CN114492956B - Working condition prediction method, equipment and storage medium based on historical data search - Google Patents
Working condition prediction method, equipment and storage medium based on historical data search Download PDFInfo
- Publication number
- CN114492956B CN114492956B CN202210016479.2A CN202210016479A CN114492956B CN 114492956 B CN114492956 B CN 114492956B CN 202210016479 A CN202210016479 A CN 202210016479A CN 114492956 B CN114492956 B CN 114492956B
- Authority
- CN
- China
- Prior art keywords
- working condition
- historical
- data
- space vector
- time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000003860 storage Methods 0.000 title claims abstract description 8
- 239000013598 vector Substances 0.000 claims abstract description 52
- 238000004364 calculation method Methods 0.000 claims abstract description 21
- 238000012163 sequencing technique Methods 0.000 claims abstract description 12
- 230000002159 abnormal effect Effects 0.000 claims description 15
- 238000009826 distribution Methods 0.000 claims description 11
- 230000008569 process Effects 0.000 abstract description 15
- 230000007246 mechanism Effects 0.000 abstract description 9
- 238000010801 machine learning Methods 0.000 abstract description 8
- 238000005457 optimization Methods 0.000 abstract description 4
- 238000013136 deep learning model Methods 0.000 abstract description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 23
- 238000002347 injection Methods 0.000 description 11
- 239000007924 injection Substances 0.000 description 11
- 230000008859 change Effects 0.000 description 10
- 238000000342 Monte Carlo simulation Methods 0.000 description 5
- 230000005856 abnormality Effects 0.000 description 5
- 230000009471 action Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 4
- 238000004088 simulation Methods 0.000 description 4
- 239000000243 solution Substances 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000001174 ascending effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000003054 catalyst Substances 0.000 description 1
- 238000004523 catalytic cracking Methods 0.000 description 1
- 238000012824 chemical production Methods 0.000 description 1
- 239000010779 crude oil Substances 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 206010027175 memory impairment Diseases 0.000 description 1
- 238000003801 milling Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of process parameter optimization, in particular to a working condition prediction method, device and storage medium based on historical data search, which comprises the steps of inputting current working condition data; expanding the current working condition data into working condition space vectors; calculating the similarity of the current working condition space vector and the historical working condition space vector, and obtaining a calculation result; sorting the calculation results to form a candidate set, and returning a sorting result; step S5, executing service operation according to the sequencing result; compared with a model based on a mechanism, the invention is closer to the actual working condition, thus having the strongest interpretation and being capable of helping an expert and an automatic control system to determine the optimal operation; the space vector is adopted, the space vector calculation is not limited by the number of samples, and all historical data participate in the space distance calculation, so that the method has no disastrous forgetting compared with the machine learning and deep learning model schemes; the method fully utilizes the existing data, does not depend on mechanism modeling and machine learning parameter tuning, and has large-scale application capability.
Description
Technical Field
The invention relates to the technical field of process parameter optimization, in particular to a working condition prediction method, equipment and a storage medium based on historical data search.
Background
The process industrial production pursues stable control, and operation parameters need to be adjusted in time according to on-line working conditions. Taking milling as an example, raw material variations can lead to product quality variations. In order to achieve the self-control industry, two technical paths are provided:
control based on mechanism. Such as optimizing catalytic cracking reactions based on the original database, etc. The disadvantage of this solution is that the mechanism model is different from the actual production environment and cannot meet the requirements of fine control. In addition, the data of the mechanism model come from long-term engineering time, and the cost is very high.
Control based on machine learning. The principle inputs different sample data to the model, so that the algorithm automatically obtains and records data characteristics which have obvious influence on the result. The disadvantages of this solution are:
the interpretability is poor. The output model does not understand the meaning of the correlation between the physical quantities, and the weighted meaning of the model features is difficult for humans to understand. The parameters of the model tuning are therefore difficult to maintain.
Catastrophic forgetfulness cannot be avoided. As previously mentioned, model training is based on a large amount of sample data. Most of the data of the actual working conditions are stable and have no fluctuation, and the sample data for training is very few. In this case, the model can learn only the rule of the general case, and neglect the fluctuating data, so that it is difficult to draw conclusions to improve the production efficiency.
The state history hypothesis (ergodic hypothesis) is a hypothesis that attempts to restore statistical regularity to mechanical regularity, proposed by l. He believes that an isolated system will experience all possible microscopic states from either initial state after a sufficient period of time.
