Prediction Intervals for Neural Networks via Nonlinear Regression

Multivariable Generalized Minimum Variance Control Based on Artificial Neural Networks and Gaussian Process Models

“…Notice that equations (18) and (19) have additional terms compared to (9) and (10). As pointed out in [6] these terms provide extra robustness in the face of uncertainty in the output estimation.…”

“…The uncertainty associated with each output can be estimated by several standard confidence bound estimation algorithms [7], [10] which normally gave satisfactory results. However, the size of the estimated confidence interval could be biased [11].…”

Solid fuel decomposition modelling for the design of biomass gasification systems

“…Secondly, as in Duret et al (2005), a stoichiometry is defined, and fitted to the product distribution by letting reaction equilibrium temperatures vary from the measured gasification temperature. A complete stoichiometry is generated by writing the molecule and element formula matrix in its reduced row echelon form (Smith and Missen, 1982),…”

“…Duret et al, 2005). A major impediment to CHP remains equipment fouling problems related, namely, to the production of a wide array of condensable species commonly termed tar.…”

“…In this respect, reaction modelling is a challenging task, because of the difficulty of identifying and quantifying heavier products, and the lack of thermophysical properties for a variety of feedstocks, chars, and tars. In addition, the quantity and the nature of tars vary according to several factors such as reactor temperature and pressure (Shafizadeh, 1982;Evans and Milne, 1987), the nature and ratio of oxidising gases (Kinoshita et al, 1994), the nature of biomass feedstocks (Fagbemi et al, 2001) especially the catalytic effect of their inorganic content (Shafizadeh, 1982), and even the types of gasifiers (Buekens and Schoeters;1985). These issues have an impact on choices between design alternatives that affect the overall efficiency of process design.…”