Power demand control algorithm for single controllable load with parametric and variable load factors

Abstract

Summary Power demand control is required at steelmaking plants and other energy intensive industrial plants where the electrical bills include high demand charges and are subject to demand limit contracts. This paper introduces a novel and simplified algorithm for the implementation of a power demand controller. The presented algorithm does not require a model of the electrical load since it is based on a geometrical approach for the estimation of the disconnection time of the controllable load and only requires the nominal power values and the load factors of both the total (full) load and the noncontrollable (base) load as parameters. The power demand control method presented is verified with a computational simulation, and precise control of the demand limit is achieved under random variations of the base and full loads. The proposed control algorithm is simpler than commercial methods and achieves a competitive performance. Simulation results show a better performance by calculating the disconnection times based on load parameters specified as mean values between the average and nominal powers of the full and base loads. Some considerations to tune‐up the load factor parameters are also discussed. The presented control method has been successfully implemented on industrial controllers in two steelmaking facilities in Mexico, having an electric arc furnace as controllable load in both cases. This novel demand control is adequate for industrial plants with electrical loads that can be disconnected without affecting the quality of the product and with minor impact on productivity.

A Peak Demand Control Algorithm for Multiple Controllable Loads in Industrial Processes

“…The control algorithm presented in this work is an enhancement of the algorithm presented by Martell et al in [28]. The contribution of the present algorithm is sustained in the fact that it adds functionality for the control by considering not one, but multiple loads through a dynamic computation of the multi-load parameters.…”

“…where P T , P NC and P C are the nominal power value of the total, non-controllable and controllable loads respectively; LF T , LF NC and LF C are the load factor values of the total, non-controllable and controllable loads respectively, while t t is the time length of the complete interval control (typically 900 seconds). According to [28] this parametric approach is the principal advantage of the algorithm as it can improve the accuracy of the disconnection time computation depending on which of both parameters, nominal or average power, is used. When using the nominal power, the disconnection time can be shortened (over anticipated); while the use of the average power may imply exceeding the demand limit, due to the variation in the behavior of the loads.…”

“…This could be an inconvenient when implementing this type of algorithm on field because of the specific knowledge or extra time needed to model the loads. In [28], the authors addressed this issue by proposing a parametric algorithm that instead of modelling the load, it utilizes basic electrical parameters such as the load factor for computing the precise disconnection time before the maximum demand value is reached; however, such algorithm only works for single load scenarios, wherewith its applicability is very limited, as most industrial processes are composed of multiple electrical loads.…”