CN108375908B - Bisphenol A crystallization process rolling optimization method based on system operation mode - Google Patents
Bisphenol A crystallization process rolling optimization method based on system operation mode Download PDFInfo
- Publication number
- CN108375908B CN108375908B CN201810319806.5A CN201810319806A CN108375908B CN 108375908 B CN108375908 B CN 108375908B CN 201810319806 A CN201810319806 A CN 201810319806A CN 108375908 B CN108375908 B CN 108375908B
- Authority
- CN
- China
- Prior art keywords
- data
- mode
- variable
- batch
- quality index
- 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
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- General Factory Administration (AREA)
- Feedback Control In General (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a rolling optimization method for a bisphenol A crystallization process based on a system operation mode, and belongs to the field of chemical process manufacturing. The rolling optimization method for the bisphenol A crystallization process is divided into four steps, wherein the first step is collection and pretreatment of process data, the second step is construction of a system operation mode, the third step is batch data alignment based on the mode, the fourth step is establishment of a quantitative relation between the mode and a quality index, optimal operation conditions are solved by using a rolling optimization strategy, a statistical analysis algorithm is used, the concept of the mode is introduced into optimization of the bisphenol A crystallization process, the traditional optimization method based on an original variable form is converted into an optimization method based on a variable implicit process operation mode, the defects of the traditional optimization method in data utilization are overcome, and then the operation conditions are corrected in time to adapt to complex working conditions through a rolling optimization strategy which is predicted and corrected.
Description
Technical Field
The invention relates to a rolling optimization method for a bisphenol A crystallization process based on a system operation mode, and belongs to the field of chemical process manufacturing.
Background
The bisphenol A crystallization process belongs to intermittent industrial production, has no stable working point in the whole production process, and has a plurality of operation stages and remarkable nonlinearity. The quality indexes (such as crystal purity, crystal product granularity, raw material conversion rate and the like) of the bisphenol A crystallization process are closely related to the set operation state or operation trajectory. In order to enable the bisphenol A crystallization process to be in the optimal operation condition and achieve the expected control variable index, the traditional optimization strategy is to establish a mechanism model of the process and optimize the operation condition.
However, bisphenol A is complex to crystallize, difficult and expensive to model accurately. The intermittent production mode also makes most of the optimization methods difficult to directly use. Therefore, finding an effective optimization method becomes a general requirement for enterprises.
Under the background of big data, the storage capacity of industrial operation data is continuously improved, and the data mining technology is continuously developed, so the invention provides a bisphenol A crystallization process rolling optimization method based on a mode by utilizing a large amount of data in the production process, introduces the concept of a system operation mode into the optimization of the bisphenol A crystallization process by utilizing a statistical analysis algorithm, converts the traditional optimization method based on process variables into the optimization method based on the system operation mode, makes up the deficiency of the traditional optimization method in data utilization, and then corrects the operation conditions in time to adapt to complex working conditions through a rolling optimization strategy of prediction and correction. The system running mode not only considers the effect of the set operation condition on the quality index, but also considers the effect of other process variables on the quality index under the operation condition, and the relation between the quality index and the operation condition can be more accurately established due to the fact that more process information is contained, so that the optimization effect is ensured, the product quality is stabilized, the economic benefit is improved, and the application of the data-based optimization technology in the bisphenol A crystallization process is promoted.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a bisphenol A crystallization process rolling optimization method based on a system operation mode, wherein the idea of optimizing by using the system operation mode is applied to the bisphenol A crystallization process, and a mode for revealing essential characteristic states of physics, chemistry and the like in the process is extracted from set operation trajectory data and process variable data of the process; then, a quantitative relation between the mode and the quality index is established, and the optimal operating condition is solved by utilizing a rolling optimization strategy.
