CN102889598A - Control method for assisting stable combustion of garbage by predicting garbage calorific value - Google Patents
Control method for assisting stable combustion of garbage by predicting garbage calorific value Download PDFInfo
- Publication number
- CN102889598A CN102889598A CN2012104118041A CN201210411804A CN102889598A CN 102889598 A CN102889598 A CN 102889598A CN 2012104118041 A CN2012104118041 A CN 2012104118041A CN 201210411804 A CN201210411804 A CN 201210411804A CN 102889598 A CN102889598 A CN 102889598A
- Authority
- CN
- China
- Prior art keywords
- value
- garbage
- parameter
- control method
- genetic algorithm
- 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.)
- Pending
Links
Images
Landscapes
- Incineration Of Waste (AREA)
Abstract
The invention discloses a control method for assisting stable combustion of garbage by predicting a garbage calorific value. The control method comprises the following steps of: converting operational parameter signals into input parameters from a garbage incineration control system; screening the parameters which are relatively close to the garbage calorific value by using a Garson algorithm; performing dimensionality reduction treatment on the acquired parameters through principal constituents to remove redundant information; establishing genetic algorithm optimized neural network model to read and process data which is acquired through the dimensionality reduction so as to train, and outputting a predicted garbage colorific value; and performing signal conversion on the garbage colorific value, and then feeding the garbage colorific value back to a controller to guarantee that the garbage incineration control system can adjust relevant parameters according to commands of the controller and eliminate adverse effects caused by fluctuation of the colorific value. By the control method, online prediction of the garbage colorific value can be realized, corresponding prediction accuracy can be improved, and hysteretic quality caused by measurement of the conventional method or large deviation caused by simple neural network prediction is avoided.
Description
Technical field
The present invention relates to the boiler combustion monitoring field of Thermal Power Engineering, particularly a kind of control method of utilizing the auxiliary rubbish smooth combustion of refuse thermal value prediction.
Background technology
The variation that enters the stove refuse thermal value makes a big impact to the stable operation meeting of incinerator.Present waste incineration control is all take fuel quantity and air distribution as the control and regulation parameter, and the fluctuation of refuse thermal value and overall operation parameter are closely related, singly from fuel quantity and the very difficult associated adjustment of carrying out exactly system of air distribution.And the calculating of conventional garbage calorific value and prediction mainly be the physics that consists of by rubbish as the input of prediction, adopt the methods such as neutral net, linear regression to predict refuse thermal value.Yet on-the-spot the operation is difficult to quantitatively component of refuse be analyzed, if by artificial supposition or replacement table sample analysis, because the unstability of component of refuse all can cause relatively large deviation, be difficult to realize on-line measurement.
Utilize the on-line measurement data of incinerator as input parameter, can avoid the off-line measurement of component of refuse.The operational factor of refuse thermal value and incinerator is the coupled relation of nonlinearity, is difficult to judge exactly which parameter and rubbish caloric value contact more tight.But all relevant parameters as the prediction input, can be caused institute's established model structure complicated, affect its training time and effect, this can cause certain hysteresis quality to on-line prediction.If only rule of thumb choose several exemplary parameter as input parameter, tend to cause the degree of accuracy that predicts the outcome not high.
Summary of the invention
The object of the invention is to overcome incinerator and enter stove rubbish and be difficult to realize online and comparatively present situation and the shortcoming of Measurement accuracy, a kind of control method of utilizing the auxiliary rubbish smooth combustion of refuse thermal value prediction is provided.