According to the parameter optimization method based on the working condition retrieval, the process system device is assumed to have each state history, namely after a period of observation, all possible working conditions of the process can be observed, so that the retrieval algorithm can always find data similar to the current working condition in the historical working conditions.
Meng Te Carlo method (Monte Carlo method), also called statistical simulation method, is a very important numerical calculation method guided by probability statistical theory, which is proposed in the middle of the forty-th century due to the development of science and technology and the invention of electronic computer. Refers to a method of solving many of the computational problems using random numbers (or more commonly pseudo-random numbers). Corresponding to it is a deterministic algorithm. Meng Te the Carlo method is widely applied in the fields of financial engineering, macro-economics, computational physics (such as particle transport computation, quantum thermodynamic computation, aerodynamic computation) and the like.
The parameter optimization method based on the condition retrieval uses Monte Carlo method ideas to predict the condition evolution:
in order to predict the evolution result of the current working condition after the t time period, multiple simulation experiments are required to be carried out by taking the current working condition as an initial condition;
according to the various state experience hypotheses, the current working condition can be approximately considered to be repeatedly appeared in history;
by the searching method, a history working condition similar to the current working condition is found out to replace the simulation experiment in the Monte Carlo method;
the monte carlo method is used to give a prediction of the evolution result and the associated probability distribution. Because the history working condition is used for replacing the simulation experiment, the search method predicts the evolution result more truly and reliably. And the historical data can provide very strong interpretation of the results.
The states are not strictly required to be established, the method always gives a predicted result, and the confidence of the predicted result can be judged according to the quality of the search result.
Disclosure of Invention
Object of the invention
In order to solve the problems in the background art, the invention provides a working condition prediction method, equipment and a storage medium based on historical data search.
(II) technical scheme
In order to solve the above problems, the first aspect of the present invention provides a method for predicting a working condition based on historical data search, including the following steps:
step S1, inputting current working condition data;
s2, expanding current working condition data into working condition space vectors;
step S3, calculating the similarity between the current working condition space vector and the historical working condition space vector, and obtaining a calculation result;
s4, sorting the calculation results to form a candidate set, and returning a sorting result;
and S5, executing service operation according to the sequencing result.
Preferably, the method further comprises reading historical working condition data and expanding the historical working condition data into a historical working condition space vector.
Preferably, the historical working condition space vector is a full-historical moment working condition space vector or a partial-historical moment working condition space vector.
Preferably, executing the business operation according to the sorting result includes:
taking a certain historical moment t in the returned sequencing result as a starting point, and loading the historical data in the n time periods after loading;
in the n time period, calculating the time difference from the t moment to the abnormal moment and the occurrence times of the abnormal moment in the n time period;
and summarizing the result of the last step, and counting the probability and time difference distribution of abnormal time occurrence, wherein the probability and time difference distribution are used for predicting whether alarm information is sent out.
A second aspect of the invention provides a computing device comprising:
one or more processors; and
a memory; and
one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for the history data search based operating condition prediction method described above.
A third aspect of the present invention provides a computer readable storage medium having stored thereon one or more programs, the one or more programs comprising instructions adapted to be loaded from a memory and to perform the above-described history data search based operating condition prediction method.
The technical scheme of the invention has the following beneficial technical effects:
compared with a model based on a mechanism, the invention is closer to the actual working condition, thus having the strongest interpretation and being capable of helping an expert and an automatic control system to determine the optimal operation;
the invention adopts the space vector, the space vector calculation is not limited by the number of samples, and all historical data participate in the space distance calculation, so that compared with a machine learning and deep learning model scheme, the invention has no disastrous forgetting;
the invention fully utilizes the existing data, does not depend on mechanism modeling and machine learning parameter tuning, and therefore has the capacity of large-scale application.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a diagram of alarm data according to an embodiment of the present invention;
FIG. 3 is a diagram of time A according to an embodiment of the present invention;
fig. 4 is a diagram of time B data according to an embodiment of the present invention.
Detailed Description
The objects, technical solutions and advantages of the present invention will become more apparent by the following detailed description of the present invention with reference to the accompanying drawings. It should be understood that the description is only illustrative and is not intended to limit the scope of the invention. In addition, in the following description, descriptions of well-known structures and techniques are omitted so as not to unnecessarily obscure the present invention.
