CN115755608A - Energy consumption optimization decision method for high-pressure roller mill - Google Patents
Energy consumption optimization decision method for high-pressure roller mill Download PDFInfo
- Publication number
- CN115755608A CN115755608A CN202211445935.1A CN202211445935A CN115755608A CN 115755608 A CN115755608 A CN 115755608A CN 202211445935 A CN202211445935 A CN 202211445935A CN 115755608 A CN115755608 A CN 115755608A
- Authority
- CN
- China
- Prior art keywords
- data
- energy consumption
- roller mill
- pressure roller
- roll gap
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
The invention relates to the field of mineral processing industry control parameter decision, in particular to a high-pressure roller mill energy consumption optimization decision method. The method comprises the following steps: s1, acquiring required data characteristics by a field sensor, and storing real-time data into a database; s2: preprocessing the acquired data, removing the data acquired in non-working time and the data which do not meet the requirements of roll deflection, adding two characteristics (throughput and energy consumption), and filtering; s3, clustering by using a K-means clustering method and taking the processing capacity as a standard, and selecting data with the lowest energy consumption in each category to be placed in a case library; s4: and (3) preprocessing the data at the current moment in the same way as S2, and finding the closest data in the case base by taking the similarity as a standard to serve as output, wherein the similarity is calculated by adopting a maximum correlation coefficient to determine the weighted Euclidean distance of the weight. The invention combines historical experience, considers the processing capacity and energy consumption at the same time, intelligently selects the optimal parameters, and can continuously learn and update the data in the case base.
Description
Technical Field
The invention relates to the field of control parameter decision of the mineral processing industry, in particular to an energy consumption optimization decision method of a high-pressure roller mill.
Background
The high-pressure roller mill is widely applied to different fields of metallurgy, chemical industry, cement and the like, and especially plays a main role in the mineral separation process. The operation process of the high-pressure roller mill is still based on the traditional control theory at present, and for parameter setting, the problems of large result error, unstable working condition and the like exist depending on manual experience. Meanwhile, the setting of parameters is mostly aimed at high yield values, and cost problems and environmental problems caused by high energy consumption are ignored.
Therefore, the set values of the parameters need to be intelligently selected with the aim of reducing the energy consumption, the energy consumption is reduced while the output value is ensured, the cost of an enterprise can be reduced, and the best effort of environmental protection can also be realized.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides an energy consumption optimization decision method for a high-pressure roller mill, which reduces the energy consumption of the high-pressure roller mill under the condition of ensuring the treatment capacity. Therefore, a case base is established by adopting a K-means clustering method, an optimal parameter is searched according to the weighted similarity based on the maximum correlation coefficient and is input into a control system, and the final aim is achieved.
In order to achieve the purpose, the invention adopts the following technical scheme, comprising the following steps:
Step 2: preprocessing the acquired data, removing the data acquired in non-working time and the data which are not in accordance with the requirements of roll deviation, adding two characteristics of processing capacity and energy consumption, reconstructing the data characteristics, and finally filtering the data.
And 3, clustering by using a K-means clustering method and taking the processing capacity as a standard, and selecting the data with the lowest energy consumption in each category to be put into a case library.
And 4, preprocessing the data at the current moment in the same way as the data in the step 2, and finding the closest data in the case base by taking the similarity as a standard to serve as output, wherein the similarity is calculated by adopting a maximum correlation coefficient (MIC) to determine the weighted Euclidean distance of the weight.
Further, in the step 2, as for the sequence of data processing, firstly, removing the data which are collected in the non-working time and do not meet the requirements of the roll deviation, then adding two data characteristics of the processing amount Q and the energy consumption P according to the formulas (1.1) and (1.2), then obtaining a roll gap value S according to the formula (1.3), replacing a roll gap at a driving side and a roll gap at a non-driving side in a database, and finally performing Kalman filtering processing on the data to obtain input data after the data characteristics are recombined; wherein:
Q＝3600nDSρ (1.1)
where n is the rotational speed, D is the roll diameter, S is the roll gap, and ρ is the bulk density.
