Combined heat and power (CHP) production is a prominent technique for producing heat and electric power in an integrated process. While the produced heat is used locally, for district heating or for industrial processes, the electricity can be transmitted over long distances by the grid. The advantage of CHP is much higher energy efficiency than separate heat and power production. Therefore, CHP offers also significant potential for reducing emissions. At the same time, the global trend of increasing intermittent renewable energy forms is leading to imbalance between power supply and demand. While CHP production is difficult to adjust for non-coincident heat and power demand, the flexibility of a CHP system can be improved by including also separate power and heat production plants. Energy storages and power transmission between multiple areas can also improve the system flexibility and provide economic benefits. Cost efficient operation of such CHP systems can be determined by optimization models.
This study presents optimization models for different CHP planning problems, with different time horizons (short to long-term), single area and multiple areas, and with heat storages, power storages, or both. Depending on which components are included to the model, dedicated decomposition-based techniques have been developed for solving each model efficiently. An integrated model is decomposed into sub-models including both local and multi-area models. The first model is a multi-period local CHP model with heat storage. This model is solved by a generic linear programming (LP) algorithm, but a special extreme point formulation is applied in the modelling. The second model is an hourly multi-area model. A two-phase de-composition method including local and network models is proposed to optimize hourly multi-area energy production with power transmission. The third model extends the decomposition model with power storages for long-term problems. Last, an iterative process based on decomposition method is developed to include also heat storages. To produce realistic test data, a method was developed to generate heat demand with proper spatial and temporal variation also for areas where hourly historical data is unavailable.
The models have been validated by comparison with existing solution techniques for different problem sizes and time horizons. The results indicate that developed methods have high accuracy and fast solution time for long-term problems that is useful in solving hourly large-scale energy systems including thousands of variables. The speed advantage of the decomposition method improves with model size.The methods can be used for long-term planning of CHP system operation to support investment decisions, and for simulating system extension.
|Publication status||Published - 2019|
|MoE publication type||G5 Doctoral dissertation (article)|
- combined heat and power, optimization, linear programming, power transmission, energy storage