Substation renovation project demonstration takes into account both purchase and operational costs.
Since the major outages in Europe in 2004, grid system reliability has become a key topic as society now seeks improved standards of supply reliability. Substations are generally regarded as the sensitive parts of the grid system where reliability can only be improved by installing redundant equipment. The (N-1) rule is one criterion to guarantee a certain level of reliability, but it fails to provide guidelines to address the questions of where and how to install redundancies in substations. Furthermore, utilities consistently seek to optimize costs.
Introduced in 1933, the concept of life-cycle cost (LCC) can be applied for the assessment of substations together with a standard calculation procedure. Following the introduction of energy markets, the LCC assessment now has to take into consideration operational costs. System outages and the need to replace components that fail in service can result in costly penalties, undelivered energy and a reduction in profits.
A pilot project with the Swiss Federal Institute of Technology, ABB, AREVA T&D and National Grid was started in 2003 to develop a universal substation planning tool, which combines substation LCC assessment and substation reliability analyses. The mathematical and scientific methodology was developed in a high-voltage laboratory, and the practical knowledge was provided by the industrial partners.
Planning Substation Replacement
To design a substation and optimize LCC, it is necessary to consider the balance between investment and operational costs. Fig. 1 shows the installed switchgear selected as a result of the LCC components for a standard substation based on IEC 60300-3-6 LCC guide. A substation renovation project was chosen to demonstrate partial application and suitability of the developed methodology. To illustrate this application of LCC, consider the replacement of a 220/110-kV substation, as shown in the schematic in Fig. 2:
- Two incoming 220-kV bays
- One 220-kV busbar
- Two 220-kV power transformer bays
- Two power transformer groups 220/110 kV (160 MVA)
- Two 110-kV power transformer bays
- One 110-kV busbar
- Two outgoing 110-kV bays.
This equipment is based on air-insulated switchgear (AIS) inspected and maintained in accordance with the following time-based schedule:
- Visual inspection every year
- Maintenance every seven years
- Reconditioning every 14 years.
The new substation must fulfill the same transmission functions as the time-expired substation, but may differ in the layout, type of substation equipment and utility maintenance strategy. The table shows the time-based maintenance schedule for the equipment based on current technology, where the first maintenance strategy has the highest intensity and leads to the lowest component failure probability. Maintenance strategies two and three have longer time intervals, while strategies four and five make no provision for maintenance but have a higher component failure probability.
The LCC study assumes each component can be executed in AIS or gas-insulated switchgear (GIS) technology, can be installed as single, redundant or multi-redundant components and can be maintained in accordance with one of the five maintenance strategies. The new substation has two incoming and two outgoing bays. The substation layout that generates the lowest LCC is required.
The design team faces a multiplicity of possible substation solutions. Also, the individual cost parameters influence each other. For example, a substation with a high level of redundancy generates a high investment cost, and as all components need maintenance, that cost is also increased. Higher levels of redundancy improve reliability, minimizing outages and penalty costs. To optimize LCC, the design team must decide on the following:
- Redundant components to be installed
- Technology to be employed
- Maintenance strategy to adopt and apply.
This decision-making process can be simplified by executing an intelligent algorithm developed to identify the substation design solution with the lowest LCC.
The LCC calculation differs between the determinable cost (e.g., the investment, scheduled maintenance and renewal cost) and the probabilistic cost (e.g., the failure penalty cost and losses as a consequence of a failure) illustrated in Figs. 3 and 4. The latter requires a reliability analysis of the substation layout, which is provided by the method of reliability block diagram. Every substation component is defined by its specific failure rates (hazard curve). The reliability block diagram provides the failure rate of the entire substation layout with due consideration of all redundancies.
The substation design team can then calculate the LCC of each possible substation layout or, preferably, apply an optimization algorithm to save time in identifying the best solution. Different optimization algorithms have been analyzed to determine their application and suitability for the optimization of the substation LCC.
The genetic algorithm fulfills all requirements for the implementation. The substation algorithm has to identify each component with its LCC parameters:
- Acquisition cost
- Maintenance cost
- Parameters of the component hazard curve
- Component replacement time
- Parameters of different maintenance strategies.