The technical scheme adopted by the invention is as follows:
the rolling optimization method of the bisphenol A crystallization process is divided into four steps, wherein the first step is the collection and pretreatment of process data, the second step is the construction of a system operation mode, the third step is the alignment of batch data based on the mode, and the fourth step is the establishment of a quantitative relation between the mode and a quality index, and the optimal operation condition is solved by using a rolling optimization strategy; the method comprises the following specific steps:
step 1): process data collection and preprocessing
Acquiring set operation trajectory data, process variable data and quality index data of a bisphenol A crystallization process, and performing data preprocessing in a mode of eliminating abnormal data and complementing missing data and standardized data; the method comprises the steps of collecting set operation trajectory data of a bisphenol A crystallization process as temperature data, and process variable data of crystallization tank temperature, crystallization tank pressure, pipeline temperature and the like;
step 2): construction of system operation mode
(1) Performing variable selection on the data preprocessed in the step one, and selecting variables which contribute more to the quality index, so as to avoid the covering of secondary variable information on key variable information; the variable selection method can be a mutual information entropy method, a correlation coefficient method, an LASSO regularization method, an elastic network, an orthogonal signal correction method and the like;
(2) acquiring comprehensive characteristics of the process by using a variable extraction method to obtain a system operation mode P ═ P1,P2,…,Pk]Wherein
The method is a kth batch mode, a is the number of comprehensive variables, a is linearly independent, a is smaller than the data dimension before the variable extraction method, and E is the number of sampling points of a batch; the variable extraction method refers to the technologies of principal component analysis, probability principal component analysis, dynamic principal component analysis, multi-directional principal component analysis and the like;
step 3): and aligning the data of each batch by taking the mode similarity of the data of each batch as a standard, wherein a calculation formula of the similarity is as follows:
wherein d is the similarity between two patterns, the smaller the value of d, the more similar the two patterns are,
is a reference pattern as an alignment reference, F is the number of sampling points of the reference pattern,
is composed of
To middle
The w-th element of the composite variable, w ∈ {1,2, …, E },
is the kth batch data pattern
The v-th element of the composite variable, v ∈ {1,2, …, F }.
Step 4): establishing a quantitative relation between the mode and the quality index, and solving the optimal operating condition by using a rolling optimization strategy
(1) Setting the process quality index as Y ═ Y1,y2,…,yk]Obtaining system operation mode P ═ P by using existing modeling method1,P2,…,Pk]Quantitative relationship with process quality index Y:
Y＝ψ(P) (B)；
where ψ (-) is a dynamic characteristic function describing the process;
(2) combining formula (B) with the desired quality index y*Establishing a minimum variance objective function to solve the pattern correction of the k-th batch
(3) For the optimal mode
Performing reconstruction to obtain influence
To optimally set the operation trajectory
(4) In order to overcome dynamic control deviation and uncontrollable random interference, the optimal setting operation rail obtained through the stepsThread
When the method is put into practical application, the optimal set operation trajectory is acquired in real time
Process variable data of;
(5) setting some decision points in the process, obtaining corresponding system operation mode data by using a variable extraction method according to the obtained process variable data and the set operation trajectory data when the production process reaches each decision point, and complementing the system operation mode data behind the current decision point according to the change trend of the system operation mode of the historical batch to obtain
(6) Predicting the quality index in the current batch mode according to the formula (B)
When in use
Not less than y*Then, for the setting operation trajectory
Without correction, when
Less than y*Then according to formula (C) pair
And (6) correcting.
ΔS(i)k+1＝ψ(Δyk+1) (D)；
Wherein the content of the first and second substances,
is as followsi+1 decision point modified set operation trajectory,
setting an operational trajectory for the ith decision point, Δ S (i)k+1To set the operating trajectory modifier for the ith decision point,
the deviation between the predicted quality index and the expected quality index in the current batch mode is shown.
(7) The (k + 1) th batch production is finished, and the measured actual process data is processed by a variable extraction method to obtain Pk+1。
(8) And (5) repeating the steps (2) to (7), so that the operation condition is repeatedly predicted and corrected at each decision point of each batch and is optimized in a rolling mode.
In the step 1), the mode of rejecting abnormal data is a 3 sigma criterion, the mode of filling missing data is a mean filling method, and the selection of the comprehensive variable number a is based on the accumulated contribution rate (for example, 80%).
