Economic Production Optimization of a Power Plant with Constraints
Author
Thorsteinsson, Simon
Term
4. term
Education
Publication year
2015
Submitted on
2015-06-03
Pages
110
Abstract
Dette projekt undersøger, om Model Predictive Control (MPC) kan øge indtjeningen ved at planlægge produktionen i et kraftvarmeværk (Combined Heat and Power, CHP), der leverer fjernvarme og el. Arbejdet er et casestudie af Horsens Kraftvarmeværk. MPC udarbejder en produktionsplan ved at løse et optimeringsproblem, der tager højde for anlæggets model, elpriser, efterspørgsel efter fjernvarme og produktionsomkostninger. Anlægget er beskrevet med en hybridtilstandsmodel på et højt abstraktionsniveau. MPC’en er testet og evalueret i simulation med empiriske data fra DONG Energy. Der er gennemført to analyser: en deterministisk analyse (med antagelse om perfekt viden om fremtiden) og en prognosebaseret analyse (med brug af forudsigelser). Resultaterne viser, at den prognosebaserede analyse opnåede 99% af den indtjening, som den deterministiske analyse angiver. Det er positivt, fordi den deterministiske analyse fungerer som en teoretisk øvre grænse for MPC-algoritmens præstation.
This project examines whether Model Predictive Control (MPC) can improve profits by planning production in a Combined Heat and Power (CHP) plant that supplies district heating and electricity. The work is a case study of Horsens Kraftvarmeværk. The MPC controller creates a production plan by solving an optimization problem that accounts for the plant model, electricity prices, district heating demand, and production costs. The plant is represented by a high-level, hybrid state model. The MPC was tested and evaluated in simulation using empirical data provided by DONG Energy. Two analyses were performed: a deterministic analysis (assuming perfect knowledge of future conditions) and a forecast-based analysis (using predicted values). The forecast-based approach achieved 99% of the profit of the deterministic analysis, which is encouraging because the deterministic analysis represents a theoretical upper bound for the MPC algorithm.
[This abstract was generated with the help of AI]
Keywords
Documents