شبیه سازی داده های مربوطه سرعت باد برای ارزیابی پخش بار اقتصادی (ED)

ABSTRACT Simulation of Correlated Wind Speed Data for Economic Dispatch Evaluation

ECONOMIC dispatch (ED) co power to be generated among available generators, consists of dispatching the in order to obtain the most efficient, low-cost, and reliable operation of a power system. It considers operating limits, availability, and reliability, and minimizes costs so that both load demand and losses are supplied [1]. It, therefore, plays a key role in power system planning and operation. As wind power increases its share ratio in electrical networks, the ED problem must consider scheduled wind power and its costs. Even though wind power forecasting methodologies [2] have been considerably improved during the last years, wind power cannot be scheduled with total accuracy. The generally used models for wind power forecasting are based on several factors: current data and atmospheric behavior, such as the numerical weather prediction (NWP); historical data, such as autoregressive (AR), autoregressive moving average (ARMA), etc.; spatial correlation; artificial intelligence; or a combination of these [3]–[6]. Therefore, taking into account that scheduled and available wind power may not coincide, the costs to be considered in the ED problem are different in nature, as will be explained later. The ED problem is solved a priori, which means that the power to be produced is scheduled. Flexible plants can set output power to the required value, so both the scheduled and the produced power values coincide unless the unit is malfunctioning. There are other things to consider when wind power is taken into account because the available and scheduled powers at the corresponding wind turbine (WT) may differ and give rise to costs that must be evaluated. Several types of procedures can be applied to solve the ED problem. Analytical solutions like those in [7] are very difficult to apply when considering wind power correlation so numerical solutions like the ones in [8] are preferred. Therefore, the ED evaluation carried out in this paper is based on four terms that depend on the following: 1) the scheduled power of flexible plants; 2) the scheduled wind power; 3) the difference between available and scheduled wind power, when it is positive; 4) the difference between scheduled and available wind power, when it is positive. From the point of view of the system operator, there is a first term for the cost due to the sum of the power provided by the conventional generators. The second one is due to the amount to be paid to wind power producers according to agreements established with them. A term is included that evaluates the cost of the difference between available and scheduled wind power. These costs include the payment to the wind power producers for not using the available power, which is wasted or diverted to another generation facility like a hydro pumping station. Therefore, the third term is basically related to penalty costs due to not using all the available power in the network, i.e., it evaluates the cost of under-forecasting wind power. On the other hand, if the wind power is over-forecasted, the power requested by the load demand has to be supplied anyway, so the power must be purchased from an alternative source. Therefore, the cost of the reserve power or the cost of the power purchased through an interconnection is evaluated in the fourth term. So, these four terms evaluate the costs in order to solve the ED problem. However, this is a general model to evaluate ED and it is adaptable to all possible situations, so any term can be removed depending on specific cases. Notice that if the total power generated and demanded in an electrical network were not equal, then steady-state security could be affected, because this balance is needed for keeping adequate operating conditions. Thus, the electrical network could be seriously influenced by under- or over-forecasting wind power if the steps towards achieving the balance were not taken into account.