The invention provides a rolling optimization method for a bisphenol A crystallization process based on a mode, which utilizes a statistical analysis algorithm to introduce the concept of the mode into the optimization of the bisphenol A crystallization process, converts the traditional optimization method based on an original variable form into an optimization method based on a process operation mode with hidden variables, makes up the deficiency of the traditional optimization method in data utilization, and then corrects the operation conditions in time to adapt to complex working conditions through a rolling optimization strategy of prediction and correction. The mode not only considers the action of the set operation condition on the quality index, but also considers the action of other process variables on the quality index under the operation condition, so that the mode containing more process information can more accurately establish the relation between the quality index and the operation condition, ensure the optimization effect, stabilize the product quality, improve the economic benefit and promote the application of the optimization technology in the bisphenol A crystallization process.
Drawings
FIG. 1 is a block flow diagram of an embodiment of the present invention;
FIG. 2 is a graph of the temperature operating variables of the present invention;
FIG. 3 is a process variable diagram of the present invention;
FIG. 4 is a graph of the cumulative contribution of the present invention;
FIG. 5 is a schematic view of the present invention;
FIG. 6 is a schematic diagram of the operational trajectory scrolling optimization of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention provides a rolling optimization method for a bisphenol A crystallization process, which is shown in figures 1-6 and comprises four steps, wherein the first step is collection and pretreatment of process data, the second step is construction of a system operation mode, the third step is batch data alignment based on a mode, the fourth step is establishment of a quantitative relation between the mode and a quality index, and optimal operation conditions are solved by using a rolling optimization strategy; the method comprises the following specific steps:
the method comprises the following steps: process data collection and preprocessing
Acquiring set operation trajectory data, process variable data and quality index data of a bisphenol A crystallization process, and performing data preprocessing in a mode of eliminating abnormal data and complementing missing data and standardized data; the method comprises the steps of collecting set operation trajectory data of a bisphenol A crystallization process as temperature data, and process variable data of crystallization tank temperature, crystallization tank pressure, pipeline temperature and the like;
step two: construction of system operation mode
(1) Performing variable selection on the data preprocessed in the step one, and selecting variables which contribute more to the quality index, so as to avoid the covering of secondary variable information on key variable information; the variable selection method can be a mutual information entropy method, a correlation coefficient method, an LASSO regularization method, an elastic network, an orthogonal signal correction method and the like;
(2) obtaining the comprehensive characteristics of the process by using an extraction method to obtain a system operation mode P ═ P1,P2,…,Pk]Wherein
The method is a kth batch mode, a is the number of comprehensive variables, a is linearly independent, a is smaller than the data dimension before the variable extraction method, and E is the number of sampling points of a batch; the variable extraction method refers to the technologies of principal component analysis, probability principal component analysis, dynamic principal component analysis, multi-directional principal component analysis and the like;
step three: and aligning the data of each batch by taking the mode similarity of the data of each batch as a standard, wherein a calculation formula of the similarity is as follows:
wherein d is the similarity between two modes, the smaller the value of d, the more similar the two modes are,
is a reference pattern as an alignment reference, F is the number of sampling points of the reference pattern,
is composed of
To middle
The w-th element of the composite variable, w ∈ {1,2, …, E },
is the kth batch data pattern
The v-th element of the composite variable, v ∈ {1,2, …, F }.
Step four: establishing a quantitative relation between the mode and the quality index, and solving the optimal operating condition by using a rolling optimization strategy
(1) Setting the process quality index as Y ═ Y1,y2,…,yk]Obtaining system operation mode P ═ P by using existing modeling method1,P2,…,Pk]Quantitative relationship with process quality index Y:
Y＝ψ(P) (B)；
where ψ (-) is a dynamic characteristic function describing the process;
(2) combining formula (B) with the desired quality index y*Establishing a minimum variance objective function to solve the pattern correction of the k-th batch
(3) For the optimal mode
Performing reconstruction to obtain influence
To optimally set the operation trajectory
(4) In order to overcome dynamic control deviation and uncontrollable random interference, the optimal setting operation trajectory obtained through the steps
When the method is put into practical application, the optimal set operation trajectory is acquired in real time
Process variable data of;
(5) setting some decision points in the process, obtaining corresponding system operation mode data by using a variable extraction method according to the obtained process variable data and the set operation trajectory data when the production process reaches each decision point, and complementing the system operation mode data behind the current decision point according to the change trend of the system operation mode of the historical batch to obtain
(6) Predicting the quality index in the current batch mode according to the formula (B)
When in use
Not less than y*Then to the setting operation trajectory
Without correction, when
Less than y*Then according to formula (3) to
The correction is carried out so that the correction is carried out,
ΔS(i)k+1＝ψ(Δyk+1) (D)；
wherein the content of the first and second substances,
the set operation trajectory for the i +1 th decision point modification,
setting an operational trajectory for the ith decision point, Δ S (i)k+1To be aligned withThe i decision points set the operating trajectory correction,
the deviation between the predicted quality index and the expected quality index in the current batch mode is shown.
