CN105278527B - A kind of real-time performance evaluation method suitable for tobacco processing course single loop control system - Google Patents
A kind of real-time performance evaluation method suitable for tobacco processing course single loop control system Download PDFInfo
- Publication number
- CN105278527B CN105278527B CN201510864765.4A CN201510864765A CN105278527B CN 105278527 B CN105278527 B CN 105278527B CN 201510864765 A CN201510864765 A CN 201510864765A CN 105278527 B CN105278527 B CN 105278527B
- Authority
- CN
- China
- Prior art keywords
- data
- time
- real
- performance evaluation
- control system
- 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
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
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Manufacture Of Tobacco Products (AREA)
- Stored Programmes (AREA)
Abstract
The present invention relates to the correlative technology field of tobacco processing course control, specifically a kind of method of evaluating performance for tobacco processing course single loop control system, it is characterised in that：A certain single loop control system in selected tobacco work flow, by using the time series models for carrying out real-time identification peration data with the least square method for becoming forgetting factor, improves the accuracy of Model Distinguish as object to be evaluated；And in performance evaluation, real-time rolling analysis is carried out to peration data using the form of sliding window, the real-time of performance evaluation is improved.Using method proposed by the present invention, forgetting factor changes with the change of models fitting error.When larger from anticipation error, the appropriate value for reducing forgetting factor improves identification sensitivity；When smaller from anticipation error, the value of appropriate increase forgetting factor improves identification precision.By the real-time identification ability of this method, the performance evaluation level of tobacco processing course can be lifted.
Description
Technical field
It is specifically a kind of processed for tobacco the present invention relates to the correlative technology field of tobacco processing course control
The method of evaluating performance of journey single loop control system.
Background technology
Tobacco processing course is continuous flow procedure, after prolonged production run, and each control loop performance can not
With the decline of degree, how control performance to be evaluated using process creation data, the latent of control system performance is found in time
Degenerating, and then taking preventive and improved measures, improving process control level and production maintenance efficiency, be process control personnel urgently
The problem of solution.
Tobacco processing course process is numerous, and each operation includes many different types of control loops.Single loop is still it
In most basic, most common control loop.Harris passes through to side of the single loop routine operation data using time series analysis
Method, finds feedback invariant and as the minimum variance benchmark for evaluating control system（Can J Chem Eng, 1989(67):
856-861）.On this basis, various types of evaluation method and computational methods are also suggested evaluation control system
Energy.Although method of evaluating performance is a lot, minimum variance method of evaluating performance is because the procedural knowledge of needs is few, it is relatively simple to calculate
It is single, it is still present most popular evaluation method.
But this conventional minimum variance method of evaluating performance faces following subject matter in actual applications：One is this
Method is to utilize conventional discrimination method（Such as：Least square；Burg methods etc.）The time series models of peration data are carried out
Identification, the degree of accuracy of identification gained model is not high, have impact on the accuracy of performance evaluation；Two be due to that discrimination method is limited, and is led to
Be often that peration data is analyzed and evaluated in batches, lack to the real-time rolling analysis of creation data, performance evaluation it is real-time
Property is not high.
The content of the invention
The problem of purpose of the present invention exactly exists for above-mentioned existing minimum variance Performance Evaluation Technique, it is proposed that a kind of
It is adapted to the real-time performance evaluation method of tobacco processing course single loop control system.By using with the most young waiter in a wineshop or an inn for becoming forgetting factor
Multiplication carrys out the time series models of real-time identification peration data, improves the accuracy of Model Distinguish；And in performance evaluation, use
The form of sliding window carries out real-time rolling analysis to peration data, improves the real-time of performance evaluation.