The present invention is achieved through the following technical solutions:
A kind of control method of utilizing the auxiliary rubbish smooth combustion of refuse thermal value prediction comprises the steps:
(1) utilize the Garson method to filter out from the parameter of Control System of Incinerator data streams read, this parameter comprise garbage treatment quantity (t), fire box temperature (℃), combustion chamber draft (Pa), superheater inlet temperature (℃), the economizer exit temperature (℃), main steam flow (t/d), main steam pressure (MPa), main steam temperature (℃), a wind flow (km
3/ h), secondary air flow (km
3/ h), one-level desuperheating water flow (t/d), secondary desuperheating water flow (t/d), feed temperature (℃), one-level steam air heater steam inlet flow (t/d), secondary steam air heater steam inlet flow (t/d); According to the number of parameter that Control System of Incinerator reads, the hidden layer number of selection is set up neutral net and is trained, until reach training error; Read the weights of each articulamentum of neutral net, the connection weight that the connection weight of each hidden layer, output layer node is assigned to hidden layer, input layer gets on; Selected parameter is carried out principal component analysis, remove redundancy, reduce the dimension of input parameter, accelerate predetermined speed; Choose contribution rate of accumulative total〉85% characteristic value, try to achieve corresponding principal component value, at last this signal component value is represented original input information;
(2) set up three-layer neural network, determine its structure according to step () gained input parameter, and utilize threshold value and the weights of Genetic Algorithm Optimized Neural Network, be absorbed in local minimum when preventing neural metwork training, its process is as follows:
(1) according to the dimension of input parameter, hidden layer node number and output vector dimension are determined the chromosome length of genetic algorithm;
(2) the definition fitness function is the inverse of quadratic sum of the difference of actual value and predicted value, genetic algorithm according to this function select, crossover and mutation carries out iteration, seeks its minimum of a value;
Corresponding weights and threshold value were assigned to neutral net and train when (3) genetic algorithm was minimum of a value, if neutral net still can't reach the desired value of error, then returned to step (1) and carried out iteration.
Above-mentioned steps (two) genetic algorithm specifically adopts the real coding form, is provided with the real coding population of n individuality, its chromosome length L=S
1* R+R*S
2+ R+S
2, wherein R is BP neutral net input dimension of network, S
1, S
2Be respectively the number of hidden nodes and output vector dimension;
The definition fitness function
T wherein
r, t
oBe respectively output actual value and training output valve.
The present invention adopts the principal component analysis rule can introduce more multi-parameter, afterwards as the input of BP neutral net, both can avoid too much parameter input by the conversion dimensionality reduction, also can improve the precision of prediction of institute's established model, reduce predicted time, realize the on-line measurement of refuse thermal value.The refuse thermal value that incinerator obtains according to on-line prediction can be used as auxiliary its smooth combustion of feedback signal.
The present invention both can realize the on-line prediction of refuse thermal value, also can improve corresponding precision of prediction, had avoided the hysteresis quality that adopts the conventional method measurement to cause, the relatively large deviation of perhaps utilizing simple neural network prediction to cause.
Description of drawings
Fig. 1 is the flow chart of control method of the present invention
Fig. 2 is that refuse thermal value of the present invention is predicted being seen neural network model
Fig. 3 is the particular flow sheet that the present invention utilizes Genetic Algorithm Optimized Neural Network prediction refuse thermal value
Fig. 4 is the regression result figure of institute of the present invention established model training sample
Fig. 5 is model of the present invention and other model prediction result's relative error absolute value figure
The specific embodiment
Below in conjunction with specific embodiment the present invention is done further concrete detailed description the in detail.
Embodiment
Such as Fig. 1, Fig. 2, shown in Figure 3, the present invention utilizes the control method of the auxiliary rubbish smooth combustion of refuse thermal value prediction, it is characterized in that, comprises the steps:
(1) utilize the Garson method to filter out from the parameter of Control System of Incinerator data streams read, this parameter comprise garbage treatment quantity (t), fire box temperature (℃), combustion chamber draft (Pa), superheater inlet temperature (℃), the economizer exit temperature (℃), main steam flow (t/d), main steam pressure (MPa), main steam temperature (℃), a wind flow (km
3/ h), secondary air flow (km
3/ h), one-level desuperheating water flow (t/d), secondary desuperheating water flow (t/d), feed temperature (℃), one-level steam air heater steam inlet flow (t/d), secondary steam air heater steam inlet flow (t/d); Analyze selected parameter, if stable combustion condition mainly is reflected in main steam temperature, flow and pressure, the flow of desuperheating water also can be followed the variation of main steam condition and be done relevant the adjustment; The increase of refuse thermal value, the flue-gas temperature of burner hearth can increase, and is the direct reflection that refuse thermal value changes; The variation of air quantity can cause fire box temperature to produce fluctuation, can indirectly embody the variation of refuse thermal value.This shows that parameters obtained follows actual conditions to coincide, it can be represented initial data as input.