(7) The (k + 1) th batch production is finished, and the measured actual process data is processed by a variable extraction method to obtain Pk+1。
(8) And (5) repeating the steps (2) to (7), so that the operation condition is repeatedly predicted and corrected at each decision point of each batch and is optimized in a rolling mode.
As shown in fig. 1, the specific implementation steps of the present invention are as follows:
step 1: in the bisphenol A crystallization process, the temperature is set operation data, the track is shown in figure 2, 20 variables such as the temperature of a crystallization tank, the pressure of the crystallization tank, the temperature of a pipeline and the like are selected as process variables related to quality indexes, part of variable trajectories are shown in figure 3, the conversion rate of raw materials is quality index data, a total of 400 batches of sample data are collected, and the expected conversion rate of the raw materials is 90%. And (3) according to a Lauda criterion, respectively carrying out abnormal value elimination pretreatment on the sample data set in the step (1) according to a 3 sigma criterion, filling missing values generated after elimination according to a mean filling method, and carrying out standardization treatment on quality index data.
Step 2: expanding the operation data and 20 process variable data set in the step 1 according to the variable direction, carrying out standardization processing, then carrying out principal component analysis to obtain 21 comprehensive variables,
and step 3: the cumulative contribution rate is calculated, as shown in FIG. 4, dimension 2 of the selected pattern.
And 4, step 4: the pattern of the samples obtained in batches according to step 3 is shown in FIG. 5, according to
Calculating the kth batch data pattern and the reference pattern
And the similarity between the two sets of the model data is aligned with the model data of each batch at the same stage.
And 5: a linear relation model between the system operation mode and the quality index (raw material conversion rate) of 400 batches of data is established by adopting a partial least square algorithm.
Step 6: establishing a minimum variance objective function between the expected quality index and the predicted quality index in combination with the step 5, and solving the mode correction quantity
k is 401,402, …, N, and the optimal mode is further determined
Reconstructing the original variable form to obtain the optimal temperature operation data
And 7: temperature operating data to implement step 6 solution
And collects the process variable data under the set value in real time.
And 8: 3 decision points are set in the whole production process, mode data are obtained by subjecting the obtained process variable data and the set temperature data to principal component analysis projection at each decision point, and then unknown mode data of the batch are complemented according to the mode change trend of historical batch data to obtain
And step 9: predict feedstock conversion in this mode according to step 5
When the predicted value is not less than 90% of the expected conversion rate of the raw materials, the temperature operation curve is not corrected until the next decision point, if the predicted value is less than 90%, the deviation between the predicted value and the expected conversion rate of the raw materials is calculated
Correction quantity delta S (i) of operation trajectoryk+1＝ψ(Δyk+1) Finally according to the formula
And obtaining an optimal operation curve.
Step 10: when the current batch production is finished, the measured actual process data is processed to obtain P by a variable extraction methodk+1
Step 11: steps 6-10 are repeated, thus iteratively predicting and modifying, at each decision point of each batch, the operating conditions for the rolling optimization, as shown in FIG. 6.