The purpose of the present invention is achieved through the following technical solutions：
A kind of real-time performance evaluation method suitable for tobacco processing course single loop control system, specifically includes following step
Suddenly：
（1）For the particular process process in tobacco work flow, single loop control system to be evaluated is selected, phase is determined
The controlled variable answered（Peration data）；
（2）Determine collection rule：Including sampling time, evaluating data scope, initial samples scope.It is processed according to tobacco
The characteristics of journey, the sampling time is 5-15 seconds；Evaluating data scope chooses the data segment of stable production process operation（Removing material head,
Expect the creation data of tail and failure phase）；Initial samples scope is 100-200 data.
（3）Peration data to collection is pre-processed, including singular value is rejected and stationary test.If stationarity is examined
Test and do not pass through, then data can be carried out once or second order difference is handled, and peration data is subtracted into its setting value with being used as identification
Model data；
（4）Autoregression is carried out to Identification Data using with the least square method for becoming forgetting factor（AR）Time series models
Modeling；
（5）The variance of noise or residual error is obtained by " albefaction " by time series models above；Utilize priori
Or peration data, estimate the time delay d of control system；
（6）Binding time postpones d, asks for control feedback invariant from the time series models obtained using long division；
And combine noise variance, ask for the minimum variance under peration data minimum variance meaning；
（7）The realized variance of the output of process is asked for out of peration data sample range, so as to ask for evaluation, this is adopted
The Performance Evaluating Indexes of sample scope controlling system；
（8）Using the form of sliding window, rolling analysis is carried out to whole evaluating data scope, continuous performance is obtained and comments
Valency index.
Step（3）In, singular value, which is rejected, to be carried out manually to substantially not meeting the data of steady production process or faulty section
Reject；Stationary test uses non-parametric test method ADF；Data once or second order difference processing is such as following formula：
Step（4）In, autoregression is carried out to Identification Data using with the least square method for becoming forgetting factor（AR）Time
Series model, the autoregression model structure after identification is shown below：
The identification algorithm used for：
Wherein,For the output peration data at current time；For the estimation coefficient of autoregression model；For noise；For intermediate variable, initial value is set according to actual conditions；For forgetting factor, following fair curve is met：
Wherein,Respectively the minimum value and maximum of forgetting factor, set depending on concrete condition.
Step（5）In, varianceThe autoregression model " albefaction " obtained by identification is obtained, i.e., obtained by following formula：
Step（6）In, the minimum variance under peration data minimum variance meaningAsked for by following methods：
First, by time series models（AR）Expression formula is expressed as the moving average model of a unlimited item：
Wherein d postpones for control time.Then, d control feedback invariants before asking for, then under LMS control most
Small variance is：
Step（7）In, the realized variance of the output of processCalculated by following formula：
Wherein, n is the number of evaluating data.
Step（8）In, as a result of the mode of Dynamic Identification model, the data progressive in whole range of value
Rolling Calculation Analysis is carried out when can evaluate in the form of sliding window.
Using method proposed by the present invention, forgetting factor changes with the change of models fitting error.Missed when from expectation
When difference is larger, suitably reduce the value of forgetting factor to improve identification sensitivity；When smaller from anticipation error, appropriate increase is forgotten
The value of the factor improves identification precision.Also, the real-time identification ability based on this method, is adopted in performance evaluation to peration data
Analyzed in real time with the form of sliding window, so that the degree of accuracy and the real-time of performance evaluation are improved, lifting tobacco process
Performance evaluation level.
Brief description of the drawings
Fig. 1 is common single loop control system；
Fig. 2 is discrete adaptive wave filter；
Fig. 3 is minimum variance performance indications rolling calculation flow chart of the present invention（The figure is used as Figure of abstract）；
Fig. 4 is minimum variance performance rolling analysis figure of the embodiment SH93 moisture control loop using the present invention；
Fig. 5 is minimum variance performance batch analysis figure of the embodiment SH93 moisture control loop using former method.