According to the number of parameter that Control System of Incinerator reads, the hidden layer number of selection is set up neutral net and is trained, until reach training error; Read the weights of each articulamentum of neutral net, the connection weight that the connection weight of each hidden layer, output layer node is assigned to hidden layer, input layer gets on;
Selected parameter is carried out principal component analysis, remove redundancy, reduce the dimension of input parameter, accelerate predetermined speed; Choose contribution rate of accumulative total〉85% characteristic value, try to achieve corresponding principal component value, at last this signal component value is represented original input information;
(2) set up three-layer neural network, determine its structure according to step () gained input parameter, and utilize threshold value and the weights of Genetic Algorithm Optimized Neural Network, be absorbed in local minimum when preventing neural metwork training, its process is as follows:
(1) according to the dimension of input parameter, hidden layer node number and output vector dimension are determined the chromosome length of genetic algorithm;
(2) the definition fitness function is the inverse of quadratic sum of the difference of actual value and predicted value, genetic algorithm according to this function select, crossover and mutation carries out iteration, seeks its minimum of a value;
Corresponding weights and threshold value were assigned to neutral net and train when (3) genetic algorithm was minimum of a value, if neutral net still can't reach the desired value of error, then returned to step (1) and carried out iteration.
Described step (two) genetic algorithm specifically adopts the real coding form, is provided with the real coding population of n individuality, its chromosome length L=S
1* R+R*S
2+ R+S
2, wherein R is BP neutral net input dimension of network, S
1, S
2Be respectively the number of hidden nodes and output vector dimension;
The definition fitness function
T wherein
r, t
oBe respectively output actual value and training output valve.Setting initial population is 50, and genetic algebra is 100, and by selecting of given fitness function f (x), crossover and mutation carries out iteration.When chromosomal average fitness tends towards stability, just can obtain optimum initial weight and the threshold value of network after the decoding, and it is given to network trains, if still can't reaching target error, network then repeats this process, until network convergence.Its flow chart is seen Fig. 3.
The training of refuse thermal value on-line measurement model and predict as follows:
Gather 50 groups of data in certain Control System of Incinerator, wherein select 40 groups as training sample, 10 groups is test samples.The parameter of part sample sees Table 1.The training sample fan-in network is trained, and in order to check the height of training quality, as abscissa, trained values is as ordinate with measured value.With least square method one-variable linear regression method the training the data obtained is carried out linear regression, get straight line shown in Figure 4.The best linear fit result of training result is substantially identical with the ideal curve of " measured value=trained values ".Carry out simultaneously correlation analysis and get coefficient R=0.99961, show that training quality is higher, institute's established model is suitable for the prediction into stove rubbish.
The relative error absolute value of 10 groups of test samples prediction data is seen Fig. 5.The present invention sets up the PCA-GA-BP model relative error minimum that enters stove rubbish control system for incinerator as can be seen from Figure 5, fluctuation range is also minimum, compare the as a result precision that other model obtains higher, therefore this model can effectively provide reliable refuse thermal value feedback signal for incinerator.
Table 1
The PCA-GA-BP forecast model that the present invention is set up embeds in the waste incineration control system, is this model conversion the input signal of waste incineration control system, makes its concrete height according to refuse thermal value make corresponding adjustment.Higher such as refuse thermal value, then should reduce the total amount into stove rubbish, reduce primary air flow, improve the concrete measures such as secondary air flow; If refuse thermal value is on the low side, can be corresponding make opposite operation, auxiliary incinerator stable operation.
As mentioned above, just can realize preferably the present invention.
