Meta-heuristics Applied to Power Systems

Electric Power Systems Research, 2010

This paper describes an approach to address the generation expansion-planning problem in order to help generation companies to decide whether to invest on new assets. This approach was developed in the scope of the implementation of electricity markets that eliminated the traditional centralized planning and lead to the creation of several generation companies competing for the delivery of power. As a result, this activity is more risky than in the past and so it is important to develop decision support tools to help generation companies to adequately analyse the available investment options in view of the possible behavior of other competitors. The developed model aims at maximizing the expected revenues of a generation company while ensuring the safe operation of the power system and incorporating uncertainties related with price volatility, with the reliability of generation units, with the demand evolution and with investment and operation costs. These uncertainties are modeled by pdf functions and the solution approach is based on Genetic Algorithms. Finally, the paper includes a Case Study to illustrate the application and interest of the developed approach.

International Journal of Electrical Power & Energy Systems, 2019

The fast construction times of projects based on variable generation technologies (VGTs) such as photovoltaic and wind generation, together with growing difficulties for building new transmission lines due to socio-environmental requirements, have opened new challenges in the development of sustainable power systems. Due to the complexity of the Transmission Network Expansion Planning (TNEP) problem, current models are usually oversimplified and do not always meet the requirements needed for the practical application. Examples of these simplifications are the use of reduced network equivalents, limiting the planning horizon to one or a few years and limiting the expansion options to adding new lines in given corridors. To meet the new challenges and achieve a time-effective increase of the transmission capacity for the integration of VGTs, improved models and algorithms capable of taking into account a higher degree of detail in the TNEP problem are needed. In this article, a novel meta-heuristic multi-year TNEP model based on Ant Colony Optimization (ACO) is presented. One of the main characteristics of the model is that it enables us to consider simultaneously further expansion options such as line reconductoring, voltage uprating, and adding series compensation to lines. We tested the proposed ACO model with the Garver's 6-bus test system and a modified version of the IEEE 118-bus test system, assuming a significant incorporation of VGTs. The results obtained for a 15-year planning task show (i) an excellent performance of the model in terms of quality of the obtained solution and computational times, compared to the traditional MILP approach, and (ii) including line uprating options within the multi-year TNEP brings significant benefits such as reducing the total investment and congestion costs of the system as well as the number of lines to be built.

18th International Conference and Exhibition on Electricity Distribution (CIRED 2005), 2005

This paper presents a new optimization procedure for multi-year distribution network planning based on dynamic programming, local network concept and mixed integer model. By introducing principles of dynamic programming and local network concept the multi-year planning problem of real-size networks is divided into sequences of sub problems with significantly reduced dimensions, enabling application of complex mixed integer programming model in each sub problem. The results obtained on several real-life distribution networks have shown that suggested procedure enables more efficient use of existing capacities and reduction of future expansion costs.

2011

This paper presents a novel planning approach which optimizes size and location of new transmission substation (TS) investments considering capacity expansion of the existing TSs based on primary distribution network investment requirements. Mixed Integer Programming (MIP) is utilized and the problem is decomposed into investment and feasibility check subproblems. The algorithm is formulated to minimize total investment cost while supplying the spatial forecasted load considering a set of system constraints. The results of two numerical examples indicate that presented algorithm is adequate for determining requirements of new transmission substations and/or capacity expansions together with new HV/MV lines, via appropriate selection of candidates. Visualization of the planning algorithm results, in an iterative manner, gives very important verification signals regarding the necessity of the proposed investments.

Technological and Economic Development of Economy, 2010

Although the problem of rational power generation has been extensively studied, traditional approaches for power optimization do not offer good solutions to this purpose, especially in a competitive electricity market environment where many factors are uncertain. In this paper, within the framework of two‐stage linear stochastic programming, the method for power planning has been developed, with uncertain factors taken into account, through a continuously distributed set of scenarios. The objective is to find the structure of the power plants capacity in the region which minimizes the sum of the investment and the expected operating costs over the long‐term planning horizon, taking into account the environmental impact. The structure of the considered task corresponds to a power investment planning problem that often arises in the developing regions. The method is developed for solving the stochastic optimization problem by the sequence of Monte‐Carlo sampling estimators. The proced...

Complexity

This work is dedicated to the economic scheduling of the required electric stations in the upcoming 10-year long-term plan. The calculation of the required electric stations is carried out by estimating the yearly consumption of electricity over a long-time plan and then determining the required number of stations. The aim is to minimize the total establishing and operating costs of the stations based on a mathematical programming model with nonlinear objective function and integer decision variables. The introduced model is applied for a real practical case study to conclude the number of yearly constructed stations over a long-term plan in the electricity sector in Jeddah City, Saudi Arabia. The current planning method is based only on intuition by constructing the same number of required stations in each year without searching for better solutions. To solve the introduced mathematical model, a novel recent gaining sharing knowledge-based algorithm, named GSK, has been used. The A...