Wherein, P Liquid for treating urinary tract infection The constant value is different according to different models of the roller mill; p is Press roll Then according to the collected current I of the movable roller and the fixed roller Movable part 、I Stator Real-time changes of;
S＝(SND+SD)/2 (1.3)
wherein SD is a driving side roll gap, and SND is a non-driving side roll gap.
Further, in step 2, kalman filtering is adopted for data filtering, and for two data characteristics of fixed roll material saving valve displacement LS and moving roll material saving valve displacement LD, two noise parameters in the kalman filtering are as follows: q =0.001, r =0.1, and the filter parameters for the remaining data characteristics are q =0.001, r =1.
Further, in step 3, the data used by the K-means clustering method needs to be two-dimensional or more, so two data features of the processing capacity and the roll gap are selected for clustering, the types of the processing capacity smaller than the bottom line value are removed, and the sample with the lowest energy consumption is selected from the rest types and put into the case library to form the initial case library required in the subsequent algorithm.
Further, in step 4, for the calculation of the similarity SIM, combining the maximum correlation coefficient and the euclidean distance, calculating MIC values between different features and the total power according to a formula (1.5), determining a weight of each feature according to a formula (1.6), and finally calculating the SIM by using a formula (1.16); wherein:
where X and Y are the number of bins in the X and Y directions of the sample, respectively, and B is typically n as a rule of thumb 0.5～0.6 And n is the number of samples.
Wherein, ω is j Weight of the jth feature, MIC j Is the MIC value of the jth feature.
Wherein m is the number of features, x j For the j-th feature of the input, x kj Is the jth feature of the kth case.
Compared with the prior art, the invention has the beneficial effects.
The case base combines historical experience, can intelligently select optimal parameters, continuously learns and updates data in the case base, and avoids errors caused by manual parameter adjustment. Meanwhile, the energy consumption and the processing capacity are considered, and the aim of the invention is achieved.
Drawings
The invention is further described with reference to the following figures and detailed description. The scope of the invention is not limited to the following expressions.
Fig. 1 is a general flow diagram of an optimization decision control system.
Fig. 2 is a decision data output flow chart.
Detailed Description
As shown in fig. 1-2, the present invention comprises the steps of:
s1: the data characteristics required by the user are acquired by a field sensor, real-time data are stored in a database through communication between the PLC and an upper computer, and the data are acquired every two seconds. The collected related data comprises a moving roll current ID, a fixed roll current IS, a bin weight BW, a driving side roll gap SD, a non-driving side roll gap SND, a locking pressure FP, a fixed roll material saving valve displacement LS and a moving roll material saving valve displacement LD.
S2: in order to enable the next case base to be established more accurately and comprehensively, the invention needs to carry out some processing on the collected data. Firstly, removing data collected in non-working time, and ensuring that the data are all live data of work;
firstly, according to the working principle of the high-pressure roller mill, calculating the throughput Q and the total energy consumption P of the high-pressure roller mill as new characteristics:
Q＝3600nDSρ (1.1)。
where n is the rotational speed, D is the roll diameter, S is the roll gap, and ρ is the bulk density.
The total energy consumption mainly comprises a hydraulic system and a compression roller system of the high-pressure roller mill, the power of the hydraulic system can be calculated by adopting the installed power, so that P in the formula (1.2) Liquid for treating urinary tract infection The constant value is different according to different models of the roller mill. P Press roll The current changes in real time according to the collected current of the movable roller and the fixed roller, so that the P is mainly influenced by the current of the movable roller and the fixed roller;
further, in order to ensure the ore grinding effect and prolong the service life of the high-pressure roller mill, the roll deviation of the high-pressure roller mill needs to be within a certain range, so that the difference value of the roll gap SD of the driving side and the roll gap SND of the non-driving side is smaller than deltas, data which do not meet the conditions are deleted, the roll gap value is calculated according to the formula (1.3), and the value of the roll gap of the non-driving side, which is in close contact with the roll gap of the driving side, is replaced.
S＝(SND+SD)/2 (1.3)。
Preferably, kalman filtering is performed on the new feature combination to remove noise present in the data, and the specific method is as follows:
Step 2, calculating the estimation value of this time
And step 3: updating Kalman gain
And 4, step 4: updating estimated covariance
And actual covariance
Go to step 2.
