Multi Echelon Distribution Model for Electric Market Deregulation Collaboration Strategy in East Kalimantan

Electric market deregulation aims to provide flexibility for customers to have many suppliers and low prices. However, market deregulation also provides opportunity for certain parties to manipulate supply, so there is a scarcity of products that result in price increases. Therefore, this research proposes a collaborative strategy for electricity market deregulation using the Multi Echelon distribution model applied in East Kalimantan with dummy data. Collaboration strategy made based on optimization of mathematical models in two stages and three scenarios. The simulation was carried out using Excel Solver covering three regions, dynamic time and estimated price fluctuations over three periods. As a result, Gencos gets the biggest profit when serving basic load. Whereas wheeling occurs, the D4 region is a strategic area that generates the largest profit compared to the D1 and D6 regions.


Introduction
Electricity market deregulation has changed the vertical integration of electric power production systems into three separate systems namely Gencos, Transcos and Discos. Gencos are companies that have power plant to produce electricity, while Transcos are those that own transmission network to send electricity from producers to consumers. The last part of this system is Discos, which consist of companies that are responsible in distributing the electricity to end customer. The separation of these three systems are aimed to create competition among Genco companies so that electricity can be produced at low cost, and ensure continuity of electric supply.
Based on literature electricity rates are influenced by three factors, i.e., fuel costs, losses and transmission costs [1], while the continuity of supply depends on the amount of production reserves at the power plant [2]. Fuel costs depend on the characteristics of the power plant [3]. The difference in power plant characteristics among Gencos companies causes the allocation of loads needs to be integrated. This integration can be done by using Economic Dispatch process as in [1]- [3].
Electricity produced by the power plant is sent to customers through Transcos. Gencos should achieve cost minimization in order to avoid additional cost in the form of losses. System integration between Gencos and Discos can be done with Price Based Dynamic Economic Dispatch (PBDED) [7], [8]. PBDED has succeeded in creating integration between the two with low cost, even though the results are still lacking in detail [9]. More detailed PBDED obtained by combining PBDED models with multi . Merging these two models successfully integrates Gencos and Transcos to deliver electricity to customers (DISCOS). The result of application of this model is very dependent on the characteristics of supply and demand in respective area.
This paper applied a combined PBDED model with multi echelon in the East Kalimantan electricity system. The electricity system in respective area consists of interconnected power plants through 150 kv and 20 kv transmission networks. Each power plants are situated in sparsely area as well as serving wide range location of end customer. The result of this research is providing infomation of optimal allocation of power plants with parameters of costs and emissions using Excel Solver.

Methods
This research was conducted using 5 stages: Problem formulation, Literature review, modeling, Simulation, and Analysis ( Figure 1). Formulation of the problem. What is the best collaboration of power suppliers in East Kalimantan to get the smallest total cost and emission?
Literature study. All of journals used as literatur in this research published by Elsevier on Economic Dispatch, optimization, and collaboration topics.
Modeling. There are three stages of modeling. First, build the conceptual model to facilitate in designing details of mathematical models; Second, provide mathematical model to illustrate supplier collaboration. Third, Verification and validation. So, there are four mathematical models in this research: SEC + RE, SEC + IPP + RE, SEC + Rent + RE, and SEC + EC + RE.

PBDED Model and Multi Echelon Distribution
The PBDED model as in [4] is aimed to maximize profit with the following objective functions: The purpose of the PBDED model is to maximize profit not just minimizing fuel costs as in (Columbus & Simon, 2013 While additional constraints are as in [5]: Equation (7) is used to model the cost of fuel with a regular load, whilst equation (8) applied to accomodate the cost of fuel when wheeling occurs. The 9 th equation is aimed to calculate the transmission cost. In order to model the fuel cost we used a quadratic function in equation (10). Whereas the last two equation i.e. equation (11) and (12) were MW-mile method to calculate the transmission of 150 KV and 20 KV Simulation. Simulation in this research is divided into two stages. In the first stage we did two calculation, which is allocation and distribution of electricity using basic loads, and calculation of allocation and distribution of electricity with additional demand loads. The second stages consist of collaborative scenarios. Scenario 1, we put additional demand of power in D1 region; Scenario 2, additional demand in the D4 region; and Scenario 3, additional demand in D6 region.
Analysis. This section contains Results and Discussion. In Results section we provide data and computational process whilst in Discussion various finding will be presenter.

