Academic literature on the topic 'Hydro Unit Commitment'

In this paper, a novel approach of hybridization of two efficient metaheuristic algorithms is proposed for energy system analysis and modelling based on a hydro and thermal based power system in both single and multiobjective environment. The scheduling of hydro and thermal power is modelled descriptively including the handling method of various practical nonlinear constraints. The main goal for the proposed modelling is to minimize the total production cost (which is highly nonlinear and nonconvex problem) and emission while satisfying involved hydro and thermal unit commitment limitations. The cascaded hydro reservoirs of hydro subsystem and intertemporal constraints regarding thermal units along with nonlinear nonconvex, mixed-integer mixed-binary objective function make the search space highly complex. To solve such a complicated system, a hybridization of Gray Wolf Optimization and Artificial Bee Colony algorithm, that is, h-ABC/GWO, is used for better exploration and exploitation in the multidimensional search space. Two different test systems are used for modelling and analysis. Experimental results demonstrate the superior performance of the proposed algorithm as compared to other recently reported ones in terms of convergence and better quality of solutions.

As we move towards electrical networks with a growing presence of renewable generation, the representation of the electrical components becomes more important. In hydro-dominated power systems, modelling the forbidden zones of hydro plants becomes increasingly challenging as the number of plants increases. Such zones are ranges of generation that either should be avoided or are altogether unreachable. However, because representing the forbidden zones introduces a substantial computational burden, hydrothermal unit-commitment problems (HTUC) for large systems are usually formulated ignoring the forbidden zones. Nonetheless, this simplification may demand adjustments to the solution of the HTUC, because the generation of the hydro stations may fall in forbidden zones. In practice, the adjustments are usually performed based on the experience of system operators and, then, can be far from an optimal correction. In this paper, we study the impact of explicitly representing the hydro-generation forbidden zones in a large-scale system with more than 7000 buses, 10,000 lines, and 700 hydro units. Our findings show that the simplified model that is current used can deviate significantly from the model with forbidden zones, both in terms of the generation of hydro plants, as well as the generation of thermal plants and the system marginal costs.

In recent years the deregulation of energy markets and expansion of volatile renewable energy supplies has triggered an increased interest in stochastic optimization models for thermal and hydro-thermal scheduling. Several studies have modelled this as stochastic linear or mixed-integer optimization problems. Although a variety of efficient solution techniques have been developed for these models, little is published about the added value of stochastic models over deterministic ones. In the context of day-ahead and intraday unit commitment under wind uncertainty, we compare two-stage and multi-stage stochastic models to deterministic ones and quantify their added value. We show that stochastic optimization models achieve minimal operational cost without having to tune reserve margins in advance, and that their superiority over deterministic models grows with the amount of uncertainty in the relevant wind forecasts. We present a modification of the WILMAR scenario generation technique designed to match the properties of the errors in our wind forcasts, and show that this is needed to make the stochastic approach worthwhile. Our evaluation is done in a rolling horizon fashion over the course of two years, using a 2020 central scheduling model of the British National Grid with transmission constraints and a detailed model of pump storage operation and system-wide reserve and response provision. Solving stochastic problems directly is computationally intractable for large instances, and alternative approaches are required. In this study we use a Dantzig-Wolfe reformulation to decompose the problem by scenarios. We derive and implement a column generation method with dual stabilisation and novel primal and dual initialisation techniques. A fast, novel schedule combination heuristic is used to construct an optimal primal solution, and numerical results show that knowing this solution from the start also improves the convergence of the lower bound in the column generation method significantly. We test this method on instances of our British model and illustrate that convergence to within 0.1% of optimality can be achieved quickly.

Output power of hydro power plant was modelled and an optimization algorithm was implemented in a tool for optimizing hydro power plants. The tool maximizes power output of a hydro power plant by distributing water over a set of active units in the power plant which will be used in planning of electricity production. This tool was built in a MATLAB environment, using the optimization toolbox, and a GUI was developed for Vattenfall. The optimization tool was based on the same architecture as the current tool used for this kind of optimization which is to be replaced by the work presented in this thesis. Therefore, the goal was to achieve the same optimal results as the current optimization tool. Power output of three of Vattenfall’s hydro power plants were computed and two of these plants were optimized. These power output results were compared to results from the optimization tool currently used. This showed differences within the inaccuracy of measurements of ≤ 0.3%. These three power plants proved that the new tool is sufficient to replace the current tool but further testing is recommended to be conducted on more of Vattenfall’s hydro power plants to prove its consistency.

