Abstract
Two interesting techniques for on-line operation optimisation are tested and further evaluated. The two techniques are multivariate modelling with PLS (Project of Latent Structures, the Simca-method), and process identification (the Eurosim-method). The purpose of the test was to for the Simca-method to check the ability of a multivariate model to predict NO{sub x}-SO{sub x}- and CO-levels in the fuel gas and the combustion efficiency. The ability of the model to find important key-parameters were also investigated as how the model were influenced by non-controllable parameters. The project includes designing of tests, evaluation with PLS and presentation of the results gained from the model. The predictions of the model have been verified in separate tests. The tests of the Simca-method have shown that in the plant in Naessjoe the model could be used for prediction of NO{sub x} and combustion efficiency with good accuracy. Usage of the technique can be done in two ways: * In large plants with a well defined process and a ABB-master control-system. On-line usage and prediction of process data for optimisation. * In plants where the process performs poorly and the knowledge of process parameters are small. Modelling and parameter-study off-line for rough optimisation and parameter
More>>
Wilde, J;
Harnevie, H;
[1]
Bergdahl, B G
[2]
- Vattenfall Utveckling AB, Vaellingby (Sweden)
- Eurosim AB, Nykoeping (Sweden)
Citation Formats
Wilde, J, Harnevie, H, and Bergdahl, B G.
Testing of on-line operation optimization; Test av on-line driftoptimering, etapp 2.
Sweden: N. p.,
1994.
Web.
Wilde, J, Harnevie, H, & Bergdahl, B G.
Testing of on-line operation optimization; Test av on-line driftoptimering, etapp 2.
Sweden.
Wilde, J, Harnevie, H, and Bergdahl, B G.
1994.
"Testing of on-line operation optimization; Test av on-line driftoptimering, etapp 2."
Sweden.
@misc{etde_10142860,
title = {Testing of on-line operation optimization; Test av on-line driftoptimering, etapp 2}
author = {Wilde, J, Harnevie, H, and Bergdahl, B G}
abstractNote = {Two interesting techniques for on-line operation optimisation are tested and further evaluated. The two techniques are multivariate modelling with PLS (Project of Latent Structures, the Simca-method), and process identification (the Eurosim-method). The purpose of the test was to for the Simca-method to check the ability of a multivariate model to predict NO{sub x}-SO{sub x}- and CO-levels in the fuel gas and the combustion efficiency. The ability of the model to find important key-parameters were also investigated as how the model were influenced by non-controllable parameters. The project includes designing of tests, evaluation with PLS and presentation of the results gained from the model. The predictions of the model have been verified in separate tests. The tests of the Simca-method have shown that in the plant in Naessjoe the model could be used for prediction of NO{sub x} and combustion efficiency with good accuracy. Usage of the technique can be done in two ways: * In large plants with a well defined process and a ABB-master control-system. On-line usage and prediction of process data for optimisation. * In plants where the process performs poorly and the knowledge of process parameters are small. Modelling and parameter-study off-line for rough optimisation and parameter knowledge. The tests of the Eurosim-method have shown that process-identification is a powerful tool for mainly NO{sub x}-minimisation. The method is a tool to automatically, on-line predict the actual relationship between NO{sub x} and O{sub 2} with continuous upgrading. The results of the method can be either an on-line system where the result is given on a screen to operator or a system where the O{sub 2}-level is automatically adjusted to reach minimal NO{sub x}-level from given limitations in for example CO-level. 3 refs, 29 figs, 5 tabs}
place = {Sweden}
year = {1994}
month = {Jan}
}
title = {Testing of on-line operation optimization; Test av on-line driftoptimering, etapp 2}
author = {Wilde, J, Harnevie, H, and Bergdahl, B G}
abstractNote = {Two interesting techniques for on-line operation optimisation are tested and further evaluated. The two techniques are multivariate modelling with PLS (Project of Latent Structures, the Simca-method), and process identification (the Eurosim-method). The purpose of the test was to for the Simca-method to check the ability of a multivariate model to predict NO{sub x}-SO{sub x}- and CO-levels in the fuel gas and the combustion efficiency. The ability of the model to find important key-parameters were also investigated as how the model were influenced by non-controllable parameters. The project includes designing of tests, evaluation with PLS and presentation of the results gained from the model. The predictions of the model have been verified in separate tests. The tests of the Simca-method have shown that in the plant in Naessjoe the model could be used for prediction of NO{sub x} and combustion efficiency with good accuracy. Usage of the technique can be done in two ways: * In large plants with a well defined process and a ABB-master control-system. On-line usage and prediction of process data for optimisation. * In plants where the process performs poorly and the knowledge of process parameters are small. Modelling and parameter-study off-line for rough optimisation and parameter knowledge. The tests of the Eurosim-method have shown that process-identification is a powerful tool for mainly NO{sub x}-minimisation. The method is a tool to automatically, on-line predict the actual relationship between NO{sub x} and O{sub 2} with continuous upgrading. The results of the method can be either an on-line system where the result is given on a screen to operator or a system where the O{sub 2}-level is automatically adjusted to reach minimal NO{sub x}-level from given limitations in for example CO-level. 3 refs, 29 figs, 5 tabs}
place = {Sweden}
year = {1994}
month = {Jan}
}