Referring to fig. 1, the first aspect of the present invention proposes a method for predicting a working condition based on historical data search, including the following steps: step S1, inputting current working condition data;
s2, expanding current working condition data into working condition space vectors; the purpose is to describe the current working condition state, the trend characteristics including the change, the adjustment action and the adjustment result of the control system, the action and the adjustment result of the manual control by using the data;
step S3, calculating the similarity between the current working condition space vector and the historical working condition space vector, and obtaining a calculation result; the aim is to convert the similarity into a quantifiable value. Similarity includes all-round comparisons such as current operating conditions, recent changes, control system actions and adjustment results, manually controlled actions and results, and the like. Assuming that the input current working condition space vector is A1, the calculation process is as follows:
the working condition space vectors in the candidate set are sequentially traversed or approximately traversed, wherein the historical working condition space vectors are abbreviated as B1 (hereinafter abbreviated as B1). And executing the subsequent steps circularly until all vectors in the candidate set and the current working condition space vector A1 finish calculation.
Calculating the vector space similarity of the current working condition space vectors A1 and B1, wherein the similarity can be measured by using methods such as Euclidean distance, vector included angle and the like;
recording a historical working condition space vector B1 which is found in the two steps and is relatively similar to the current working condition space vector A1;
when the data volume of the historical working condition is large, the calculation process can be accelerated in a clustering, quantization or graph theory mode, and the similar working condition space vector can be quickly found through approximate calculation or a partial traversal method;
the method specifically comprises the following steps: the definition of the similarity is the quantification of the similarity of two working conditions at different moments, and the similarity can be measured by using methods such as Euclidean distance, vector included angle and the like.
The calculation process is to compare the similarity between the input working condition vector and the historical data working condition vector.
S4, sorting the calculation results to form a candidate set, and returning a sorting result; the purpose is to select the nearest working condition according to the similarity.
The method comprises the following steps:
according to the calculation result of the step S3, working conditions which are closer to the current working conditions are obtained;
selecting similar working conditions according to the similarity and actual application, for example, directly selecting K most similar historical working conditions, sampling importance of the similar working conditions according to the service, or filtering conditions according to the service, and the like;
ordering according to the similarity, and returning the sequence result to the external system. The external system executes subsequent business steps according to the sequencing result, including but not limited to alarming, predicting working conditions, adjusting control parameters and the like.
The method aims at screening out the most similar working conditions by using the ordering and supporting other business applications by using the change process of the working conditions.
And S5, executing service operation according to the sequencing result.
The method aims at predicting according to the practical requirement of application and applying the prediction result to production, and comprises the following steps:
selecting a prediction target according to application requirements, such as system stability after a period of time, chemical synthesis product quality, process energy consumption and the like;
predicting the predicted target according to the similar conditions selected in step S4 includes, but is not limited to, the following methods:
based on the Monte Carlo method, the similar working conditions given in the step S4 are regarded as random experiments, and the statistical indexes (mean, variance and the like) of the targets and the prediction of probability distribution on possible values are given;
modeling similar working conditions based on a machine learning method, and giving predictions of statistical indexes (mean, variance and the like) and probability distribution on possible values of targets;
it is further contemplated that the method further includes reading the historical operating condition data and expanding to a historical operating condition space vector. The historical working condition space vector can be a full historical moment working condition space vector or a partial historical moment working condition space vector.
Wherein, executing the business operation according to the sorting result comprises: taking a certain historical moment t in the returned sequencing result as a starting point, and loading the historical data in the n time periods after loading; in the n time period, calculating the time difference from the t moment to the abnormal moment and the occurrence times of the abnormal moment in the n time period; and summarizing the result of the last step, and counting the probability and time difference distribution of abnormal time occurrence, wherein the probability and time difference distribution are used for predicting whether alarm information is sent out.
For a better understanding of the present invention, practical examples are provided below for further explanation of the present invention.
Referring to fig. 2, the target: predicting whether the device voltage will be below 1kV (kilovolt) is illustrative of the working process of the present application:
control target: interface in the reactor.
Influence factors: feed rate, water injection rate, temperature, voltage, current, catalyst, and feed properties.
Inputting working condition data:
the system monitors the on-line data in real time. The input process is that the program is automatically executed every minute. The system reads all relevant working condition data from the online data interface. Such as voltage, current, interface, water injection valve, crude oil properties. The actual working conditions are many, and only the data with the greatest influence on the example is used as an illustration. The data related to this example includes current, voltage, device water content indicator (load_water_l1, derivative indicator), fill valve flow rate. 18 in fig. 2: 04 (abbreviated as time S) the subsystem gives an alarm, at which:
current = 286A;
voltage = 2.05kV;
water content in the apparatus = 11084kg;
1# fill valve (X label) =0 kg/hour;
0# fill valve (Y label) =5521 kg/hour;
2# water injection valve (Z label) =20009 kg/hour;
the expansion is as working condition space vector:
near 30 minutes of data were read, 17: 04-18:04.
