This paper presents a hybrid approach consisting of Modified Priority List and Sequential Approach (MPLSA) for solving thermal Unit Commitment (UC) problems. The solution methodology comprises two-stage processes of unit selection and economic dispatch of the committed units. The Modified Priority List (MPL) was used for unit selection, which was developed by sorting the units in descending order in respect of their Contribution Factor (CF). The CF is a ratio of the maximum generation capacity to the total installed capacity of system. Sequential Approach with Matrix Framework (SAMF) was applied to obtain the optimal dispatches of the committed units. The proposed methodology was validated by solving the standard ten-unit system, and the effectiveness of Prohibited Operating Zones (POZ) in UC problem was also studied. The initial state and minimum up and down time of units and spinning reserve requirement constraints were also considered. Numerical simulation results indicate an improvement in the solution quality. The solution methodologies are entirely based on generation limits that narrow the solution space which directs to the optimal unit selection and economic scheduling. The proposed method can be extended to solve security-constrained UC problems.

Unit Commitment (UC) is an important optimization problem in power system operation that aims to schedule the generating units online or offline over a scheduling horizon such that the power production cost is minimized, with the load demand fully met and the operation constraints satisfied. The UC problem is combinatorial in nature and consists of many hard constraints. In solving this problem, generator schedules are first found and their costs are evaluated through the economic dispatch calculations (Lee, 1988; and Wood and Wollenberg, 2002).

The solution methods being used to solve the UC problem are Priority List (PL) method (Lee, 1988; Sheble, 1990; and Wood and Wollenberg, 2002), Dynamic Programing (DP) (Ouyang and Shahidehpour, 1991), Branch and Bound (BB) method (Cohen and Yoshimura, 1983), Lagrangian Relaxation (LR) method (Zhuang and Galiana, 1988; and Virmani et al., 1989), Mixed Integer Programing (MIP) method (Dillion et al., 1983) and artificial intelligence methods. Senjyu et al. (2006) have proposed Extended Priority List (EPL) approach for solving UC problems. In the PL methods, the units are committed in the ascending order of the unit full load cost. The most economic base load units are committed first and the peaking units are last in order to meet the load demand. The PL methods are very fast but they are highly heuristic and give schedules with relatively high operation cost. The DP method has been extensively used for the solution of the UC problem but it has curse of dimensionality. The BB method has the danger of deficiency of storage capacity, with increasing the calculation time enormously being a large-scale problem. The BB method adopts a linear function to represent the fuel consumption and time-dependent start cost and obtains the required lower and upper bounds. The shortcoming of the BB method is the exponential growth in the execution time with the size of the UC problem. The LR method concentrates on finding an appropriate coordination technique for generating feasible primal solutions, while minimizing the duality gap. The main problem with the LR methods is the difficulty encountered in obtaining feasible solutions. The MIP method employs linear programing technique to solve the UC problem but it also suffers from an excessive computational time requirement. Artificial neural network-based method was developed to solve UC problem (Sasaki et al., 1992), but it can be limited in its applications to large-scale UC problem. Fuzzy logic programing and control logic programing were suggested to solve the UC problems (Saneifard et al., 1997; and Huang et al., 1998). The modified or hybrid approaches of DP, LR and MIP were developed to solve the thermal UC problems (Su and Hsu, 1991; Ouyang and Shahidehpour, 1992; Madrigal and Quintana, 2000; Takirti and Birge, 2000; and Carrion and Arroyo, 2006). The methods based on Enhanced Lagrangian Relaxation (ELR) and Augmented Lagrangian Relaxation (ALR) were presented for the solution of UC problem (Ongsakul and Petcharaks, 2004).

Electrical and Electronics Engineering Journal, Unit Commitment (UC), Prohibited Operating Zones (POZ), Generation
scheduling, Economic dispatch.