Decision Making Support for Selecting Structure of Generating Capacities at Development of Local Power Supply Systems

. This paper considers the problem of multi-criteria option selection for the purpose of developing generating capacities for local power supply systems. The authors suggest a 2-stage method of multi-criteria analysis. At the 1st stage one conducts a multi-criteria selection of the most prospective power generation technology. At the 2nd stage one conducts a multi-criteria assessment of options for power station capacity ratio on the basis of the selected efficient technology of power generation. At the first stage the authors suggest a modified method of hierarchy analysis allowing for the decrease in the number of requests to a decision-maker (DM). At the second stage the authors use the modified method TOPSIS with the application of value functions for calculating non-linear change of the DM preferences in relation to the assessment of options by the criteria.


Introduction
In compliance with the Energy Strategy of Russia for the period until 2035, an "optimistic scenario" of the energy policy stipulates for a more comprehensive use of the power sector to boost the economic growth and improving the welfare of the Russian population. According to this scenario a higher pace of social and economic development is suggested, especially for the regions of the Russian Far East and Eastern Siberia. This suggestion is connected to the implementation of additional number (more than 100) of investment projects which will involve the growth of demand for energy resources and require making up new energy and industrial complexes [1].
Due to this the priority direction of the energy strategy is forming a sustainable system of energy development of remote territories on the basis of local fuel and energy and renewable resources [1]. The development of eastern region infrastructure requires building local power supply systems (LPSS). This brings about the necessity to select an efficient technology of power production as well as reasonable proportions of power capacities on the basis of selected technology.
At present finding a rationale for developing power supply systems is a multi-stage process, and on each process subsequent stage the previously obtained solutions are specified. The complex nature of this process is defined by the necessity of taking into account multiple significant factors and the uncertainty of future conditions of the power industry developing.
The methodology of validating the development of power industry transfers from centralized state planning to a new system of a multilateral process of validating solutions and making up the mechanisms of their implementation under uncertainty, multiple criteria and multiple diverging interests [2]. These criteria are also fair for validating the options of local power supply systems, which in its turn requires developing an efficient method of selecting the structure of generating capacities (SGC) of LPSS.
In the framework of the described problem this article suggests a 2-stage method of SGC selection which allows excluding inefficient options of power generation during preliminary research and forming the first insight on a SGC with the consideration of multiplicity of purposes and criteria.

2-stage method of multi-criteria selection of LPSS SGC
While selecting a SGC, depending on the number of available kinds of energy sources and infrastructure regional development a need may arise to compare a larger number of power generation technologies. As a result of such multiple criteria a DM or a group of DM face a complex tasks needing a systematic approach. That is why it is suggested to resolve the problem in 2 stages. Fig. 1 shows the main stages of a 2-stage method of the LPSS SGC multi-criteria selection [3]. At the first stage the authors select the most prospective technologies of power generation from multiple options. This stage includes reference data gathering: defining the value of aggregate electrical load; forming alternative options for power generation technologies which could comprise power plants working at organic fuel (TPP), hydraulic power plants (HPP), wind power plants (WPP), solar power plants (SPP) and other. Further, with the help of a DM one forms a set of criteria and assesses alternative options on the selected criteria. As a result, a set of reference data is formed for subsequent multi-criteria analysis leading to the selection of the most prospective power generation technology.
At the second stage the selection of criteria occurs with the account of the peculiarities of selected power generation technology defined at the first stage. Forming the alternative options of a SGC occurs on the basis of various ratios of the identified capacities of the selected prospective technology and is conducted as follows (Fig. 2). The level of basic power plant capacity with a small manoeuvring ability working with continuous load. The second curve part is covered with power plants the energy resource intake of which cannot be managed. The remaining curve part is covered by manoeuvring stations.
Further, the authors assess the formed SGC options by the criteria with the help of objective and subjective models. The analysis of the SGC economic efficiency is conducted on the basis of assessment of joint power generation by power stations. For example, in case a SGC for power plants is selected in the framework of HPP, TPP, SPP and WPP, power generation at the moment i, providing for the fulfillment of HPP technological constraints in terms of capacity, discharge and head-flow characteristics, is defined with the help of the equation systems [3]:    The approaches towards the alternative options by other possible criteria is presented in [3]. As a result, the authors form a set of assessed alternative options by the criteria. Further, the authors conduct a multi-criteria analysis with the help of the selected method. It resulted in the selection of the most efficient ratios of the established power station capacities. Consider the methods used in the research for multi-criteria analysis of alternative options (steps 1.5, 2.4, Fig.1)

