Network Constrained Economic Dispatch of Integrated Heat and Electricity Systems through Mixed Integer Conic Programming

Shaojun Huang*, Weichu Tang, Qiuwei Wu, Canbing Li

*Corresponding author for this work

    Research output: Contribution to journalJournal articleResearchpeer-review

    447 Downloads (Pure)


    This paper proposes an economic dispatch method for an integrated heat and electricity system with respect to network constraints. Network constraints are usually nonlinear and can cause severe difficulties for optimization solvers. Particularly, in a heating network, the mass flow mixing at each node and the pressure and temperature drop along each pipe involve both hydraulic and thermal processes, which can cause high nonlinearity that has not been properly modeled. This paper will firstly model the power and heat network constraints by using a nonlinear model, which is accurate but hard to solve. Then, simplification and convexification will be employed to reform the nonlinear constraints to linear and conic ones. Consequently, the entire economic dispatch problem will be modeled as a mixed integer conic programming problem. Because the proposed model allows for the changes of the mass flow rate and direction, an optimal mass flow profile can be achieved along with the solution of the economic dispatch. Case studies on an integrated district heating and power system with a portfolio of power and heat sources show that the proposed economic dispatch model can handle the complexity of the network constraints and make optimal dispatch plans for multi-energy systems.
    Original languageEnglish
    Pages (from-to)464-474
    Number of pages18
    Publication statusPublished - 2019


    • Integrated heat and electricity system
    • District heating network
    • Combined heat and power (CHP)
    • Economic dispatch
    • Multi-energy systems (MES)
    • Mixed integer conic programming (MICP)


    Dive into the research topics of 'Network Constrained Economic Dispatch of Integrated Heat and Electricity Systems through Mixed Integer Conic Programming'. Together they form a unique fingerprint.

    Cite this