Embodiment
The invention will be further described with reference to the accompanying drawings and examples：
With certain cigar mill SH93 type air flow type cut tobacco driers（Scale flow 5100kg/h before drying）Moisture control loop is evaluation object
（Control system is as shown in Figure 1）.Specific implementation steps are as follows（As shown in Figure 3）：
（1）Selected moisture control loop is evaluation object；Determine controlled variable（Peration data）To dry silk moisture content of outlet；
（2）After certain trade mark pipe tobacco is completely into cut-tobacco drier, after after system stable operation 5 minutes, take 10 minutes it is later
Outlet moisture in cut tobacco data proceed by rolling real-time performance evaluation, and the length of evaluating data is 120（Sampling time is 10
Second）, the scope of evaluating data is the data in 60 minutes.
（3）Data to collection are pre-processed, including singular value is rejected and stationary test.Eliminate 3 it is unusual
Data, carry out stationary test with ADF and pass through；And setting value 14.2% is subtracted with the data obtained as identification model data；
（4）Auto-regressive time series model modeling is carried out to Identification Data using with the least square method for becoming forgetting factor
（Sef-adapting filter is as shown in Figure 2）.Here forgetting factorMaximum and minimum value be taken as 0.95 and 0.1 respectively；Initially
Variable；；；
（5）The variance of noise or residual error is obtained by " albefaction " by time series models above；According to
Prior process knowledge, obtain procedures system time delay be；
（6）Binding time postpones d, asks for control feedback invariant from the time series models obtained using long division；
And combine noise variance, ask for minimum variance；
（7）The realized variance of the output of process is asked for from 117 current output peration datas, so that
Ask for evaluating the Performance Evaluating Indexes of the sample range controlling system；
（8）Repeat（4）-（7）Step, every 5 data, rolls and solves minimum variance Performance Evaluating Indexes（Fig. 4）.
Former minimum variance Evaluation results as shown in figure 5, due to no rolling calculation, in the range of whole evaluating data only
There are 3 evaluation indexes, thus the index change in the range of each evaluating data is short in understanding, the real-time of performance evaluation is not
It is high；With reference to real process control effect, there is the trend of performance degradation in the middle of batch, and former method fails to reflect this deterioration in time
Trend.As can be seen here, rolling analysis method proposed by the present invention improves the degree of accuracy and the real-time of performance evaluation, is conducive to cigarette
The Real-Time Evaluation of careless process control performance, improves the fault diagnosis and the level of IT application of process.
Claims (6)
1. a kind of real-time performance evaluation method suitable for tobacco processing course single loop control system, it is characterised in that：Specifically
Comprise the following steps：
(1) the particular process process in tobacco work flow is directed to, single loop control system to be evaluated is selected, it is determined that accordingly
Controlled variable is peration data；
(2) collection rule is determined：Including sampling time, evaluating data scope, initial samples scope, according to tobacco processing course
Feature, the sampling time is 5-15 seconds；Evaluating data scope chooses the data segment of stable production process operation, and initial samples scope is
100-200 data；
(3) peration data to collection is pre-processed, including singular value is rejected and stationary test, if stationary test is not
Pass through, then data can be carried out once or second order difference is handled, and peration data is subtracted into its setting value with being used as identification model
Use data；
(4) autoregression is carried out to Identification Data using with the least square method for becoming forgetting factor（AR）Time series models are built
Mould；Autoregression model structure after identification is shown below：
The identification algorithm used for：
Wherein,For the output peration data at current time；For the estimation coefficient of autoregression model；For noise；For intermediate variable, initial value is set according to actual conditions；For forgetting factor, following fair curve is met：
Wherein,Respectively the minimum value and maximum of forgetting factor, set depending on concrete condition；
(5) variance of noise or residual error is obtained by " albefaction " by time series models above；Utilize priori or behaviour
Make data, estimate the time delay d of control system；
(6) binding time delay d, control feedback invariant is asked for using long division from the time series models obtained；And tie
Close noise variance, ask for the minimum variance under peration data minimum variance meaning；
(7) realized variance of the output of process is asked for out of peration data sample range, so as to ask for evaluating the sampling model
Enclose the Performance Evaluating Indexes of controlling system；
(8) using the form of sliding window, rolling analysis is carried out to whole evaluating data scope, continuous performance evaluation is obtained and refers to
Mark.