Le problème de gestion de production hydro-électrique (HUC) est un problème difficile, qui joue un rôle important dans la gestion de production électrique journalière à EDF. Dans cette thèse, nous étudions différents modèles et algorithmes pour résoudre un cas particulier du problème HUC à une usine (1-HUC). Etudier ce cas particulier présente plusieurs intérêts. D'une part, il existe des instances réelles à une usine mal résolues par les approches actuelles. D'autre part, cela nous permet d'étudier plus spécifiquement deux sources de difficultés séparément. L'une provient de la présence de non-linéarités, notamment la puissance qui est une fonction non-convexe et non-concave du débit d'eau et du volume des réservoirs. L'autre est due à l'ensemble des contraintes hydrauliques, notamment des volumes minimaux et maximaux ainsi que des volumes cibles des réservoirs.Dans une première partie, différentes alternatives de modélisation des non-linéarités du 1-HUC, plus particulièrement sur la puissance, sont proposées. L'objectif est d'identifier un modèle pouvant être résolu efficacement avec un bon compromis entre temps de calcul et précision. Les sept alternatives proposées couvrent un large spectre de familles de modélisation. Elles sont comparées sur un jeu d'instances présentant des variations sur cinq caractéristiques qui impactent la résolution. Cette étude comparative permet d'identifier trois types de modèles pertinents: un modèle avec des fonctions polynomiales de degré 2, un modèle avec une fonction linéaire par morceaux, et un modèle utilisant un ensemble fini de débits. Ce dernier modèle étant similaire à la modélisation actuelle à EDF, nous proposons dans la suite des algorithmes dédiés à celui-ci.Dans une deuxième partie, une étude polyédrale est proposée pour améliorer la résolution du problème 1-HUC. L'idée est de focaliser sur le coeur combinatoire, ce qui revient à considérer le lien entre les contraintes sur les volumes et l'ensemble discret des débits. Pour celà, nous définissons une variante du problème du sac-à-dos avec chaines de précédence et poids symétriques (SCPKP). Pour le SCPKP, nous définissons des conditions nécessaires de facettes, qui sont aussi prouvées suffisantes dans certains cas. Un algorithme de branch-and-cut en deux phases s'appuyant sur ces conditions et sur l'aspect symétrique du SCPKP est mis au point. L'efficacité de cet algorithme est ensuite montrée numériquement face à des algorithmes de l'état de l'art. Les résultats de cette analyse polyédrale du SCPKP, ainsi que l'algorithme de résolution proposé sont ensuite étendus au problème 1-HUC.Dans une troisième partie, une technique de résolution efficace est proposée pour prendre en compte les contraintes hydrauliques en s'appuyant sur la représentation du problème 1-HUC par un graphe. Cela permet de se ramener à un cas particulier du problème de plus court chemin avec fenêtres de ressource (RWSPP). Nous proposons deux algorithmes de graphes. Le premier algorithme est une variante exacte de l'algorithme A*, utilisant une borne duale dédiée au problème 1-HUC. En comparaison avec deux approches de l'état de l'art, nous montrons numériquement que cet algorithme est plus efficace pour traîter un cas spécifique du 1-HUC. L'objectif du second algorithme est de prendre en compte davantage de contraintes hydrauliques. Le principe s'appuie sur le concept d'optimisation bi-objectif pour lequel le second objectif correspond à une relaxation du volume d'eau. L'avantage par rapport à une optimisation bi-objectif classique est qu'il est possible d'utiliser les volumes minimaux et maximaux pour réduire l'espace de recherche et diriger l'énumération de solutions. Nous montrons numériquement, sur un grand jeu d'instances réelles, que cet algorithme est plus performant que trois approches de l'état de l'art. Même si cet algorithme a été conçu pour résoudre le 1-HUC, nous le définissons de manière générique pour tout RWSPP avec une ressource
The Hydro Unit Commitment problem (HUC) is a difficult problem playing a major role in the scheduling of daily electricity production at EDF. In this thesis, we study different models and algorithms to solve the special case of the single-unit HUC problem (1-HUC). Studying this case is relevant for the following reasons. On the one hand, there are real world instances of the 1-HUC problem which cannot be solved efficiently by current approaches. On the other hand, it makes it possible to study individually two particular sources of difficulty. One stems from the presence of non-linearities, in particular the power which is a non-convex non-concave function of the flow and the reservoirs' volume. The other is due to the set of hydraulic constraints, specifically the volume minimum and maximum bounds, as well as target volumes for the reservoirs.In a first part, modeling alternatives for the non-linear 1-HUC, focusing on the power function, are proposed. The aim is to identify a model which can be solved efficiently, with a good trade-off between computational time and precision. The seven proposed modeling alternatives cover a large panel of modeling families. These models are compared on a set of instances with variations on five features that impact the solution. This comparative study enables us to identify three efficient types of model: a model with polynomial functions of degree 2, a model with a piecewise linear function, and a model using a finite set of flows. As the latter model is similar to the current model at EDF, in the following we present algorithms dedicated to it.In the second part, a polyhedral study is proposed to improve the solving approach of the 1-HUC problem. The idea is to focus on the combinatorial aspects, which means considering the relationship between the bounds on volumes and the discrete set of flows. For this purpose, we introduce a variant of the knapsack problem, with Symmetric weight and Chain Precedences (SCPKP). For the SCPKP, we characterize necessary facet-defining conditions, which are also proven to be sufficient in some cases. A two-phase branch-and-cut algorithm based on these conditions and on the symmetric aspect of the SCPKP is devised. The efficiency of this algorithm is then shown experimentally against state-of-the-art algorithms. The results of this polyhedral study of the SCPKP, as well as the proposed algorithms, are then extended to the 1-HUC problem.In the third part, an efficient solving technique based on a graph representation of the 1-HUC problem is proposed, taking into account of the hydraulic constraints. It appears that the 1-HUC problem is a special case of the Shortest Path Problem with Resource Windows (RWSPP). We propose two graph algorithms. The first one is an exact variant of the A* algorithm, using a dual bound dedicated to the 1-HUC problem. In comparison with two state-of-the-art approaches, we show numerically that this algorithm is more efficient for handling a specific case of 1-HUC. The aim of the second algorithm is to take into account a wider set of hydraulic constraints. The idea is based on the concept of bi-objective optimization, for which the second objective corresponds to a relaxation of the volume. The advantage compared to a classical bi-objective optimization is that it is possible to use the minimum and maximum bounds on the volume to reduce the search space and to guide the enumeration of solutions. We show numerically, on a large set of real instances, that this algorithm outperforms three state-of-the-art approaches. Although this algorithm was designed to solve the 1-HUC problem, it is defined in a generic way for any RWSPP with a single resource