And 5: and outputting the estimated value.
Preferably, the final database values are obtained.
The final decision is determined by the data in the case library, and in order to ensure the final result, the invention adopts a K-means clustering method to search the case with the lowest energy consumption under the condition of ensuring the processing capacity.
Firstly, selecting roll gaps and treatment capacity in a database as a two-dimensional sample set, setting categories as k, and setting corresponding category sets as C = { C = b B =1, 2.., k }, after which the center of each class is initialized to f b (b＝1，2，...，k)。
Preferably, the euclidean distance is taken as a criterion for similarity and distance determination.
Preferably, the final indicator of the end of the classification is the order of the objective function
And minimum.
Preferably, after the classification is completed, the processing amount is less than Q min Samples (the bottom line value of the processing amount) are abandoned, the rest samples are classified and concentrated, and the sample with the lowest energy consumption is selected to be placed in the case base to form the case base required by the next step.
S4: the data at the current moment is preprocessed in the same way as S2, enters a database, and is searched for the data with the maximum similarity as output.
First, MIC values of different characteristics with respect to the movable and fixed roller currents are calculated according to the definitional equation (1.5):
where X and Y are the number of bins in the X and Y directions of the sample, respectively, and B is typically n as a rule of thumb 0.5～0.6 (n is the number of samples).
Preferably, different features are given different weights according to MIC values between the different features and the total power, as follows:
wherein, ω is j Is the weight of the jth feature, MIC j Is the MIC value of the jth feature.
Preferably, the similarity is calculated
The value of (c).
Preferably, the cases with similarity degree greater than 85% are used as output, and the cases with similarity degree less than 85% are updated into the case base, and the output is the input data per se.
Preferably, the bin weight, the roll gap and the locking pressure in the output data are used as the input of the control system.
The method comprises the steps of firstly establishing new characteristics according to the working mechanism of the high-pressure roller mill, and filtering characteristic data after removing data which do not meet conditions. And establishing a case base by adopting a K-means clustering method, wherein the rule of selecting the case base is that the required energy consumption is the lowest on the premise of ensuring the minimum processing capacity. After the case base is established, calculating the correlation between the current of a movable fixed roller and other characteristics for the weight value of subsequent similarity calculation. And for the data input into the case base, matching the data in the case base according to the similarity by adopting the same processing mode as before, outputting the data with the similarity reaching the standard, and updating the data which does not reach the standard into the case base.
Example 1 is as follows:
the first step is as follows: and (5) data acquisition and transmission.
The data of the high-pressure roller mill is measured by a field sensor, the data is communicated with an upper computer through a PLC (programmable logic controller), the real-time data is stored in a database, and the data is sampled every two seconds. The method comprises the steps that new materials and return materials enter a bin through a belt, the amount of the materials entering a high-pressure roller mill IS controlled by the displacement of a material saving valve, and in the process, collected related data comprise moving roll current ID, fixed roll current IS, bin weight BW, driving side roll gap SD, non-driving side roll gap SND, locking pressure FP, fixed roll material saving valve displacement LS and moving roll material saving valve displacement LD.
The second step: and (4) preprocessing process data.
1. And (4) establishing a database.
The invention aims to reduce the energy consumption of the high-pressure roller mill, and the actually consumed power of the high-pressure roller mill mainly comprises two major parts, namely a hydraulic system and a compression roller system, which are defined as follows:
meanwhile, to ensure the processing capacity of the high-pressure roller mill, the calculation formula is as follows:
Q＝3600nDSρ。
where n is the rotational speed, D is the roll diameter, S is the roll gap, and ρ is the bulk density.
In the invention, data acquisition is carried out in real time, but the machine is not always in a working state, so that the data in a non-working period is removed firstly, and the energy consumption and the processing capacity at each moment are calculated according to a formula.
In order to ensure the stability of the working condition, the roll deviation should be smaller than Δ S, namely the difference between the driving side roll gap SD and the non-driving side roll gap SND should be smaller than Δ S, data which do not meet the condition should be removed, and the value of the roll gap S is calculated according to the formula (1.3). Finally, the characteristics in the database include a moving roll current ID, a fixed roll current IS, a bin weight BW, a roll gap S, a locking pressure FP, a fixed roll material saving valve displacement LS, a moving roll material saving valve displacement LD, a handling capacity Q and a total power P.
