Search

• The algorithm of recursive modified partial least square (RMPLS) is proposed and employed to enhance the precision of prediction with low computation complexity. • The problem of signal prediction is equivalently transformed into parameter prediction with the help of locally weighted projection regression (LWPR), which makes the nonlinear quadratic problem of predictive controller for nonlinear process solvable. • The proposed predictive control strategy is updated by learning from the online measurable data. It is therefore suitable for the tracking control of chemical industry with complex and changeable work conditions. In this paper, a data-driven predictive control strategy for nonlinear system is proposed and testified on a continuous stirred tank heater (CSTH) benchmark. A recursive modified partial least square (RMPLS) algorithm is employed to regress the local linear model. The algorithm of locally weighted projection regression (LWPR) is then leveraged to build the predictive model, based on which a novel data-driven predictive control strategy is put forward. The proposed predictive controller has the ability to deal with changing working conditions, benefiting from the incremental learning ability of RMPLS and LWPR. The performance of the proposed control strategy is demonstrated with the CSTH while the superiority is illustrated by comparison with an existing model-free adaptive control approach. [ABSTRACT FROM AUTHOR]

Abstract: This paper presents a distributed model predictive control (DMPC) algorithm based on Benders’ decomposition for temperature regulation in buildings using multiple heating sources. It is well known that the main objective of this control problem is to minimize the heating energy bills while maintaining a certain indoor thermal comfort. To reduce the heating costs, many buildings are equipped with different heating sources. An example is the use of a hot water based central heating system and local electric convectors as complementary heating sources. Using a linear system model of the controlled process, the MPC optimization problem can be solved by linear programming. In order to reduce the computational demand required to solve the minimization problem, the authors propose a DMPC algorithm based on the Benders’ decomposition. The effectiveness of the proposed approach compared to the currently used PI-based control is illustrated in a simulation study. [Copyright &y& Elsevier]