Modelling District Heating and Combined Heat and Power

Tingting Fang

Research output: ThesisDoctoral ThesisCollection of Articles


Combined heat and power (CHP) and district heating (DH) systems are becoming the main-stream technology due to its outstanding energy efficiency and environmental friendliness. In order to better use CHP and DH, we develop a set of models to manage CHP-DH systems. The models can be used sequentially or independently to serve different purposes. We first design a model for the DH systems to forecast heat demand based on the historical customer consumption and weather information. The model was built based on linear regression where outdoor temperature and wind speed predict heat demand. Accuracy of the model was significantly improved by adding different weekly rhythms of heat consumption to the model. The results indicate that the proposed linear regression model (T168h) using 168-hour demand pattern with midweek holidays classified as Saturdays or Sundays gives the highest accuracy among all the tested models. The forecasting model can be the premise of the production planning for DH or CHP systems. In parallel, another model is designed to estimate the water flows, temperatures, and heat losses in different parts of the DH network using automated hourly meter readings from customers. The model can be applied to arbitrary DH networks. According to the uncertainty analysis, the estimation model is robust with respect to the measurement uncertainty. Reliable and automatic DH network state estimation is a prerequisite for various operational network management and optimization tasks. On top of the abovementioned 2 models, optimization is developed in two directions: one direction is to optimize the heat production simultaneously at multiple heat plants at different locations of a DH network in order to minimize the production and distribution costs. The model is based on the static DH system model; it can determine the optimal supply temperatures at different heat plants and optimal load allocation between the plants. Because the objective function is a non-convex and non-smooth function of the decision variables, the genetic algorithm (GA) is applied to solve the problem. Optimization can result in savings in fuel and pumping costs. The other direction is to optimize the heat production between a CHP system and heat storage in order to minimize the net operating cost based on a sliding time window method which considers the uncertainty of the forecasts instead of assuming perfect information of heat demand and power sales price for the planning horizon. The model can be applied for both optimally operating heat storage and supporting investment planning for a new storage. The CHP optimization model based on sliding time window method can achieve 90% of the theoretically possible savings. Overall the key value of this thesis is to enhance the design and operation of CHP and DH systems.
Translated title of the contributionKaukolämmön ja yhteistuotannon mallinnus
Original languageEnglish
QualificationDoctor's degree
Awarding Institution
  • Aalto University
  • Lahdelma, Risto, Supervising Professor
  • Lahdelma, Risto, Thesis Advisor
Print ISBNs978-952-60-6843-5
Electronic ISBNs978-952-60-6844-2
Publication statusPublished - 2016
MoE publication typeG5 Doctoral dissertation (article)


  • district heating
  • combined heat and power
  • optimization
  • production planning
  • heat demand forecast
  • state estimation


Dive into the research topics of 'Modelling District Heating and Combined Heat and Power'. Together they form a unique fingerprint.

Cite this