2. the real-time performance evaluation method according to claim 1 suitable for tobacco processing course single loop control system,
It is characterized in that：Step（3）In, it is to enter pedestrian to the data for substantially not meeting steady production process or faulty section that singular value, which is rejected,
Work is rejected；Stationary test uses non-parametric test method ADF；Data once or second order difference processing is such as following formula：
。
3. the real-time performance evaluation method according to claim 1 suitable for tobacco processing course single loop control system,
It is characterized in that：Step（5）In, varianceThe autoregression model " albefaction " obtained by identification is obtained, i.e., obtained by following formula：
。
4. the real-time performance evaluation method according to claim 1 suitable for tobacco processing course single loop control system,
It is characterized in that：Step（6）In, the minimum variance under peration data minimum variance meaningAsked for by following methods：
First, by time series models（AR）Expression formula is expressed as the moving average model of a unlimited item：
Wherein d postpones for control time, then, d control feedback invariants before asking for, then the minimum side under LMS control
Difference is：
。
5. the real-time performance evaluation method according to claim 1 suitable for tobacco processing course single loop control system,
It is characterized in that：Step（7）In, the realized variance of the output of processCalculated by following formula：
Wherein, n is the number of evaluating data.
6. the real-time performance evaluation method according to claim 1 suitable for tobacco processing course single loop control system,
It is characterized in that：Step（8）In, as a result of the mode of Dynamic Identification model, the data in whole range of value are entered
Rolling Calculation Analysis is carried out during row performance evaluation in the form of sliding window.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510864765.4A CN105278527B (en) | 2015-12-01 | 2015-12-01 | A kind of real-time performance evaluation method suitable for tobacco processing course single loop control system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510864765.4A CN105278527B (en) | 2015-12-01 | 2015-12-01 | A kind of real-time performance evaluation method suitable for tobacco processing course single loop control system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105278527A CN105278527A (en) | 2016-01-27 |
| CN105278527B true CN105278527B (en) | 2017-10-10 |
Family
ID=55147707
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510864765.4A Active CN105278527B (en) | 2015-12-01 | 2015-12-01 | A kind of real-time performance evaluation method suitable for tobacco processing course single loop control system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105278527B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108121215B (en) * | 2017-09-12 | 2018-11-16 | 山东科技大学 | Process control loops method of evaluating performance and device based on full loop reconstruct emulation |
| CN107966976B (en) * | 2017-12-06 | 2019-07-02 | 中南大学 | A kind of baking silk moisture control loop performance evaluation of data-driven and adjustment system |
| CN109062182B (en) * | 2018-07-27 | 2020-09-18 | 东北大学秦皇岛分校 | Efficient fact evaluation method and device based on minimum evaluation window |
| CN109032117B (en) * | 2018-09-06 | 2021-04-06 | 华北电力大学（保定） | ARMA model-based single-loop control system performance evaluation method |
| CN109508895A (en) * | 2018-11-29 | 2019-03-22 | 福建中烟工业有限责任公司 | Control performance assessment device, method and the storage medium of tobacco cutting equipment |
| CN111983997B (en) * | 2020-08-31 | 2021-07-20 | 北京清大华亿科技有限公司 | Coupling analysis-based control loop performance monitoring method and system |
| CN115903727A (en) * | 2022-10-10 | 2023-04-04 | 乌海宝化万辰煤化工有限责任公司 | DCS control system-based PID control loop performance evaluation system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6459939B1 (en) * | 1999-06-29 | 2002-10-01 | Alan J. Hugo | Performance assessment of model predictive controllers |
| CN1656946A (en) * | 2004-12-31 | 2005-08-24 | 龙岩卷烟厂 | Tobacco shreds drying and moisture moving control method |
