Multi-Objective Hydro-Thermal Scheduling Problem Using Two Novel Optimization Techniques

Introduction

Hydrothermal scheduling (HTS) is an essential organizing task in power system operation. Generally, thermal system scheduling is less complex than that the optimal scheduling of hydrothermal power system. HTS is mostly a non-linear problem including non-linear objective function and a combination of linear and non-linear constraints. Purchaser load requirements in electric power systems are subject to differ because human accomplishments exhibit an agenda of 1 day or 1 week while satisfying several limitations on hydraulic and thermal power system network. Meanwhile, the marginal production cost of hydroelectric plants is negligible, but one of the manufacturing problems is well organized usage of available water. Maximum generation from hydroelectric plants decrease the cost of thermal generation considering the presence of nonlinearities, large number of decision variables and a set of constraints coupled in time periods. Economic load dispatch (ELD) is a significant technique in the process of a thermal power system. Conventionally, the hydrothermal power systems are worked in such a manner that the total fuel cost is diminished irrespective of emissions created. The problem being undertaken in this paper is one of the short term optimum scheduling of generation units in order to minimize power fabrication cost and emission levels over a given period of time. Load demand is circulated among generating units in ELD (Wood & Wollenberg, 1994; Happ, 1997; Chowdhury & Rahman, 1990; Liu & Cai, 2005) problem which will satisfy generation limit, prohibited operating zone, ramp rate, etc., considering transmission loss at every time interval such that the over-all cost is minimum. Fossil fuel produces various pollutants, like nitrogen oxides, carbon dioxide, sulfur oxides, etc., into the atmosphere at the time of generation of electricity from thermal power plant. The power companies have to assured standards concerning about the emission levels of pollutants for the strict government guidelines on ecological protection. Since the text of the Clean Air Act Amendments of 1990 and similar acts by European and Japanese governments, environmental constraints have surpassed the list of effectiveness management concerns (Yalcinoz, Altun & Hasan, 2002). The consciousness due to the rising concern over pollutants, society hassles suitable and safe electricity not only at the inexpensive probable price, nevertheless at lowest level of pollution. Recently, ELD problem has been combined with emission dispatch (CEED) problem (Roy, Ghoshal & Thakur, 2010a; Roy, Ghoshal & Thakur, 2010b; Zhang, Luh & Zhang, 1999); where many researchers consider emission as an additional constraints or second objective function with minimize the cost economy. In the instance of short term hydro thermal load managing, it is typically expected that the volume of water entrances essential to encounter the load requirements and load demand are known with certainty. The several restrictions that cannot be disrupted HTS is an intricate decision making procedure. Water discharge rate, upper and lower bounds on reservoir volumes, hydraulic continuity restriction, operating capacity limits of hydro plant and water spillage constraints are make the HTS problem as a complicated constrained optimization problem whose viable solution space is enormous.

For the resolution of the hydrothermal scheduling problem (Zhang, Luh & Zhang, 1999; Al-Agtash, 2001), several mathematical optimization algorithms have been employed. Various classical techniques, such as decomposition approach (DA) (Pereira & Pinto, 1983), progressive optimality algorithm (Turgeon, 1981), linear programming (LP) (Mohan, Kuppusamy & Khan, 1992) and dynamic programming (DP) (Yang & Chen, 1989) have been deployed to solve the HTS problem. The main disadvantage with the majority of these techniques is the complicatedness of treating large-scale systems. The main disadvantage of the DP method is that the dimensional requirements and computational grow severely with increasing system size and planning horizon.