on support vector machine and equipment thereof |
| CN104899425A (en) * | 2015-05-07 | 2015-09-09 | 浙江大学 | Variable selection and forecast method of silicon content in molten iron of blast furnace |
| CN106249724A (en) * | 2016-09-14 | 2016-12-21 | 东北大学 | A kind of blast furnace polynary molten steel quality forecast Control Algorithm and system |
| CN108940596A (en) * | 2018-07-12 | 2018-12-07 | 浙江大学 | A kind of electric dust removing system optimal control method based on mixed model |
-
2019
- 2019-03-01 CN CN201910156054.XA patent/CN109858709B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101541030A (en) * | 2009-05-06 | 2009-09-23 | 华为技术有限公司 | Method for predicting data based on support vector machine and equipment thereof |
| CN104899425A (en) * | 2015-05-07 | 2015-09-09 | 浙江大学 | Variable selection and forecast method of silicon content in molten iron of blast furnace |
| CN106249724A (en) * | 2016-09-14 | 2016-12-21 | 东北大学 | A kind of blast furnace polynary molten steel quality forecast Control Algorithm and system |
| CN108940596A (en) * | 2018-07-12 | 2018-12-07 | 浙江大学 | A kind of electric dust removing system optimal control method based on mixed model |
Non-Patent Citations (2)
| Title |
|---|
| 刘淑清: "火电厂动力煤配煤优化方法与系统研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
| 赖旭芝等: "基于多目标遗传算法的炼焦生产过程优化控制", 《计算机集成制造系统》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110276564A (en) * | 2019-07-10 | 2019-09-24 | 成渝钒钛科技有限公司 | For vanadium titanium blast furnace second level coke sulphur analysis control method |
| CN111240282A (en) * | 2019-12-31 | 2020-06-05 | 联想(北京)有限公司 | Process optimization method, device, equipment and computer readable storage medium |
| CN111240282B (en) * | 2019-12-31 | 2021-12-24 | 联想(北京)有限公司 | Process optimization method, device, equipment and computer readable storage medium |
| CN113110306A (en) * | 2020-01-10 | 2021-07-13 | 台达电子工业股份有限公司 | Machine table and operation method thereof |
| CN113110306B (en) * | 2020-01-10 | 2022-08-02 | 台达电子工业股份有限公司 | Machine table and operation method thereof |
| CN111461481A (en) * | 2020-02-25 | 2020-07-28 | 国网河南省电力公司电力科学研究院 | Power transmission cable quality analysis method based on neural network |
| CN111950854B (en) * | 2020-07-14 | 2022-04-19 | 包头钢铁（集团）有限责任公司 | Coke quality index prediction method based on multilayer neural network |
| CN111950854A (en) * | 2020-07-14 | 2020-11-17 | 包头钢铁（集团）有限责任公司 | Coke quality index prediction method based on multilayer neural network |
| CN112098263A (en) * | 2020-09-14 | 2020-12-18 | 山西亚鑫新能科技有限公司 | Method for parameter comprehensive prediction of coke thermal strength model |
| JPWO2023276523A1 (en) * | 2021-06-28 | 2023-01-05 | ||
| WO2023276523A1 (en) * | 2021-06-28 | 2023-01-05 | Ｊｆｅスチール株式会社 | Method for generating post-reaction coke strength inference model, method for inferring post-reaction strength of coke, and method for producing coke |
| JP7332054B2 (en) | 2021-06-28 | 2023-08-23 | Ｊｆｅスチール株式会社 | Coke post-reaction strength estimation model generation method, post-reaction strength estimation method of coke, and coke production method |
| CN114266412A (en) * | 2021-12-29 | 2022-04-01 | 浙江中控技术股份有限公司 | Optimization method and device for coking production, electronic equipment and storage medium |
| CN114565150A (en) * | 2022-02-25 | 2022-05-31 | 青岛海尔工业智能研究院有限公司 | Method, device and equipment for optimizing foaming process and storage medium |
| CN114662763A (en) * | 2022-03-24 | 2022-06-24 | 包头钢铁（集团）有限责任公司 | Method and system for evaluating cost performance of single coal for coking coal blending |
| CN115034465A (en) * | 2022-06-01 | 2022-09-09 | 黄河科技集团创新有限公司 | Coke quality prediction method based on artificial intelligence |
| CN118331210A (en) * | 2024-06-12 | 2024-07-12 | 合肥城市云数据中心股份有限公司 | Intelligent optimization method for coking production process based on knowledge distillation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109858709B (en) | 2022-08-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109858709A (en) | A kind of method, apparatus and equipment optimizing coke production | |
| CN108459955B (en) | Software defect prediction method based on deep self-coding network | |
| Ratledge et al. | Using machine learning to assess the livelihood impact of electricity access | |
| CN111785329A (en) | Single-cell RNA sequencing clustering method based on confrontation automatic encoder | |
| Huang | A hybrid particle swarm optimization approach for clustering and classification of datasets | |
| CN112819604A (en) | Personal credit evaluation method and system based on fusion neural network feature mining | |
| Shingari et al. | A review of applications of data mining techniques for prediction of students’ performance in higher education | |
| Tosun | Using data envelopment analysis–neural network model to evaluate hospital efficiency | |
| CN117522607A (en) | Enterprise financial management system | |
| Saliminezhad et al. | Validity of unbalanced growth theory and sectoral investment priorities in Indonesia: Application of feature ranking methods | |
| Singh et al. | Feature selection based classifier combination approach for handwritten Devanagari numeral recognition | |
| CN118016280A (en) | Breast cancer clinical data analysis diagnosis and treatment platform based on artificial intelligence large language model | |
| CN110738245A (en) | automatic clustering algorithm selection system and method for scientific data analysis | |
| Wang et al. | Risk assessment of customer churn in telco using FCLCNN-LSTM model | |
| Sun et al. | Improved Commodity Supply Chain Performance Through AI and Computer Vision Techniques | |
| KR101768807B1 (en) | Method for technology forecasting using patent analysis and expert survey | |
| Moodi et al. | Evaluation of feature selection performance for identification of best effective technical indicators on stock market price prediction | |
| CN115660221B (en) | Oil and gas reservoir economic recoverable reserve assessment method and system based on hybrid neural network | |
| Xiang et al. | Exploring influential factors of CO2 emissions in China’s cities using machine learning techniques | |
| Kang et al. | Predicting Stock Closing Price with Stock Network Public Opinion Based on AdaBoost-IWOA-Elman Model and CEEMDAN Algorithm | |
| Alexandridis et al. | Forecasting crude oil prices using wavelet neural networks | |
| CN106446160A (en) | Content polymerization method and system oriented to mobile internet self-adaptive increments | |
| Huang et al. | An enhanced approach to optimizing the stock portfolio selection based on Modified Markowitz MV Method | |
| Mazumder et al. | Rectification-based knowledge retention for task incremental learning | |
| Farfan et al. | End-to-end dimensionality reduction and regression from 3D geological uncertainties to estimate oil reservoir simulations |
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 |