Wir betrachten ein Kraftwerkssystem mit thermischen Blöcken und Pumpspeicherwerken und entwickeln dafür ein Modell für den kostenoptimalen Wochenbetrieb. Auf Grund der Ungewißheit des Bedarfs an elektrischer Energie ist das mathematische Modell ein mehrstufiges stochastisches Problem. Dieses Modell beinhaltet viele gemischt-ganzzahlige stochastische Entscheidungsvariablen. Die Variablen einzelner Einheiten sind aber nur durch wenige Nebenbedingungen miteinander verbunden, welches die Zerlegung in stochastische Teilprobleme erleichtert. Diese stochastischen Teilprobleme besitzen deterministische Analoga, deren Lösungsverfahren entsprechend erweitert werden können. In dieser Arbeit werden ein Abstiegsverfahren für stochastische Speicherprobleme und eine Erweiterung der dynamischen Programmierung auf stochastische Probleme betrachtet. Die Lösung des dualen Problems führt zu Schattenpreisen, die bestimmte Einsatzentscheidungen bevorteilen. Die Heuristik zur Suche von primalen zulässigen Punkten wertet eine Folge von zugeordneten Economic-Dispatch-Problemen aus. Die Kombination der Einschränkung auf dual bevorzugte Fahrweisen (Lagrangian reduction) mit der Auswertung einer Folge von Economic-Dispatch-Problemen (Facettensuche) führt zu einem effizienten Verfahren. Die numerischen Ergebnisse an Hand realistischer Daten eines deutschen Versorgungsunternehmens rechtfertigen diesen Zugang.
We consider a power generation system comprising thermal units and pumped hydro storage plants, and introduce a model for its weekly cost-optimal operation. Due to the uncertainty of the load, the mathematical model represents a dynamic (multi-stage) stochastic program. The model involves a large number of mixed-integer (stochastic) decisions but its constraints are loosely coupled across operating power units. The coupling structure is used to design a stochastic Lagrangian relaxation method, which leads to a decomposition into stochastic single unit subproblems. The stochastic subproblems have deterministic counterparts, which makes it easy to develop algorithms for the stochastic problems. In this paper, a descent method for stochastic storage problems and an extension of dynamic programming towards stochastic programs are developed. The solution of the dual problem provides multipliers leading to preferred schedules (binary primal variables). The crossover heuristics evaluates the economic dispatch problems corresponding to a sequence of such preferred schedules. The combination of the restriction on dual preferred schedules (Lagrangian reduction) with the evaluation of a sequence (facet search) leads to an efficient method. The numerical results on realistic data of a German utility justify this approach.

