After successful completion of the course, students are able to classify the according to learning contents presented optimisation- and forecasting-methods with regard to their basic features concerning details of modelling, (simplifying) approximations, data requirements for that and accuracy of results to be expected. With the help of these knowledges they can differentiate the approaches with respect to their advantages or shortcomings in practical application and such assess their suitability for problems in electric enenrgy supply. By this it is easier for students to construct wellsuited solution strategies, to implement themselves computer programs or to select appropriate standard-software with desired accuracy of results as well as to perform additional supporting search of references in literature - what may be useful for their (future) work or research.
With the scope on electric energy supply and power economics the two topics optimization and (load-)forecasting are treated in closer detail:
After an outline of basic definitions in optimization and fundamental formulation of constrained optimization problems (modelling of goal function, inclusion of constraints, interpretation of dual "Lagrange" parameters for this purpose) various optimization approaches are presented: non linear and linear concepts. mixed-integer programing plus branch and bound. derivation free search methods (search by chance and inclusion of stochstic learning process with evolutionary or genetic algorithms), dynamic programing, decomposition by Lagrange'-Relaxation, multiobjektive optimization, decision under uncertainty with risk assessment. Provided examples of application are economic load dispatch, unit committment, extension planning, elements of hydro-thermal power system modelling.
As prerequesit for optimization an outline of (load-)prediction approaches is provided: time series forecasting by univariate autoregressive methods (exponential smoothing,ARIMA-modells, state space estimation approach by KALMAN-filtering), multi-variate causal models by multple regression, cluster-analysis for daily load curve pattern recognition, neural-network structures, inclusion of fuzzy-logic concepts.
Based on the written script to the course and on visualization with illustrative figures by means of slide presentations during the lectures optimization- and prediction-approaches as listed in the learning content are described. Starting in particular with the theoretic algorithmic numeric concept and fundamental structure of (iterative) procedure afterwards basic features are outlined with the help of exemplaric implementations in pratical cases from references in literature or research works. Discussed are advantages or shortcomings in practical application, data reqirements, necessary (simplifying) approximations in modelling and algorithmic process as well as accuracy in results to be expected. The somewhat smaller number of attending students allows to deal spontaneously with questions araised by students.
(A script to the lecture is available during the course from the lecturer.)
(price: 15 €)
Notes to references in literature on the subject are provided in the script or given by the lercturer during the course.
cause of Corona-restrictions the script is sent to registered students via email-atachment on request
please give your email-adress
The evidence of achievement has to be shown by the reply to questions of an oral examination talk whereby knowledge about optimization- or forecasting-methods,understanding of the tasks in electric energy supply and the features as well as solution procedure of approaches have to be demonstrated - eventually supported by a graphic visualisation of the answer.
(Bei der Vorbesprechung im Übungsraum CF0426)
wegen der Corona-Beschränkugen:
Anmeldung über meine Email-Adresse: mueller@ea.tuwien.ac.at
mit Angabe von Namen, Matrikel-Nr., Kennzahl (Studium) und Email-Adresse
Umfangreiches Skript wird für angemeldete Hörer über meine Email-Adresse (als Atachment) bereitgestellt - bitte zur Anforderung dazu um Angabe der Email-Adresse
Lecture notes for this course are available.
basic knowledge about problems in electric energy systems (power generation, control and transmission) as well as basic knowledge about higher mathematics (especially about numerical algorithms), basic knowledges in digital data processing and programming methods.