Computer-Aided Civil and Infrastructure Engineering, 2004

This work shows the infrastructure investment decision is essentially a 0-1, nonlinear, multiobjective knapsack problem. It argues that, without making substantial simplification on some of the practical considerations, a conventional mathematical optimization approach may not be suitable for solving the problem. Further, the use of explicit heuristics may sometimes be unwieldy dependent on the system parameters. An innovative approach based on the concept of genetic algorithms is proposed in the work to tackle the nonlinearity in the decision problem. Based on the experimental results, the proposed approach is capable of producing good quality results. The key success feature of the proposed approach resides in the unique and innovative modeling of system parameters by genetic coding in the algorithms.

IET Conference Publications, 2010

Generation expansion planning gained a new dimension with the advent of electricity markets. It is now an activity decoupled from transmission and there are several agents competing to generate electricity and aiming at maximizing their individual profits. In view of this, it becomes more important to develop tools to help generation agents to build their expansion plans, internalizing several uncertainties in the model, an being able to simulate different possible reactions of the other competitors, given their impact in the profits of the agent being modelled. In this paper, we present a long-term decision aid tool that uses System Dynamics to model the long run of electricity markets together with Genetic Algorithms to solve the individual expansion problem of generation agents given their mixed-integer nature. Apart from the detailed description of the developed approach, the paper also includes a Case Study based on a four generation agent system to illustrate its application.

E3S Web of Conferences, 2020

In order to meet the requirements of precise investment in the distribution network under the new power reform, this paper proposes an optimization model for the investment decision of the distribution network. With the largest net present value rate and the smallest comprehensive cost of grid operation as the optimization goal, and certain capital constraints as constraints, a distribution network 0-1 programming model is established and solved by LINGO software; the analysis of an example shows that this model can ensure the safety of investment in the distribution network and the precise investment allocation in the distribution network under certain financial constraints.

2021 IEEE 33rd International Conference on Tools with Artificial Intelligence (ICTAI)

We design and develop an extensible model and a decision guidance system for making actionable recommendations on investments in heterogeneous infrastructure service networks. The model expresses the cash flows, as well as performance indicators, such as total cost of ownership and carbon emissions, as a function of both investment and operational controls within physical constraints of heterogeneous infrastructures and of balancing resource flows. Uniquely, it is designed to make Pareto-optimal investment decisions under the assumption of optimal operational controls over the time horizon. We also develop an extensible library of domainspecific operational analytic models for infrastructure components, initially for desalination and water systems, including pumps, renewable energy sources, water and power storage, and Revers Osmosis desalination units. Finally, we conduct and report on a feasibility study for this domain to demonstrate the ability to solve realistic size problems.

2013

Mathematical programming has long been an important tool for electric utility planners. This paper presents a survey of state-of-the-art mathematical programming methods as applied to electric power capacity expansion planning. The focus is on modelling features which make it possible to investigate important power system issues such as reliability, uncertainty and environmental impacts which linear programming models cannot address without considerable simplification. Solution methodologies are also described.

Socio-Economic Planning Sciences, 1984

This paper reviews the use. of multiobjective decision rules for solving power plant siting problems. After a discussion of exclusionary site screening methods for bounding the decision space, classes of multiobjective and goal programming decision rules are discussed in the context of final site selection. Advantages and limitations of these methods are highlighted. Although multiobjective decision rules have seen numerous applications to power plant siting in the literature, few electric utility companies have used these methods in practice. A review of the use of multiobjective methods in actual power plant siting decisions is also presented, and reasons for the paucity of real-world applications are suggested.

Design and expansion of distribution systems seems inevitable in view of the need to satisfy the rise in energy consumption in a technical and economical way. Optimal location, sizing and determining the service area of substations is one of the principle problems in expansion of distribution systems. Also uncertainty is one of the important factors that increase risk of exact decision makings. This paper presents a fuzzy multi-objective model for HV/MV substations planning so that uncertainties are modeled using fuzzy numbers (trapezoidal form). The proposed fuzzy model is based on the risk of economic and technical objectives as well as fuzzy values of investment, operation and loss cost of the substations and primary feeders. This model determines the optimal time, location and size of substations using a multi-objective genetic algorithm (NSGA-II). The proposed model is applied on a typical distribution system to assess the efficiency of the approach.

–The problem of power system planning, due to its complexity and dimensionality aspects, is one of the most challenging aspects facing the electric power industry in developing as well as developed countries. The proposed work will attempt to describe how these aspects are analyzed and assessed based on two major considerations, namely, reliability and cost. A case study considers two separate systems in a fast-developing country, each of which must be reinforced to meet the future predicted loads. The benefits of reinforcing separately or reinforcing by interconnecting the two systems are demonstrated. Uncertainties having a significant impact upon the decision-making process in the planning process are also addressed.

Energies, 2021

Transmission expansion planning (TEP), the determination of new transmission lines to be added to an existing power network, is a key element in power system planning. Using classical optimization to define the most suitable reinforcements is the most desirable alternative. However, the extent of the under-study problems is growing, because of the uncertainties introduced by renewable generation or electric vehicles (EVs) and the larger sizes under consideration given the trends for higher renewable shares and stronger market integration. This means that classical optimization, even using efficient techniques, such as stochastic decomposition, can have issues when solving large-sized problems. This is compounded by the fact that, in many cases, it is necessary to solve a large number of instances of a problem in order to incorporate further considerations. Thus, it can be interesting to resort to metaheuristics, which can offer quick solutions at the expense of an optimality guarant...