Calculating the water content index of the device, and calculating the water content comprehensive water injection drainage and other data of the device according to an engineering formula. The moisture content index does not have a physical meaning. 18 in this example: 04 minutes moisture index = 11084.
Calculating a change trend index, for example:
moisture 20 min change = 18:04 time-17: 44 time= -310;
current 3 min change = 18: current at time 04-18: time 01 current = 0;
discretizing the derivative index and the trend index. For example, the current level is divided according to the current level
Current >300 represents a high load;
the moisture content index <13000 represents a low moisture content;
the above indices are combined to form a spatial vector. The following is a mathematical expression of the conversion from the original data to the space vector. X1- > Xn is the raw data. Xex represents a derivative index. Xext represents a change trend index. Xexc represents a discretization index.
Calculating the similarity of the current working condition space vector and the historical working condition space vector (refer to fig. 3 and 4):
the Shi Gongkuang spatial vector is traversed. Here, 14 on days 2019-12-22: 26 minutes (time A for short) and 2020-11-16 days 5:51 minutes (time B for short) illustrate the calculation process and the sorting process.
Time a data:
current flow: 278A;
voltage: 1.75kV;
the water content index: 12849;
1# water injection valve=0;
no. 0 water injection valve=7439;
2# water injection valve=0;
time B data:
current flow: 385A;
voltage: 2.19kV;
the water content index: 16044;
1# water injection valve: 0;
0# water injection valve: 7247;
2# water injection valve: 1947;
calculating the Euclidean distance between the space vector at the moment S and the space vector at the moment A to obtain 0.8142;
the Euclidean distance between the space vector at time S and the space vector at time B is calculated to obtain 1.5311.
Sorting the calculation results according to the similarity:
sequencing from small to large according to Euclidean distance;
results: { time A, time B, … };
returning a sequencing result;
and executing business operation according to the sequencing result:
in this example, it is predicted whether the voltage is abnormal or not less than 1 kV.
Starting from a time stamp t at a certain historical moment in the returned sequencing result. A piece of history data after loading from t. In this example, the target gives an alarm 2 hours in advance. Thus loading the history data from t to t+2 hours, including current, voltage, water content index, water injection.
Calculating by taking the voltage dip as an abnormal signal:
time difference from time t to abnormal time;
the number of times of occurrence of abnormal time (no occurrence is recorded as 0) within 2 hours;
and summarizing the result of the last step, and counting the probability and time difference distribution of the occurrence of the abnormality. In this example, 97.34% of cases were abnormal. Therefore, the probability of occurrence of abnormality is 97.34%, and an alarm needs to be given. 70% of anomalies occur at 90 min plus or minus 10 min, so anomalies are expected to occur at 19:24 minutes.
And sending out alarm information to prompt an operator that the voltage is abnormal.
The invention has the technical effects that: operability interpretation:
in this example, the voltage abnormality is predicted 101 minutes in advance. In this example, the time a causes voltage abnormality for the same reason, wherein the water content index is closest to the time S value and belongs to the water content lower section. The system always operates stably under the high-load condition at the moment B, wherein the water content index is greatly different from the moment S in value, and the change trend is opposite. Thus, although the current at time A, B is in a high load state, both are different operating conditions than at time S. The method is used for accurately finding the time A and predicting that the system is about to generate voltage abnormality according to the voltage change trend of the time A.
Meanwhile, the advantages of the invention compared with other technical schemes are fully reflected:
aiming at a mechanism model: the scheme is driven by data, does not need laboratory modeling, and has the advantage of large-scale application.
For machine learning models:
the time A obtained by searching, namely the subsequent voltage change data, has good interpretation. The water content at the time A is lower, and the water content at the time B is normal and in an ascending trend can be intuitively found through the chart. A voltage anomaly occurs at time a. The occurrence of anomalies at time S and time a is already sufficiently similar that voltage anomalies must occur over time.