Above-described embodiment only is the better embodiment of the present invention; but embodiments of the present invention are not restricted to the described embodiments; other are any not to deviate from change, the modification done under Spirit Essence of the present invention and the principle, substitute, combination, simplify; all should be the substitute mode of equivalence, be included within protection scope of the present invention.
Claims (2)
1. a control method of utilizing the auxiliary rubbish smooth combustion of refuse thermal value prediction is characterized in that, comprises the steps:
(1) utilize the Garson method to filter out from the parameter of Control System of Incinerator data streams read, this parameter comprise garbage treatment quantity (t), fire box temperature (℃), combustion chamber draft (Pa), superheater inlet temperature (℃), the economizer exit temperature (℃), main steam flow (t/d), main steam pressure (MPa), main steam temperature (℃), a wind flow (km
3/ h), secondary air flow (km
3/ h), one-level desuperheating water flow (t/d), secondary desuperheating water flow (t/d), feed temperature (℃), one-level steam air heater steam inlet flow (t/d), secondary steam air heater steam inlet flow (t/d); According to the number of parameter that Control System of Incinerator reads, the hidden layer number of selection is set up neutral net and is trained, until reach training error; Read the weights of each articulamentum of neutral net, the connection weight that the connection weight of each hidden layer, output layer node is assigned to hidden layer, input layer gets on; Selected parameter is carried out principal component analysis, remove redundancy, reduce the dimension of input parameter, accelerate predetermined speed; Choose contribution rate of accumulative total〉85% characteristic value, try to achieve corresponding principal component value, at last this signal component value is represented original input information;
(2) set up three-layer neural network, determine its structure according to step () gained input parameter, and utilize threshold value and the weights of Genetic Algorithm Optimized Neural Network, be absorbed in local minimum when preventing neural metwork training, its process is as follows:
(1) according to the dimension of input parameter, hidden layer node number and output vector dimension are determined the chromosome length of genetic algorithm;
(2) the definition fitness function is the inverse of quadratic sum of the difference of actual value and predicted value, genetic algorithm according to this function select, crossover and mutation carries out iteration, seeks its minimum of a value;
Corresponding weights and threshold value were assigned to neutral net and train when (3) genetic algorithm was minimum of a value, if neutral net still can't reach the desired value of error, then returned to step (1) and carried out iteration.
2. the control method of the auxiliary rubbish smooth combustion of refuse thermal value prediction of utilizing according to claim 1 is characterized in that:
Described step (two) genetic algorithm specifically adopts the real coding form, is provided with the real coding population of n individuality, its chromosome length L=S
1* R+R*S
2+ R+S
2, wherein R is BP neutral net input dimension of network, S
1, S
2Be respectively the number of hidden nodes and output vector dimension;
The definition fitness function
T wherein
r, t
oBe respectively output actual value and training output valve.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012104118041A CN102889598A (en) | 2012-10-24 | 2012-10-24 | Control method for assisting stable combustion of garbage by predicting garbage calorific value |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012104118041A CN102889598A (en) | 2012-10-24 | 2012-10-24 | Control method for assisting stable combustion of garbage by predicting garbage calorific value |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102889598A true CN102889598A (en) | 2013-01-23 |
Family
ID=47533201