Results
In the calculation we used dummy data (as shown in Table 1). Simulation carried out dynamically using three periods by dividing demand into two scenarios namely regular and wheeling. The estimated selling price of electricity varies over three periods.

Discussion
Collaboration between power plants to service basic loads requires a total fuel cost of Rp. 16,986,879. This cost required to produce 35,152 MW of electricity to meet the demand of 28,300 MW for three periods. It can be seen that that the amount of production is greater than demand due to losses occured in transmission and distribution networks (Equations 4 and 10). Under this skenario revenue gained was Rp. 33,595,000 and profit generated as much as Rp. 14,657,398. This basic load scheme is used as basis reference when we set additional demand in D1, D4, and D6 regions. Additional demand in D1 of 450 MW affected revenue to increase by Rp. 520,000. However, along with increasing in production, the cost of fuel was also rising in Rp. 11,453,890, from Rp. 16,986,879 to Rp. 28,440,769. The imbalance between increasing in fuel costs and revenue caused a decreasing in profit as much as Rp. 9,275,583. Revenue increases that are not proportional to the increase in profit are due to the equation of fuel costs (Equation 13) and the improper determination of selling prices of products. As a result tariffs for additional products must be distinguished from product rates for basic loads.
Additional demand for D4 of 450 MW caused an increasing in revenue of Rp. 520,000. The consequence was increasing in cost of production which is caused by increasing in fuel from Rp. 16,986,879 to Rp. 17,748,035 (Rp. 761,156). Compare to the first scheme, revenue gained from this area was smaller than the increase in fuel costs. However, there was an increase in profit of Rp. 1,689,459 due to additional demand in D4 region has changed the allocation structure of the power plant Additional demand for D6 of 450 MW caused revenue to increase by Rp. 520,000, but increased production also increased fuel costs from Rp. 16,986,879 to Rp. 17,748,035 (Rp. 761,156 difference). In this scheme the increasing of revenue generated was smaller than the increasing of fuel costs. However, we found significant increase in profit as much as of Rp. 1,672,939,-duet o additional demand in the D4 region has changed the allocation structure of the power plant production. The closer location of power plant in the production has caused significant decreasing in cost of transmission and distribution. The first scenario, minimizing costs. This scenario requires a total cost of Rp. 28,440,769 to produce electricity as much as 40,963 MW. Scenario 2, emission minimization. This scenario requires a fuel cost of Rp. 49,419,721 to produce electricity 38,901 MW. Scenario 3, Minimizing losses. This scenario requires a fuel cost of Rp. 38,544,976 to produce electricity as much as 37,379 MW, according to Table  3.

Wheeling in the D1 region
Additional demand causes an increase in the amount of production, and transmission / distribution costs. The first scenario requires an additional fuel cost of Rp. 11,453,890. The cost of wheeling 500 kv is Rp. 186,168, while the wheeling cost of 150 kv is Rp. 106,249. So that the total additional costs if there is wheeling in the D1 area of Rp. 11,746,307, detailed results in Table 4.  It is known that additional demand has caused an increase in amount of production, and transmission / distribution costs. The first scenario requires an additional fuel cost of Rp. 761,156. The cost of wheeling 500 kv is Rp. 4,688, while in 150 kv the cost increased to Rp. 15,420. So that the total additional costs if there is wheeling in the D4 area was Rp. 781,264 (detail calculation provided in Table  6).  The first scenario is cost minimization. This scenario required total cost of Rp. 17,748,035 to produce electricity as much as 35,704 MW. In scenario 2, emission minimization, the fuel cost expected was Rp. 41,528,235 in order to produce 35,165 MW electricity. In scenario 3, minimizing losses, we found that the system requires the fuel cost of Rp. 43,354,120 to produce electricity as much as 35,108 MW (detail calculation in Table 6).
Additional demand causes an increase in amount of production, and transmission / distribution costs. The first scenario requires an additional fuel cost of Rp. 761,156 with cost of wheeling 500 kv as much as Rp. 31,941. So that the total additional costs if there is wheeling in the D6 area of Rp. Rp. 797,785 (see Table 8)

Conclusion
Gencos gets the highest profit when serving basic load. Additional demand resulted additional revenue, however it caused decreasing in profit. This happens because additional revenue is not proportional to