Methods of multi-criteria assessment for analyzing options at selecting LPGS SGC
A traditional approach towards the consideration of multiple criteria while resolving the problems of LPSS development consists in reducing the problem to a single criterion one by means of identifying one, main criterion and transferring other criteria to a range of limits. Such approach is grounded if there is a mechanism of assigning such limitations in the form of corresponding standards. In case there are no limitations on the criteria and for some reason or another they cannot be explicitly assigned, the problem is considered as a multi-criteria one [4].
Consider the methods selected and upgraded by the authors for applying the problem of a multi-criteria selection of LPSS SGC.

Modified method of analyzing hierarchies used to select power generation technologies
At the first stage ( Figure 1) the number of alternative options is approximately 10, so, it is suggested to use the method of the analytic hierarchy process (AHP) as a tool for multi-criteria analysis [5,6]. This method is oriented towards the comparison of a smaller number of alternative options, is characterized by a simplicity of perception for a DM and is widely applied at planning the development of power system in foreign countries [7,8]. The key method concept consists in the fact that a DM conducts pair-wise comparison of options by the criteria and the criteria against each other on the basis of the comparative importance scale ( Table 1). The stages of an original method one can find in [5]. One of drawbacks of the method is a significant growth in the number of requests to a DM given the increase in the number of options and criteria. The suggested method modification consisting in building a DM preference model basing upon several requests. The model allows filling option pair-wise comparison matrices without subsequent DM involvement.
Building the preference model consists in the following. While familiarizing the grading scale by the criterion, a DM is provided with the information that the comparative importance assessment cij -is the advantage the alternative option with the assessment хi has over the assessment хj by the criterion. Further, the authors pose questions the answers to which are defined by the DM preferences in relation to the assessment "moderate advantage -3" at the alternative option comparison. As a result, one is able to build a preference model of the following type ( fig. 3). The curve (Fig. 3) ∆xL 3 , x 3 R -explicitly expressed DM preferences in relation to the assessment "moderate advantage -3" at the comparison of alternative options. 1-L, 2-L, 3-L, 4-L correspond to the assessment areas 1-3, 3-5, 5-7, 7-9 at the DM answer -∆xL 3 while 1-R, 2-R, 3-R, 4-R corresponds to the assessment areas at the DM answer -∆xR 3 .
DM preferences in relation to the assessment values 5, 7, 9 are formed on the basis of the equation system [9]:  where хaminimum assessment of the options considered; xbmaximum assessment of the options considered.
On the basis of the preference model (Fig.3) it seems possible to conduct pair-wise comparison of alternative options and get a set of global priorities ( Table 2).  Table 2 shows viL , viС , viR as local priorities of alternative options corresponding to the limits and the core of fuzzy assessment of the option xi by the criterion. w1…wn -local priorities of criteria.
Each option of the DM preferences z allows obtaining the corresponding ranking for alternative options. The best option is considered to be the case which has a higher rank at a larger number of the DM option preferences: where zinumber of preference options at which the alternative i had a higher rank. When filling pair-wise comparison matrices the authors obtain perfect match of answers which allows gaining the correct results.
Using the proposed procedure allows for the decrease in the number of requests to a DM at the comparison of power generation technology [9].