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012104118041A Pending CN102889598A (en) | 2012-10-24 | 2012-10-24 | Control method for assisting stable combustion of garbage by predicting garbage calorific value |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102889598A (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103488209A (en) * | 2013-09-22 | 2014-01-01 | 浙江大学 | System and method for optimizing furnace temperature of pesticide wastewater incinerator of intelligent support vector machine |
| CN103631140A (en) * | 2013-12-09 | 2014-03-12 | 中南大学 | Method for automatically regulating temperature of flame path of coke oven in heating and combusting procedures on basis of performance evaluation |
| CN103900092A (en) * | 2014-03-26 | 2014-07-02 | 广州环投技术设备有限公司 | Automatic combustion control system for municipal solid waste incinerator |
| CN104125584A (en) * | 2013-04-27 | 2014-10-29 | 中国移动通信集团福建有限公司 | Service index realization prediction method aiming at network service and apparatus thereof |
| CN103488208B (en) * | 2013-09-22 | 2016-01-13 | 浙江大学 | The optimizing temperature of pesticide production waste liquid incinerator system and method for least square |
| CN106016294A (en) * | 2016-05-31 | 2016-10-12 | 浙江中控技术股份有限公司 | Measuring method and device for domestic garbage calculation heat value |
| CN106051782A (en) * | 2016-06-17 | 2016-10-26 | 光大环保技术研究院（深圳）有限公司 | Method for estimating garbage heat value in real time |
| DE102017120222B3 (en) | 2017-09-01 | 2018-11-29 | Technische Universität Dresden | Method and system for predicting a fuel composition in the thermal utilization of waste |
| CN109781785A (en) * | 2019-01-31 | 2019-05-21 | 成都市兴蓉再生能源有限公司 | A kind of sludge-rubbish collaboration burning mixed material heating-value determination method |
| CN111069238A (en) * | 2019-12-24 | 2020-04-28 | 重庆工商大学环境保护研究所 | Household garbage treatment system and treatment method |
| CN112700033A (en) * | 2020-12-16 | 2021-04-23 | 浙江中控技术股份有限公司 | Gas calorific value estimation method and device based on combustion timing sequence model |
| CN113175678A (en) * | 2021-03-09 | 2021-07-27 | 广州特种承压设备检测研究院 | Method and device for monitoring garbage incineration |
| CN113926734A (en) * | 2021-11-01 | 2022-01-14 | 天津商业大学 | Sorting type garbage classification method |
| CN115358152A (en) * | 2022-08-26 | 2022-11-18 | 绍兴市再生能源发展有限公司 | Garbage incineration gas control and feedback regulation system and method |
| DE102022122433A1 (en) | 2022-09-05 | 2024-03-07 | WasteAnt GmbH | CALCULAR VALUE DETERMINATION |
-
2012
- 2012-10-24 CN CN2012104118041A patent/CN102889598A/en active Pending
Non-Patent Citations (1)
| Title |
|---|
| 马晓茜;谢泽琼: "基于BP神经网络的垃圾热值预测模型", 《科技导报》 * |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104125584A (en) * | 2013-04-27 | 2014-10-29 | 中国移动通信集团福建有限公司 | Service index realization prediction method aiming at network service and apparatus thereof |
| CN103488209A (en) * | 2013-09-22 | 2014-01-01 | 浙江大学 | System and method for optimizing furnace temperature of pesticide wastewater incinerator of intelligent support vector machine |
| CN103488208B (en) * | 2013-09-22 | 2016-01-13 | 浙江大学 | The optimizing temperature of pesticide production waste liquid incinerator system and method for least square |
| CN103488209B (en) * | 2013-09-22 | 2016-02-24 | 浙江大学 | The pesticide waste liquid incinerator furnace temperature optimization system of Intelligent Support vector machine and method |
| CN103631140A (en) * | 2013-12-09 | 2014-03-12 | 中南大学 | Method for automatically regulating temperature of flame path of coke oven in heating and combusting procedures on basis of performance evaluation |
| CN103631140B (en) * | 2013-12-09 | 2016-01-06 | 中南大学 | Based on the coke oven heating-combustion process fire path temperature Automatic adjustment method of Performance Evaluation |
| CN103900092A (en) * | 2014-03-26 | 2014-07-02 | 广州环投技术设备有限公司 | Automatic combustion control system for municipal solid waste incinerator |
| CN106016294A (en) * | 2016-05-31 | 2016-10-12 | 浙江中控技术股份有限公司 | Measuring method and device for domestic garbage calculation heat value |
| CN106051782A (en) * | 2016-06-17 | 2016-10-26 | 光大环保技术研究院（深圳）有限公司 | Method for estimating garbage heat value in real time |
| DE102017120222B3 (en) | 2017-09-01 | 2018-11-29 | Technische Universität Dresden | Method and system for predicting a fuel composition in the thermal utilization of waste |