Claims (6)
1. A rolling optimization method for a bisphenol A crystallization process based on a system operation mode comprises the following four steps:
1) collecting and preprocessing process data;
2) constructing a system operation mode;
3) batch data alignment based on run mode;
4) establishing a quantitative relation between the operation mode and the quality index, and solving an optimal operation condition by using a rolling optimization strategy;
the specific method of the step 3) is as follows: and aligning the data of each batch by taking the mode similarity of the data of each batch as a standard, wherein a calculation formula of the similarity is as follows:
wherein d is the similarity between the two operating modes,
is a reference pattern as an alignment reference, F is the number of sampling points of the reference pattern,
is composed of
To middle
The w-th element of the composite variable, w ∈ {1,2, …, E },
is the kth batch data pattern
The v-th element of the composite variable, v ∈ {1,2, …, F }; RF×aA real number vector set;
the specific method of the step 2) is as follows:
(1) performing variable selection on the data preprocessed in the step 1), selecting variables which greatly contribute to the quality index, and avoiding the covering of secondary variable information on key variable information;
(2) acquiring comprehensive characteristics of the process by using a variable extraction method to obtain a system operation mode P ═ P1,P2,…,Pk]Wherein
The method is a kth batch mode, a is the number of comprehensive variables, a is linearly independent, a is smaller than the data dimension before the variable extraction method, and E is the number of sampling points of a batch;
the specific method of the step 4) is as follows: establishing a quantitative relation between the operation mode and the quality index, and solving an optimal operation condition by using a rolling optimization strategy:
(1) setting the process quality index as Y ═ Y1,y2,…,yk]Obtaining system operation mode P ═ P by using existing modeling method1,P2,…,Pk]Quantitative relationship with process quality index Y:
where ψ (-) is a dynamic characteristic function describing the process;
Y＝ψ(P) (B)；
(2) combining formula (B) with the desired massMark y*Establishing a minimum variance objective function to solve the pattern correction of the k-th batch
(3) For the optimal mode
Performing reconstruction to obtain influence
To optimally set the operation trajectory
(4) Optimally setting an operating trajectory
When the method is put into practical application, the optimal set operation trajectory is acquired in real time
Process variable data of;
(5) setting some decision points in the process, obtaining corresponding system operation mode data by using a variable extraction method according to the obtained process variable data and the set operation trajectory data when the production process reaches each decision point, and complementing the system operation mode data behind the current decision point according to the change trend of the system operation mode of the historical batch to obtain
(6) Predicting the quality index in the current batch mode according to the formula (B)
When in use
Not less than y*Then, for the setting operation trajectory
Without correction, when
Less than y*Then according to formula (C) pair
Correcting;
ΔS(i)k+1＝ψ(Δyk+1) (D)；
wherein the content of the first and second substances,
the set operation trajectory for the i +1 th decision point modification,
setting an operational trajectory for the ith decision point, Δ S (i)k+1To set the operating trajectory modifier for the ith decision point,
the deviation between the predicted quality index and the expected quality index in the current batch mode;
(7) the (k + 1) th batch production is finished, and the measured actual process data is processed by a variable extraction method to obtain Pk+1；
(8) And (5) repeating the steps (2) to (7), so that the operation condition is repeatedly predicted and corrected at each decision point of each batch and is optimized in a rolling mode.
2. The method of claim 1, wherein the specific method in step 1) is: acquiring set operation trajectory data, process variable data and quality index data of a bisphenol A crystallization process, and performing data preprocessing in a mode of eliminating abnormal data and complementing missing data and standardized data; the method comprises the steps of collecting bisphenol A crystallization process set operation trajectory data as temperature data, and collecting process variable data as crystallization tank temperature, crystallization tank pressure and pipeline temperature.
3. The method of claim 1, further comprising: in the step 1), the mode of eliminating abnormal data is a 3 sigma criterion, and the mode of filling missing data is a mean value filling method.
4. The method of claim 1, further comprising: the selection of the comprehensive variable number a in the step 2) is based on the accumulated contribution rate.
5. The method of claim 1, wherein: in the step 2), a mutual information entropy method, a correlation coefficient method, an LASSO regularization method, an elastic network or an orthogonal signal correction method is adopted during variable selection.