Application of the TOPSIS method in case of uncertainty of preferences of the DM for selecting the ratio of power plant capacities
At the second stage (Fig.1), due to a larger number of alternative and fuzzy information on their assessment, it is suggested applying the fuzzy method TOPSIS. The main method concept consists in the measuring the distance of alternative options to an ideal positive and ideal negative solution (IPS and INS). The best option is the one that has the least distance to the first solution and, correspondingly, the longest distance to the second solution. The structure of an original method one can find in [10].
As a DM needs to express the preferences in relation to a SGC for a long-term outlook of several decades, there are difficulties occurring in relation to exact assessment. To take into account fuzzy DM preferences, the authors suggest introducing fuzzy single-criterion value functions (FSCVF) in the method. Fig. 4 presents the FSCVF example at increasing preferences. The building of such FSCVF can be implemented with the use of triangular fuzzy numbers in the DM answers by the method described in [11]. A higher index a corresponds to a left, b -to a right limit of fuzzy assessment, с -corresponds to the fuzzy assessment core. Building FSCVF is implemented with 5 check points [11]. The interval from [xi 0 , xi 1 ] defines the mean value point xi 0,5с and possible limit conditions xi 0,5a , xi 0,5b [11]. In a similar way the authors establish mean value points at the interval [xi 0 , xi 0,5c ] и [xi 0,5c , xi 1 ] with the value 25 0, i ṽ and 75 0, i ṽ correspondingly. Defining a fuzzy value by FSCVF is implemented with the help of a method described in [11].
The use of FSCVF in the TOPSIS method replaces the steps of normalizing according to the provisions and fuzzy numbers are used when accessing IPS and INS.

Application of 2-stage method for preliminary selection of LPSS SGC
Consider the application of a 2-stage method by the example of one of the prospective district of the Kamchatka Region. Stage 1 Step 1.1. The authors form a situation plan of the prospective district development on the basis of [12,13] which has a form (Fig. 5).
On the basis of the analysis conducted for similar objects [14] the authors define preliminary options of consumer capacities and form daily seasonal curves of LPSS load. The maximum load capacity amounted to 77 MW. Step 1.2. The assessment of efficiency of various power generation technologies conducted in [3] demonstrated the following results (Table 3). As this region has a well-formed system of diesel fuel supply, the alternative options for the 1st stage are: TPP, HPP, WPP, SPP, DPP.
At the stage 1.3-1.4 a DM helps to form a set of criteria. The 1st stage purpose is selecting the most efficient power generation technology. The authors suggest taking the following as the criteria: К1 -net discounted income at using the technology; К2assessing a comprehensive environment impact of the technology; К3 -assessment of public opinion on the plant type (К31 -population attitude to the manufacturing technology, К32 -provision of the population employment); К4 -assessing technical efficiency (К41 -influence of the technology on power quality, К42 -energy source manoeuvring ability). These criteria are used to assess alternative options with the help of subjective and objective models. The results are presented in Table 4. At building the preference model it is suggested that a DM answered the question in relation to the "moderate advantage" by the criteria in the following way (Table 5). Further the authors define importance factors at fuzzy preferences of a DM (Table 6). Then 81 variants of combinations of local factors were calculated. It is identified that the best alternative options are А3, А4, А1. The Table 7 shows mean values of global indices. The indicator Pi for all alternative options amounts to 1 which means that at all the DM preference options no change of power generation technology ranking occurs. Judging by the aforementioned, one can draw a conclusion on the sustainability of the obtained results towards DM preference change. Thus, to conduct the analysis at the second stage the authors adopt the alternative options HPP, SPP, TPP.
At the second stage (Fig.1) the authors corrected the structure of criteria with the account of peculiarities of the selected technology of HPP, SPP, TPP. As the assessment criterion К2 the authors selected subcriteria К21 -atmospheric pollution, К22 -assessment of biological action on flora and fauna, К23area of withdrawn lands. As the assessment of public opinion in relation to various options of the SGC it is suggested to introduce a subcriterion К33 -professional activity risk. In compliance with the method stated in [3] with the application of exact and expert models as well as the system (1) the authors define the following alternative option assessment (Table 8). Further, at conducting multi-criteria analysis with the help of a modified method TOPSIS the authors built FSCVF with the help of a DM. By the criteria К1, К31, К32 importance functions are linear. By other criteria FSCVF has the following form (Fig.6). As a result of applying the TOPSIS method the authors suggested the following option ranks (Table 9). Therefore, the best solution is the option А3: TPP -10 MW, HPP -67 MW, SPP -10 MW. The obtained solution corresponds to the DM expressed preferences.

Conclusion
This article suggests a 2-stage method of multi-criteria selection of the preliminary SGC in remote areas. As it can be seen, decision making in relation to the selection of the preliminary SGC stipulates for gathering a large amount of information and complex structure. The proposed method allows to systematize this problem, taking into account the multiplicity of purposes and criteria at the development of energy infrastructure of the newly reclaimed regions, to take into account the uncertainty of initial information and preferences of the DM.