| CN101929994A (en) * | 2010-08-24 | 2010-12-29 | 昆明理工大学 | Time sequence model and method for predicting multi-phase mixing uniformity |
| CN103472723A (en) * | 2013-08-19 | 2013-12-25 | 上海交通大学 | Predictive control method and system based on multi-model generalized predictive controller |
| CN104199296A (en) * | 2014-08-15 | 2014-12-10 | 上海交通大学 | Linear regression performance evaluation method with forgetting factor |
-
2015
- 2015-12-01 CN CN201510864765.4A patent/CN105278527B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6459939B1 (en) * | 1999-06-29 | 2002-10-01 | Alan J. Hugo | Performance assessment of model predictive controllers |
| CN1656946A (en) * | 2004-12-31 | 2005-08-24 | 龙岩卷烟厂 | Tobacco shreds drying and moisture moving control method |
| CN101929994A (en) * | 2010-08-24 | 2010-12-29 | 昆明理工大学 | Time sequence model and method for predicting multi-phase mixing uniformity |
| CN103472723A (en) * | 2013-08-19 | 2013-12-25 | 上海交通大学 | Predictive control method and system based on multi-model generalized predictive controller |
| CN104199296A (en) * | 2014-08-15 | 2014-12-10 | 上海交通大学 | Linear regression performance evaluation method with forgetting factor |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105278527A (en) | 2016-01-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105278527B (en) | A kind of real-time performance evaluation method suitable for tobacco processing course single loop control system | |
| CN109674080B (en) | Tobacco leaf conditioning water adding amount prediction method, storage medium and terminal equipment | |
| CN105512799B (en) | Power system transient stability evaluation method based on mass online historical data | |
| CN101870076B (en) | Method for predicting service life of guide pair of numerical control machine on basis of performance degradation model | |
| KR102145407B1 (en) | Method and control device for controlling technical systems | |
| CN106649755B (en) | Threshold value self-adaptive setting abnormity detection method for multi-dimensional real-time power transformation equipment data | |
| CN105341985A (en) | Moisture content control method and system of inlet cut tobaccos of cut-tobacco drier | |
| CN105334738B (en) | A kind of method of evaluating performance suitable for tobacco processing course pid control circuit | |
| CN116237817B (en) | Five-axis linkage numerical control machine intelligent monitoring system based on Internet of things | |
| CN113341883B (en) | Method and equipment for predicting machine tool machining working hours | |
| CN109508895A (en) | Control performance assessment device, method and the storage medium of tobacco cutting equipment | |
| CN113567904A (en) | Method and system suitable for metering error of capacitive mutual inductor | |
| CN113639999A (en) | Fault diagnosis method, system, medium and equipment for rolling bearing under variable rotating speed | |
| CN114115393A (en) | Method for controlling moisture and temperature at outlet of cut tobacco dryer for sheet cut tobacco making line | |
| CN110543869A (en) | Ball screw service life prediction method and device, computer equipment and storage medium | |
| CN105029672A (en) | Precise characterization method for water at outlet of baked cut tobacco | |
| CN117473833A (en) | Method and device for predicting rolling force of rod and wire based on finite element and neural network | |
| CN103267826A (en) | Soft measurement method for online detection of gelatin concentration | |
| CN110037336B (en) | Method for predicting position of actuating mechanism of cigarette circumference control system | |
| CN116400639A (en) | PLC (programmable logic controller) collected data intelligent cleaning method and system | |
| CN113435106B (en) | Transition mode operation fault detection method and system | |
| CN104980442A (en) | Network intrusion detection method based on element sample sparse representation | |
| CN109426216A (en) | A kind of rectangle plane working depth recognition methods for CNC milling machine | |
| CN113807431A (en) | Intelligent spindle state evaluation method and system based on multi-source information fusion | |
| CN113469408A (en) | Running state trend prediction method and system for phase modulator |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |