Some achievements on using MPC for system frequency Abstract-This paper deals with the application of algorithms control [18][19][20] and power plant control [21] problems have inspired by Model Predictive Control to solve voltage-related been reported. This paper, however, focuses on the control of power system control problems in both normal and emergency transmission system voltages. operating conditions. In the first part of the paper, we identifyWe first discuss what we believe are critical issues for the critical issues for a practical implementation of this methodology, and analyze how far these requirements have been met so far. In applicaton of MPG techniques to voltage control problems. the second part, we outline a voltage control scheme that For each issue, we briefly review some of the proposed hopefully addresses the above issues. The central idea of this approaches [6][7][8][9][10][11][12][13][14][15][16][17] and possibly identify some further research scheme is a static optimization to determine target control values, needs.followed by a dynamic optimization to produce a feasible The paper is organized as follows. Section 1I recalls the Model Predictive Control (MPC) techniques are both standard MPC formulation. Section III addresses some of the mature (being considered a viable control strategy, they have critical issues for MPC application to voltage problems, and been used for years in process industry [2,3]) and theoretically briefly surveys the proposed approaches. Section IV outlines a understood (see [4,5] for a good account on the main new MPC voltage control scheme while Section V offers theoretical results). It is thus somewhat surprising that up to some conclusions. now this control technique has not received more attention from power system researchers and practitioners, although II. BRIEFREVIEWOFMODALPREDICTIVEGONTROL recent references show a growing interest for this approach Standard MPC (also referred to as moving horizon or [6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21].receding horizon control) is essentially a class of computer Practical interest in using MPC is driven by the fact that control algorithms to control the future behaviour of a system industrial processes need to be operated under tighter through the use of an explicit model of the latter [4,5]. At each performance specifications and at the same time more and control step the MPC algorithm computes an open-loop more constraints (dictated by economical, environmental, and sequence of controls in order to optimize the future system safety considerations) need to be satisfied [2,3]. This fact is, behaviour. The first control action in the optimal sequence is more than ever, present in today's power systems. Precisely, applied and the entire optimization is repeated at subsequent the ability of MPC to incorporate various constraints makes it control steps. The principle of MPC is illustrated in Fig. 1. attractive in this respect. Obviously, one could also quote the It is convenient to formulate the MPC p...