碩士
國立臺灣海洋大學
電機工程學系
105
Ta-Chia River, as one of the most important river in Taiwan, has plentiful water resources and steep topography, and thus is suitable for hydro power generation. Ta-Chia River hydro power plant is accounting for about a quarter of hydro unit installed capacity and 40% of hydro unit generating capacity in Taiwan. Besides, Ta-Chia River hydro power plant can remove peak load and start to modulate frequency, and emergency backup to stabilize the power system, playing an irreplaceable role in the taipower hydroelectric power system. However, the spatial and tempora distribution of rainfall is uneven in Taiwan and the short distances of hydro power plants along the Ta-Chia River area, it is an important issue to use the limited water resources efficiently. In this study, we build a model of unit commitment with cascade hydro power plants to simulate with thermal units, and solving it by mixed integer linear programming. The systems we simulate are thermal units, combined-cycle units, and hydro units in Ta-Chia river hydraulic power plant, recording the system load by every 15 minutes in 24 hours. Systems including hydro units are far more complicated than Systems including thermal units. Except the unit operation characteristics of thermal units, a coordination of cascaded hydro units dispatch should also consider the requied amount of water, the limit of reservoir water level…etc. Under the premise that system operation security is satisfied, proving water resources use as far as possible and reduce fuel costs of thermal units can achieve maximizing the benefits by thermal units. To meet the requirements of secure operation of the system, this thesis analyzes cases of water requirement of downsterm, reservoir water level, and reserve capacity of Ching-Shan hydro units. The simulation results are justify the cases that can be applied in reality. Keywords: Unit Commitment with Cascade Hydro Power Plants, Mixed Integer Linear Programming, Unit Operation Characteristics

The randomness and intermittence of wind power require additional reserves provided by thermal generators. This creates difficult scheduling of generation, causing thermal generators to start up or shut down frequently, or to operate at low efficiency and high fuel consumption state. If not, some wind power will be curtailed and wasted. This paper examines the operation of a hybrid system made up of a wind farm, a pump storage hydro and conventional thermal generation units consisting in a combination of a heavy-duty and an aeroderivative gas turbines. In order to seek an optimal approach to deal with the uncertainty of increasing wind power and ensure both the efficient operation of thermal generators and full use of wind energy, two different control strategies have been proposed and compared: (i) a “custom” in house developed strategy and (ii) an “optimal” strategy based on Dynamic Programming. Using actual generation data of a wind farm, the operation of hydro power plant and gas turbines are obtained with the aim of compensating differences between actual wind generation and load demanded. Natural gas fuel consumption and average gas turbine efficiencies during the analyzed time period are calculated along with number of units starts-up.