6. The method according to any one of claims 3-5, wherein: in the step 2), the variable extraction method is a principal component analysis method, a probability principal component analysis method, a dynamic principal component analysis method or a multidirectional principal component analysis method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810319806.5A CN108375908B (en) | 2018-04-11 | 2018-04-11 | Bisphenol A crystallization process rolling optimization method based on system operation mode |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810319806.5A CN108375908B (en) | 2018-04-11 | 2018-04-11 | Bisphenol A crystallization process rolling optimization method based on system operation mode |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108375908A CN108375908A (en) | 2018-08-07 |
| CN108375908B true CN108375908B (en) | 2020-07-03 |
Family
ID=63032055
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810319806.5A Active CN108375908B (en) | 2018-04-11 | 2018-04-11 | Bisphenol A crystallization process rolling optimization method based on system operation mode |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108375908B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20000018385A (en) * | 1998-09-01 | 2000-04-06 | 이광순 | System and method for model based predictive control of batch process integrating quality control function |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9046882B2 (en) * | 2010-06-30 | 2015-06-02 | Rockwell Automation Technologies, Inc. | Nonlinear model predictive control of a batch reaction system |
| CN103760773A (en) * | 2014-01-22 | 2014-04-30 | 杭州电子科技大学 | Batch process PI-PD control method for state space model prediction control optimization |
| CN105334831B (en) * | 2015-11-24 | 2017-10-24 | 江南大学 | The double dimension method for on-line optimization of batch process driven by data difference |
| CN105373098B (en) * | 2015-11-26 | 2017-10-24 | 江南大学 | A kind of online rolling optimization method of batch process model-free decomposed based on the variable period |
| CN105488530A (en) * | 2015-11-30 | 2016-04-13 | 广西大学 | Method for controlling sugar boiling crystallization process based on sequential extreme learning machine |
-
2018
- 2018-04-11 CN CN201810319806.5A patent/CN108375908B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20000018385A (en) * | 1998-09-01 | 2000-04-06 | 이광순 | System and method for model based predictive control of batch process integrating quality control function |
Non-Patent Citations (1)
| Title |
|---|
| Constrained latent variable model predictive control for trajectory tracking and economic optimization in batch processes;J.L Godoy,等;《Journal of Process Control》;20161231;第45卷;第1-11页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108375908A (en) | 2018-08-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108762228B (en) | Distributed PCA-based multi-working-condition fault monitoring method | |
| CN108897286B (en) | Fault detection method based on distributed nonlinear dynamic relation model | |
| CN108375965B (en) | non-Gaussian process monitoring method based on multi-variable block cross correlation elimination | |
| CN108345284B (en) | Quality-related fault detection method based on two variable blocks | |
| CN108388234B (en) | Fault monitoring method based on relevance division multi-variable block PCA model | |
| CN107403196B (en) | Method for predicting concentration of butane at bottom of debutanizer by instant learning modeling based on spectral clustering analysis | |
| CN105607631B (en) | The weak fault model control limit method for building up of batch process and weak fault monitoring method | |
| CN111444963B (en) | SSA-SVR model-based blast furnace molten iron silicon content prediction method | |
| WO2020124779A1 (en) | Working condition state modeling and model correction method | |
| CN108445867B (en) | non-Gaussian process monitoring method based on distributed ICR model | |
| CN112904810B (en) | Process industry nonlinear process monitoring method based on effective feature selection | |
| CN111709181B (en) | Method for predicting fault of polyester filament yarn industrial production process based on principal component analysis | |
| CN111259523B (en) | Process monitoring method based on KPCA-CVA model and random algorithm | |
| CN116821832A (en) | Abnormal data identification and correction method for high-voltage industrial and commercial user power load | |
| CN109921462B (en) | New energy consumption capability assessment method and system based on LSTM | |
| CN108375908B (en) | Bisphenol A crystallization process rolling optimization method based on system operation mode | |
| CN108427398B (en) | Dynamic process monitoring method based on distributed AR-PLS model | |
| CN116341770B (en) | Production capacity prediction method and system based on polystyrene production data | |
| CN112215410A (en) | Power load prediction method based on improved deep learning | |
| CN112700050A (en) | Method and system for predicting ultra-short-term 1 st point power of photovoltaic power station | |
| CN117349595A (en) | Industrial scene-oriented extremely few sample fault diagnosis method | |
| CN112418522A (en) | Industrial heating furnace steel temperature prediction method based on three-branch integrated prediction model | |
| He et al. | Research on hybrid ARIMA and support vector machine model in short term load forecasting | |
| CN111160811A (en) | Batch process fault monitoring method based on multi-stage FOM-SAE | |
| CN115374991A (en) | Improved statistics upscale-based photovoltaic cluster output prediction method |
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 |