| CN109781785A (en) * | 2019-01-31 | 2019-05-21 | 成都市兴蓉再生能源有限公司 | A kind of sludge-rubbish collaboration burning mixed material heating-value determination method |
| CN109781785B (en) * | 2019-01-31 | 2021-06-15 | 成都市兴蓉再生能源有限公司 | Method for measuring heat value of sludge-garbage co-incineration mixed material |
| CN111069238A (en) * | 2019-12-24 | 2020-04-28 | 重庆工商大学环境保护研究所 | Household garbage treatment system and treatment method |
| CN112700033A (en) * | 2020-12-16 | 2021-04-23 | 浙江中控技术股份有限公司 | Gas calorific value estimation method and device based on combustion timing sequence model |
| CN113175678A (en) * | 2021-03-09 | 2021-07-27 | 广州特种承压设备检测研究院 | Method and device for monitoring garbage incineration |
| CN113175678B (en) * | 2021-03-09 | 2023-11-10 | 广州特种承压设备检测研究院 | Garbage incineration monitoring method and device |
| CN113926734A (en) * | 2021-11-01 | 2022-01-14 | 天津商业大学 | Sorting type garbage classification method |
| CN113926734B (en) * | 2021-11-01 | 2022-11-01 | 天津商业大学 | Sorting type garbage classification method |
| CN115358152A (en) * | 2022-08-26 | 2022-11-18 | 绍兴市再生能源发展有限公司 | Garbage incineration gas control and feedback regulation system and method |
| CN115358152B (en) * | 2022-08-26 | 2023-04-07 | 绍兴市再生能源发展有限公司 | Garbage incineration gas control and feedback regulation system and method |
| DE102022122433A1 (en) | 2022-09-05 | 2024-03-07 | WasteAnt GmbH | CALCULAR VALUE DETERMINATION |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102889598A (en) | Control method for assisting stable combustion of garbage by predicting garbage calorific value | |
| CN107726358B (en) | Boiler Combustion Optimization System based on CFD numerical simulations and intelligent modeling and method | |
| CN106842909B (en) | Power generation system for load control of power plant and control method thereof | |
| JP4993336B2 (en) | How to optimize a combined cycle / process facility | |
| US8185216B2 (en) | Plant controlling device and method, thermal power plant, and its control method | |
| CN109420424A (en) | A kind of wet desulfurization of flue gas by limestone-gypsum method energy saving of system optimization method | |
| CN113433911B (en) | Accurate control system and method for ammonia spraying of denitration device based on accurate concentration prediction | |
| Romeo et al. | Neural network for evaluating boiler behaviour | |
| CN103759290A (en) | Large coal-fired unit online monitoring and optimal control system and implementation method thereof | |
| CN107831652A (en) | A kind of unit load intelligent optimized control method based on cold end system energy storage | |
| CN109670629A (en) | Coal-burning boiler thermal efficiency forecast method based on shot and long term Memory Neural Networks | |
| CN111829003A (en) | Power plant combustion control system and control method | |
| WO2015035134A1 (en) | Combined cycle power generation optimization system | |
| CN111102559A (en) | Screen type superheater steam temperature control method based on double neural network inverse model | |
| CN105488266A (en) | Actual power generation power prediction method of thermal power generating unit | |
| Abd-Kadhum et al. | Evaluation and improvement performance of a boiler in a thermal power plant using artificial neural network | |
| Manke et al. | Application of back propagation neural network to drum level control in thermal power plants | |
| CN112396162B (en) | Neural network model for predicting wall temperature of screen type superheater of coal-fired unit | |
| CN105447256A (en) | Excitation enhancement simulation genetic optimization method | |
| KR102291800B1 (en) | Apparatus and method for deriving boiler combustion model | |
| CN110888401B (en) | Combustion control optimization method and device for thermal generator set and readable storage medium | |
| Elshafei et al. | Boilers optimal control for maximum load change rate | |
| CN113885607B (en) | Steam temperature control method and device, electronic equipment and computer storage medium | |
| CN116293623A (en) | Boiler load distribution method based on energy efficiency curve | |
| Li et al. | Life extending control of boiler–turbine systems via model predictive methods |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20130123 |