Neuro-fuzzy estimator, with complexity reduction, of the temperatures of a parabolic-trough solar field

The estimation of unobservable states of a process is important when state space control techniques are applied. These controllers assume that states values are known. When all the states are measurable, there is no need to apply any observer. The case of the solar trough plants using a distributed parameters model presents many state variables which cannot be measured with sensors. In this work an observer based on a fuzzy inference system to estimate the temperature profiles of the loops that make up the solar field is presented. A complexity reduction technique based on Functional Principal Analysis is applied to make the estimator realizable in practice without occupying much memory or spend so much time in its programming in industrial devices.

Automated summary

This paper presents an observer based on a fuzzy inference system to estimate the temperature profiles of the loops in a solar field, addressing the issue of unmeasurable state variables in solar trough plants. A complexity reduction technique is applied to make the estimator practical without consuming excessive memory or programming time in industrial devices.