Table 4.
Estimation method | Nominal | CS | NLS | ELS |
---|---|---|---|---|
Noiseless system with dense dataa | ||||
k1 | 0.0666 | 0.069 | 0.01 | 0.0666 |
k2 | 0.0333 | 0.0344 | 0.01 | 0.0333 |
Noisy systemb with dense dataa | ||||
k1 | 0.0666 | 0.0629 | 0.01 | 0.0666 |
k2 | 0.0333 | 0.0340 | 0.01 | 0.0333 |
Noiseless system with sparse datac | ||||
k1 | 0.0666 | 0.0615 | 0.01 | 0.0666 |
k2 | 0.0333 | 0.0303 | 0.01 | 0.0333 |
To estimate parameters for the noisy system we filter the data using the wden function provided by MATLAB (MathWorks: MA) version R2017a as input to the CS, NLS or ELS method
aDense data signifies data sampled at 1 min intervals
bNoisy system signifies 20% proportional error
cSparse data signifies data sampled at 45 min intervals