Guest Editorial

摘要

In modern industrial processes, aerospace systems, vehicle systems, and elsewhere there are increased demands for fuel efficiency, conservation of resources, cost and energy savings, and other optimal performance requirements. However, there is generally no dynamical model available for the process, or the process model is too complex to be tractable for controller design. Modelling and system identification are expensive and time-consuming, and models may be time-varying, or non-linear, or contain delays. The term ‘Data-driven Control’ (DDC) originated in the 1990s in Computer Science and it shares the same context as ‘big data’, ‘data mining’, and ‘data fusion’. On the other hand, Data-Based System Modelling, Monitoring, and Control are a set of topics used in the Control Systems community. The development of all these topics was driven by the huge amounts of data measured in complex process control systems, both stored historical data from prior measurements and on-line data available in real time during process runs. In these fields, the intent is to efficiently use the information in huge amounts of process input/output data to design predictors, controllers, and monitoring systems that provide guaranteed performance of the process. This Special Issue presents the latest developments on data-driven modelling and control, iterative learning control and reinforcement learning, and their applications in process industries. It contains eighteen papers, the contents of which are summarised below. A novel hybrid intelligent dynamic modelling approach is proposed by Tie et al. in their paper entitled ‘Hybrid intelligent modelling and simulation for cold tandem rolling process’. The approach combines a linearised state space model, a case-based reasoning multi models selection, a case attributes optimisation, an adaptive fractal filtering algorithm and a compensation model for the strip velocity. Shen et al. propose a model-independent approach for short-term electric load forecasting with guaranteed error convergence in their paper. The paper introduces the tracking control and Lyapunov stability theory into the load forecasting algorithm design. It can approximate the load dynamics inherent laws, without statistically learning from a certain forecasting model. Hu et al. develop a data-driven controller tuning scheme in their paper ‘Convergence performance oriented data-driven tuning method for parameterised controller design with cases investigation’. An iterative law based on the behaviour between the current parameter and the optimal parameter is proposed, which has the ability to directly seek the global optimal parameter. The paper entitled ‘MIMO system experimental validation of model-free control and virtual reference feedback tuning techniques’, by Precup et al., proposes three data-driven MIMO control system structures applied to the control of a representative non-linear MIMO system represented by the twin rotor aerodynamic system. Gao et al. address the adaptive and optimal control of a Quanser's 2-degree-of-freedom helicopter via output feedback in their paper ‘Sampled-data-based adaptive optimal output-feedback control of a 2-DOF helicopter’. The paper presents a policy iteration algorithm which yields to learn a near-optimal control gain iteratively by input/output data. The convergence is theoretically ensured and the trade-off between the optimality and the sampling period is rigorously studied as well. The authors show the performance of the proposed algorithm under bounded model uncertainties. A data-driven optimisation solution for operational index control to the selection of the set-points for a class of industrial processes is presented by Lu et al. in their paper ‘Data-driven optimal control of operational indices for a class of industrial processes’. A reinforcement learning actor-critic structure is employed to provide a data-driven optimisation. The effectiveness of the proposed method is demonstrated b