As an example, time a occurs 12 months in 2019, approximately one year from time S. Such problems account for less than 0.5% of the total production time. The method adopting machine learning can not learn the characteristics of the moment A due to huge data distribution difference, namely, catastrophic forgetting occurs. The predictive effect of the present solution cannot be achieved.
In an alternative embodiment, the present invention combines search engine technology with chemical production control.
Theoretical basis: the various states of the process are theorized, monte Carlo. The invention is based on accurate prediction by using approximate working conditions.
Because the same or enough similar history working conditions can be accurate, multiple searching/searching can deduce the change trend of the system;
because the invention makes full use of the historical data, the comparison model technical path has the characteristics of strong interpretability and no disastrous forgetting;
the invention does not depend on a mechanism model, so the invention has universality and is not limited to a specific reaction link.
A second aspect of the invention provides a computing device comprising:
one or more processors; and
a memory; and
one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for the history data search based operating condition prediction method described above.
A third aspect of the present invention provides a computer readable storage medium having stored thereon one or more programs, the one or more programs comprising instructions adapted to be loaded from a memory and to perform the above-described history data search based operating condition prediction method.
The foregoing has shown and described the basic principles, principal features and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the above-described embodiments, and that the above-described embodiments and descriptions are only preferred embodiments of the present invention, and are not intended to limit the invention, and that various changes and modifications may be made therein without departing from the spirit and scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (6)
1. The working condition prediction method based on the historical data search is characterized by comprising the following steps of:
step S1, inputting current working condition data;
s2, expanding current working condition data into working condition space vectors;
step S3, calculating the similarity between the current working condition space vector and the historical working condition space vector, and obtaining a calculation result;
s4, sorting the calculation results to form a candidate set, and returning a sorting result;
and S5, executing service operation according to the sequencing result.
2. The method of claim 1, further comprising reading historical operating condition data and expanding the historical operating condition data into a historical operating condition space vector.
3. The method for predicting the working condition based on the historical data search according to claim 1, wherein the historical working condition space vector is a full-historical time working condition space vector or a partial-historical time working condition space vector.
4. The method for predicting operating conditions based on historical data searching of claim 1, wherein,
executing the business operation according to the sorting result comprises the following steps:
taking a certain historical moment t in the returned sequencing result as a starting point, and loading the historical data in the n time periods after loading;
in the n time period, calculating the time difference from the t moment to the abnormal moment and the occurrence times of the abnormal moment in the n time period;
and summarizing the result of the last step, and counting the probability and time difference distribution of abnormal time occurrence, wherein the probability and time difference distribution are used for predicting whether alarm information is sent out.
5. A computing device, comprising:
one or more processors; and
a memory; and
one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors, the one or more programs including instructions for the historical data search based operating condition prediction method of any of claims 1-4.
6. A computer-readable storage medium having one or more programs stored thereon, characterized in that,
the one or more programs include instructions adapted to be loaded by a memory and to perform the historic data search based operating condition prediction method of any of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210016479.2A CN114492956B (en) | 2022-01-07 | 2022-01-07 | Working condition prediction method, equipment and storage medium based on historical data search |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210016479.2A CN114492956B (en) | 2022-01-07 | 2022-01-07 | Working condition prediction method, equipment and storage medium based on historical data search |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114492956A CN114492956A (en) | 2022-05-13 |
| CN114492956B true CN114492956B (en) | 2024-03-12 |
Family
ID=81509918