2. And (4) preprocessing data.
Certain noise exists in the production process data in the acquisition process, the accuracy of the subsequent algorithm of the people is affected, and therefore Kalman filtering processing needs to be carried out on the data. For the fixed roll material-saving valve displacement LS and the movable roll material-saving valve displacement LD, the parameter q =0.001, r =0.1 in Kalman filtering, and the remaining data characteristic q =0.001, r =1.
The third step: and establishing a case base based on the K-means clustering.
The control process of the high pressure roller mill needs to ensure the ore grinding throughput, so data less than Qmin are removed. Carrying out K-means clustering on the processing capacity Q of the residual data, dividing K categories, sorting according to the numerical value of the total power P, and putting the data with the lowest energy consumption into a case library to obtain the case library containing K cases.
1. Selecting a sample set from a database
G＝{g 1 ,g 2 ,...,g N }
＝{(Q 1 ,S 1 ),(Q 2 ,S 2 ),...,(Q N ,S N )}。
Randomly selecting k initial cluster centers, and recording as:
2. and calculating the distance between each sample and the k initial cluster centers, wherein in the invention, the distance calculation adopts Euclidean distance, and the formula is as follows:
wherein, N is the number of samples, k is the number of categories, and after traversing all samples, the sample data conforming to the following formula is classified into the corresponding cluster:
d_m＝min[d(g a ,f b )].
3. updating the center of each cluster according to the following formula, comparing the center with the center of the last iteration cluster, and judging whether the center changes:
wherein b is class b, l is the number of iterations, and n is the number of samples assigned to class b.
4. Calculate the value of the objective function J:
judgment of
And if the answer is not true, turning to 2, and if true, turning to 5.
5. And outputting the final classification after the iteration is finished.
6. Centralize the classification with a throughput less than Q min And (4) removing the classes, comparing the energy consumptions P corresponding to the samples in the rest classes, and selecting a group of data with the minimum energy consumption to be placed in a final case library.
The fourth step: and (4) carrying out decision optimization based on case-based reasoning.
Based on case reasoning, according to data at the current moment, a case with the highest Similarity (SIM) in a case base is searched, the numerical value of the corresponding case is output, the value at the next moment is optimized, and then the control system controls the operation of the high-pressure roller mill. In the invention, the similarity calculation adopts the weighted Euclidean distance of the weight determined by MIC, and the specific steps are as follows:
1. calculating MIC values between different characteristics and total power according to a formula (1.5), determining a weight value of each characteristic according to a formula (1.6), and calculating the following results: table 1. Weight corresponding to different features.
2. The value of the SIM is calculated according to the following equation.
Wherein the content of the first and second substances,m is a characteristic number, x j Is the j-th feature of the input, x kj Is the jth feature of the kth case.
3. And (3) taking the case characteristic with the maximum SIM value calculated in the step (2) as an output, recording the case characteristic as y, and inputting three characteristics of a roll gap S, a locking pressure FP and a bin weight BW into a control system to control the operation of the high-pressure roller mill. And if the maximum value is less than 85%, updating the data into the case library, and taking the case as output.
4. The case base is continuously updated, and the prediction precision is improved.
It should be understood that the detailed description and specific examples, while indicating preferred embodiments of the invention, are given by way of illustration only, not limitation, and it will be understood by those skilled in the art that various changes and modifications may be made therein without departing from the spirit and scope of the invention; as long as the use requirements are met, the method is within the protection scope of the invention.
Claims (5)
1. A high-pressure roller mill energy consumption optimization decision method is characterized by comprising the following steps: the method comprises the following steps:
step 1, acquiring required data characteristics, and storing real-time data into a database through communication between a PLC and an upper computer;
and 2, step: preprocessing the acquired data, removing the data acquired in non-working time and the data which are not in accordance with the requirements of roll deviation, adding two characteristics of processing capacity and energy consumption, reconstructing the data characteristics, and finally filtering the data;
step 3, clustering by using a K-means clustering method and taking the processing capacity as a standard, and selecting data with the lowest energy consumption in each category to be placed in a case library;
and 4, preprocessing the data at the current moment in the same way as in the step 2, and finding the closest data in the case base by taking the similarity as a standard to serve as output, wherein the similarity is calculated by adopting a maximum correlation coefficient to determine the weighted Euclidean distance of the weight.