The acquisition cost includes all the installation costs, namely the site, enclosure and electrical equipment. The maintenance cost includes the cost of the schedule detailed in the table. The component hazard curve is defined by the component failure rate based on CIGRÉ study committee reports and Weibull parameters presented as a component-specific file card (Fig. 5).
Intelligent LCC Optimization
In practice, the substation designer has to generate a component database with all available components for primary and secondary technology.
The genetic algorithm is an iterative algorithm, and at the start of this exercise, the substation designer is required to define the substation that needs to be optimized. The substation layout with all AIS elements is used as an example where each iteration has two main tasks: LCC calculation and variation of the substation layout (Fig. 6).
The variation task modifies the substation layout according to the components defined with component-specific file cards. As described in Fig. 4, the LCC calculation step determines the LCC value for each substation solution. The algorithm compares the LCC values and identifies which modification leads to a reduction in LCC. Application of this self-learning process, the algorithm modifies the substation to determine the LCC optimum. After several iteration steps, which the user can define, the algorithm provides the optimal composition of the substation solution.
At the start of the algorithm, the designer has to define which components are changeable and which are fixed by the existing facilities around the substation. In the actual case, the two incoming and two outgoing circuit bays were designated as fixed parameters.
The costs for the six optimization processes were considered in one model. Utilities with detailed cost data can plug them in this model as the methodology remains unchanged. Six simulations with different daily penalty cost factors (cost per day Kp) were undertaken with the following penalty cost values, Kp:
- Kp = 50
- Kp = 100
- Kp = 200
- Kp = 300
- Kp = 500
- Kp = 1000.
The third penalty cost factor of Kp = 200 seems very high in comparison with the component investment cost of a substation, but this extreme value was chosen to indicate when a utility has no alternative but to avoid an interruption in supply. Fig. 7 shows the results of the developed algorithm of the six simulations.
The optimized substation layout for penalty cost factors of Kp = 50, 300 and 1000 are shown in Fig. 8. The simulation with the lowest penalty cost factor, namely when Kp = 50, yields the most-simple layout. It consists of only single components in AIS technology (Fig. 8a). The low penalty cost factor offers the lowest system failure probability of the six cases.
The simulation with the penalty cost factor of Kp = 300 also provides a layout with single components, but the bays contain GIS technology (Fig. 8b). The higher investment cost of the GIS components compensates for the higher penalty cost with the lower failure rate.
The last simulation, with the highest penalty cost factor, provides a substation with three redundant power transformers and redundant GIS bays on the upper and lower voltage levels (Fig. 8c).
The simulation based on a penalty cost factor of Kp = 300 was relevant for the renovation of the substation. The algorithm calculated the investment strategy with the single GIS components (maintenance strategy three) and one power transformer group (maintenance strategy two) was optimum.
The genetic algorithm-based algorithm is a new approach to optimizing substation LCCs. It provides help to handle the number of cost parameters and provides a cost-optimized substation solution in a relatively short time frame, compared with the numerous cost calculations required in the traditional approach.
Furthermore, the algorithm can be used in the sensitivity analysis to identify the substation components with the highest impact on the total LCC. This approach offers a quick solution, replacing the decision-making processes normally used by experienced substation design engineers.
The author would to acknowledge the support received from AREVA T&D, ABB and National Grid for their financial support of this research project. Special thanks also to Professor Klaus Fröhlich of the High Voltage Laboratory at ETH Zurich who was the initiator and supporter of this research project.
Martin Hinow (email@example.com) received a master's degree in electrical power engineering in 2002 from Dresden University of Technology (Germany) and a Ph.D. in 2008 from the Swiss Federal Institute of Technology (Zurich, Switzerland). Currently, Hinow is working as a consultant system engineer at HIGHVOLT Prüftechnik Dresden GmbH, where he is responsible for the development and engineering of high-voltage testing systems.
|Maintenance strategy||Maintenance interval (in years)|
Enhanced Decision Support
ABB (Switzerland) www.abb.com
AREVA T&D (Switzerland) www.areva-td.com
National Grid www.nationalgrid.com