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210016479.2A Active CN114492956B (en) | 2022-01-07 | 2022-01-07 | Working condition prediction method, equipment and storage medium based on historical data search |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114492956B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116030900B (en) * | 2023-03-24 | 2023-06-16 | 安徽瑞邦数科科技服务有限公司 | Method, device, equipment and storage medium for controlling component content of chemical product |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102840571A (en) * | 2012-09-20 | 2012-12-26 | 贵州电力试验研究院 | Subspace identification based forecasting method for superheated steam output of boiler of firepower power station |
| CN103605734A (en) * | 2013-11-19 | 2014-02-26 | 广东电网公司电力科学研究院 | Characteristic vector based data transmission compression method and system |
| CN105760617A (en) * | 2016-03-07 | 2016-07-13 | 华北电力大学（保定） | Calculation method applied to multi-parameter fault prediction and judgment indexes of wind generating set |
| CN107169599A (en) * | 2017-05-12 | 2017-09-15 | 东北大学 | A kind of Multiobjective Optimal Operation method based on iron and steel enterprise's energy resource system |
| CN111521400A (en) * | 2020-06-24 | 2020-08-11 | 北京航空航天大学 | Bearing early fault diagnosis method based on EDM and spectral kurtosis |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012040916A1 (en) * | 2010-09-29 | 2012-04-05 | 东北大学 | Fault monitoring method of continuous annealing process based on recursive kernel principal component analysis |
-
2022
- 2022-01-07 CN CN202210016479.2A patent/CN114492956B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102840571A (en) * | 2012-09-20 | 2012-12-26 | 贵州电力试验研究院 | Subspace identification based forecasting method for superheated steam output of boiler of firepower power station |
| CN103605734A (en) * | 2013-11-19 | 2014-02-26 | 广东电网公司电力科学研究院 | Characteristic vector based data transmission compression method and system |
| CN105760617A (en) * | 2016-03-07 | 2016-07-13 | 华北电力大学（保定） | Calculation method applied to multi-parameter fault prediction and judgment indexes of wind generating set |
| CN107169599A (en) * | 2017-05-12 | 2017-09-15 | 东北大学 | A kind of Multiobjective Optimal Operation method based on iron and steel enterprise's energy resource system |
| CN111521400A (en) * | 2020-06-24 | 2020-08-11 | 北京航空航天大学 | Bearing early fault diagnosis method based on EDM and spectral kurtosis |
Non-Patent Citations (1)
| Title |
|---|
| 考虑风电出力和负荷不确定性的电-气合能源系统概率能流算法研究;靳康萌;中国优秀硕士学位论文全文数据库工程科技Ⅱ辑;第C038-2212页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114492956A (en) | 2022-05-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Gu et al. | A novel state-of-health estimation for the lithium-ion battery using a convolutional neural network and transformer model | |
| Zhou et al. | A model for real-time failure prognosis based on hidden Markov model and belief rule base | |
| EP2107439B1 (en) | Method and system for the design and implementation of optimal multivariable model predictive controllers for fast-sampling constrained dynamic systems | |
| Pani et al. | A survey of data treatment techniques for soft sensor design | |
| CN114492956B (en) | Working condition prediction method, equipment and storage medium based on historical data search | |
| CN113705839B (en) | Predictive maintenance method and maintenance system for fuel cell stack | |
| CN115315667A (en) | Process network with several plants | |
| Li et al. | Fault forecasting of a machining center tool magazine based on health assessment | |
| Lv et al. | Transformer based long-term prognostics for dynamic operating PEM fuel cells | |
| Liu et al. | Long–short-term memory encoder–decoder with regularized hidden dynamics for fault detection in industrial processes | |
| Qin et al. | Root cause analysis of industrial faults based on binary extreme gradient boosting and temporal causal discovery network | |
| Song et al. | Process monitoring via key principal components and local information based weights | |
| Lee et al. | Data-driven health condition and RUL prognosis for liquid filtration systems | |
| Zhang et al. | Soft sensor model development for cobalt oxalate synthesis process based on adaptive Gaussian mixture regression | |
| Wu et al. | Real-time anomaly detection in gas sensor streaming data | |
| Wang et al. | Efficient quality variable prediction of industrial process via fuzzy neural network with lightweight structure | |
| Sobaszek et al. | Survival analysis method as a tool for predicting machine failures | |
| Tian et al. | Fault diagnosis strategy for few shot industrial process based on data augmentation and depth information extraction | |
| Zhang et al. | Data-driven fault prognosis based on incomplete time slice dynamic Bayesian network | |
| Mora-López et al. | Binding statistical and machine learning models for short-term forecasting of global solar radiation | |
| Cao et al. | Research on Abnormal Diagnosis Method of Benzene-ethylene Ratio Double-layer Architecture Control System based on OHEncoding-XGBoost | |
| CN117829822B (en) | Power transformer fault early warning method and system | |
| Liu et al. | Intelligent Operational Optimization Method for Salt Lake Chemical Industrial Processes Based on Data-Driven Modeling | |
| Liu et al. | Research on Group Preventive Maintenance Method Based on Components Dependence | |
| Tian et al. | Quality Variable Prediction for Dynamic Process Based on Adaptive Principal Component Regression with Selective Integration of Multiple Local Models. |
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 | ||
| CB03 | Change of inventor or designer information |
Inventor after: Hou Jie Inventor after: Li Yuanpeng Inventor after: Zhang Xiaoming Inventor before: Hou Jie Inventor before: Li Yuanpeng |
|
| CB03 | Change of inventor or designer information |