2. The energy consumption optimization decision method of the high pressure roller mill according to claim 1, characterized in that: in the step 2, for the sequence of data processing, firstly, removing the data which are collected in the non-working time and do not meet the requirements of the roll deviation, then adding two data characteristics of the processing amount Q and the energy consumption P according to the formulas (1.1) and (1.2), then obtaining a roll gap value S according to the formula (1.3), replacing a driving side roll gap and a non-driving side roll gap in a database, and finally performing Kalman filtering processing on the data after the data characteristics are recombined to obtain input data; wherein:
Q＝3600nDSρ (1.1)
where n is the rotational speed, D is the roll diameter, S is the roll gap, and ρ is the bulk density.
Wherein, P Liquid for treating urinary tract infection The constant value is different according to different models of the roller mill; p is Press roll Then according to the collected current I of the movable roller and the fixed roller Movable part 、I Stator Real-time changes of;
S＝(SND+SD)/2 (1.3)
wherein SD is a driving side roll gap, and SND is a non-driving side roll gap.
3. The energy consumption optimization decision method of the high pressure roller mill according to claim 1, characterized in that: in the step 2, kalman filtering is adopted for data filtering, and for two data characteristics of fixed roller material-saving valve displacement LS and movable roller material-saving valve displacement LD, two noise parameters in the Kalman filtering are as follows: q =0.001, r =0.1, and the filter parameters for the remaining data features are q =0.001, r =1.
4. The energy consumption optimization decision method of the high pressure roller mill according to claim 1, characterized in that: in step 3, the data used by the K-means clustering method needs to be two-dimensional or more, so two data characteristics of the processing capacity and the roll gap are selected for clustering, the types of the processing capacity smaller than the bottom line value are removed, and the sample with the lowest energy consumption is selected from the rest types and put into a case library to form an initial case library required in a subsequent algorithm.
5. The energy consumption optimization decision method for the high-pressure roller mill according to claim 1, characterized in that: step 4, calculating the similarity SIM, combining the maximum correlation coefficient and the Euclidean distance, calculating MIC values between different characteristics and total power according to a formula (1.5), determining the weight of each characteristic according to a formula (1.6), and finally calculating the SIM with a formula (1.16); wherein:
where X and Y are the number of bins in the X and Y directions of the sample, respectively, and B is typically n as a rule of thumb 0.5～0.6 N is the number of samples;
wherein, ω is j Weight of the jth feature, MIC j Is the MIC value of the jth feature;
where m is the number of features, x j Is the j-th feature of the input, x kj Is the jth feature of the kth case.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211445935.1A CN115755608B (en) | 2022-11-18 | 2022-11-18 | Energy consumption optimization decision method for high-pressure roller mill |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211445935.1A CN115755608B (en) | 2022-11-18 | 2022-11-18 | Energy consumption optimization decision method for high-pressure roller mill |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115755608A true CN115755608A (en) | 2023-03-07 |
| CN115755608B CN115755608B (en) | 2023-05-30 |
Family
ID=85373267
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211445935.1A Active CN115755608B (en) | 2022-11-18 | 2022-11-18 | Energy consumption optimization decision method for high-pressure roller mill |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115755608B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030204484A1 (en) * | 2002-04-26 | 2003-10-30 | International Business Machines Corporation | System and method for determining internal parameters of a data clustering program |
| CN105241239A (en) * | 2015-09-10 | 2016-01-13 | 广西大学 | Intelligent optimal control method and device for sintered brick tunnel kiln roasting process |
