Deregulation and liberalization of the power sector have resulted in forming several independent companies that provide special services. The transmission system operator (TSO) provides transmission services for reliable operation of the power system. The reactive power is needed to support the system voltage profile, thus the TSO has to obtain it from producers through different mechanisms that are presented in the subsequent text. The reactive-power markets differ from the classical power markets because of certain specific characteristics of the reactive power. As it is of a local character, it can not be transmitted over long distances causing unacceptable losses. Moreover, power lines produce the reactive power in underloaded conditions. In this way, the voltage profile is supported as well. The question is how to evaluate this "service" since it requires a special organization of the reactive-power market. The answer is not straightforward, thus the TSO is commonly the only procurer of the reactive power required for adequate reserves for voltage control. Current models of pricing of the procured reactive power include the capacity and real-time payment [1]. There are at least three methods for determining the capacity (power) payment. In a cost-based-payment model producers are paid for the available reactive power capacity. The reactive power can also be secured by an auction model, which is totally a market-oriented solution. The third possibility is an obligation for providing reactive power by the generators, which is a less liberal solution. Real-time pricing involves the reactive power. There are several methods commonly used. In the first solution, a generator is not compensated when operating within the power factor or contracted range. Otherwise it may be paid for its opportunity costs. The second possibility is an auction model similar to the one for the active power. The third method uses a pricing formula announced in advance to the generators. A combination of all models is expected to be the best solution for the reactive-power pricing. The Capacity payment should cover the investment costs and the real-time payment should cover the operational costs. On the other hand, both types of the costs could be incorporated in one pricing mechanism. In addition, penalty mechanisms for non-performance should be introduces in all these models. All costs arisen from the reactive-power procurement by TSO have to be covered by consumers and producers. In most systems, those costs are bundled in total ancillary-service costs and charged by the methods used for transmission-service pricing. Those solutions allocate the costs only according to the active power and have some specific disadvantages. The postage-stamp method [2] is based on a simple key that does not provide fair allocation (Equation 3.1). The contract-path method takes into account predefined supply paths of the active power purchase, thus the actual costs are not completely compensated (Equation 3.2). The MW-mile method has similar disadvantages as the contract-path method (Equation 3.3), thus it is not an appropriate solution either. In the paper, the LGDF (Line Generation Distribution Factors) and LLDF (Line Load Distribution Factors) for power-flow tracing are used to allocate the reactive-power costs. This allocation is based on the TGDF method and it overcomes its weaknesses by introducing some simplifications presented in this paper. The method follows the Kirchhoff's Current Law and uses the Pi-model of a line shown in Figure 1 when the reactive power is concerned. Decoupling of the line power flow shown in Figure 2 is introduced in order to take into account the line losses. The matrix notation (Equation 4.1) leads to definition of factors Q(ij,p) and Q(k,p), through Equations 4.2-4.6. Q(ij,p) is the share of the load or generator p on the line i-j and Q(k,p) is the share of the load or generator p in the node k. The complete procedure is exactly described in [8] and [14]. The new method uses Equations 4.7-4.9 to allocate the reactive power costs and was tested on a simple 5-bus test system (Figure 3). Results of different methods are compared (Tables 4 and 5) as well.
