Single target for arrangement cannot fully think over the physical natures of electricity generation units and power grid safety limits, therefore it brings insufficient adjusting ability and dangerous hidden problems in real practice, hence it finally causes bad application effects. For solving this problem, this paper puts forward a method which is used for optimizing the day-ahead initial generation plan of generation units through considering many constraint conditions. First, by coupling monthly contracted electricity volume with unit start-up and shutdown status, a dynamic load factor recursive algorithm based on the weight of daily maximum generating capacity is designed to provide differentiated weights for fair allocation. Then, a multi-constraint set including unit ramp-up process, output obstruction intervals, and regional balance is constructed. Through constraint coupling and iterative relaxation mechanisms, the feasible region of unit output is accurately characterized. Finally, an iterative allocation strategy based on load factor weights is adopted, combined with a closed-loop feedback mechanism of virtual load correction and regional fixed output sequence relaxation, to correct unit output exceeding limits and eliminate power generation and consumption imbalances round by round. Under the premise of satisfying system load balance, a practically executable initial generation plan is generated. Experimental results show that after applying the proposed method, the success rate of plan preparation is improved, the system load deviation is smaller, the standard deviation of unit load factor is lower, and the application effect is better.