| CN109443766A (en) * | 2018-09-10 | 2019-03-08 | 中国人民解放军火箭军工程大学 | A kind of heavy-duty vehicle gearbox gear Safety Analysis Method |
| CN113239482A (en) * | 2021-04-23 | 2021-08-10 | 北京科技大学 | Dynamic prediction method and device for converter post-blowing carbon content |
| CN115291519A (en) * | 2022-08-16 | 2022-11-04 | 中南大学 | Intelligent optimization control method for ore grinding process |
-
2022
- 2022-11-18 CN CN202211445935.1A patent/CN115755608B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030204484A1 (en) * | 2002-04-26 | 2003-10-30 | International Business Machines Corporation | System and method for determining internal parameters of a data clustering program |
| CN105241239A (en) * | 2015-09-10 | 2016-01-13 | 广西大学 | Intelligent optimal control method and device for sintered brick tunnel kiln roasting process |
| CN109443766A (en) * | 2018-09-10 | 2019-03-08 | 中国人民解放军火箭军工程大学 | A kind of heavy-duty vehicle gearbox gear Safety Analysis Method |
| CN113239482A (en) * | 2021-04-23 | 2021-08-10 | 北京科技大学 | Dynamic prediction method and device for converter post-blowing carbon content |
| CN115291519A (en) * | 2022-08-16 | 2022-11-04 | 中南大学 | Intelligent optimization control method for ore grinding process |
Non-Patent Citations (3)
| Title |
|---|
| 关长亮;: "基于POS-BP的磨矿过程磨机负荷智能分析方法研究", 有色金属(选矿部分) * |
| 周平;柴天佑;: "基于案例推理的磨矿分级系统智能设定控制", 东北大学学报(自然科学版) * |
| 李利平;张春发;牛玉广;贡献;: "基于案例推理的热力机组在线运行优化调整决策方法", 中国电力 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115755608B (en) | 2023-05-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105278520B (en) | Based on T-KPRM complex industrial process evaluation of running status methods and application | |
| CN109597315B (en) | Method, equipment and system for identifying health degradation state of mechanical equipment | |
| WO2021036546A1 (en) | Near-infrared quantitative analysis model construction method based on biased estimation | |
| CN106181579A (en) | A kind of Tool Wear Monitoring method based on multisensor current signal | |
| CN110263474A (en) | A kind of cutter life real-time predicting method of numerically-controlled machine tool | |
| CN102169077B (en) | Hybrid intelligent soft measuring method of overflow granularity index in wet grinding process | |
| CN109847916B (en) | Energy-saving optimization method of cement raw material vertical mill system | |
| CN111633467A (en) | Cutter wear state monitoring method based on one-dimensional depth convolution automatic encoder | |
| CN111126255A (en) | Numerical control machine tool cutter wear value prediction method based on deep learning regression algorithm | |
| CN111958321B (en) | Numerical control machine tool cutter wear degree identification method based on deep neural network | |
| CN105095188B (en) | Sentence similarity computational methods and device | |
| CN108469797B (en) | Neural network and evolutionary computation based ore grinding process modeling method | |
| CN116597239B (en) | Processing method and system for recycled steel pipe for building | |
| CN1307415C (en) | Soft investigating method for overflow grain index of ore grinding system based on case inference | |
| CN110826624A (en) | Time series classification method based on deep reinforcement learning | |
| Yang et al. | A fuzzy-soft learning vector quantization for control chart pattern recognition | |
| CN115860211A (en) | Casting blank quality prediction method based on local online modeling | |
| CN112757053A (en) | Model fusion tool wear monitoring method and system based on power and vibration signals | |
| CN115755608A (en) | Energy consumption optimization decision method for high-pressure roller mill | |
| CN115169453A (en) | Hot continuous rolling width prediction method based on density clustering and depth residual error network | |
| CN111730412A (en) | Ant colony optimization algorithm-based micro milling cutter wear state monitoring method of support vector machine | |
| CN106406257A (en) | Iron ore flotation concentrate grade soft measurement method and system based on case-based reasoning | |
| CN114757087A (en) | Tool wear prediction method based on dynamic principal component analysis and LSTM | |
| CN113112121A (en) | Workshop layout scheduling optimization method based on multi-objective non-dominated sorting | |
| CN112946072A (en) | Abrasive belt wear state monitoring method based on machine learning |
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 |