Framework for dynamic optimization of district heating systems using Optimica Compiler Toolkit

Abstract

Recent studies show that district heating infrastructures should play an important role in future sustainable energy systems. Tools for dynamic optimization are required to increase the efficiency of existing systems and design new ones. This paper presents a novel framework to represent, simplify, simulate and optimize district heating systems. The framework is implemented in Python and is based on Optimica Compiler Toolkit as well as Modelon's Thermal Power Library. The high-level description of optimization problems using Optimica allows flexible optimization formulations including constraints on physically relevant variables such as supply temperature, flow rate and pressures. The benefit of new algorithms for symbolic elimination in Optimica Compiler Toolkit is also investigated. The framework is applied on a test case, which is based on a planned city district located in Graz, Austria. The results demonstrate the generality of the representation as well as the accuracy of the simplification for dynamic optimization of temperature supply and pressure control.

Equation-based modelling for dynamic optimization of district scale energy systems – a scalability study

“…The IPOPT total time and the wall-clock time scale linearly with the number of consumers. A comparison with the results in [10] shows that with the current setting (i) the total IPOPT time is between 5 and 6 times faster, (ii) the IPOPT CPU time is between 3.5 and 6.2 times faster and (iii) the wall-clock time is between 1.7 and 2.2 times faster. However, we must stress that these two studies were not carried out on the same computer nor on computers with identical computing power.…”

“…The number of states as well as the number of algebraic variables scales linearly with the number of consumers. Note that the benchmarks differ slightly from Table I; the reason for this is a better comparability with [10]. OCT supports a block-lower triangular transformation of the DAE to identify algebraic variables that only depend affinely on the corresponding block variables; further variables can be eliminated by applying tearing to handle nonlinear dependencies.…”

“…• A scalability study of Modelica-based dynamic optimization of a building using a single shooting method implemented in the Toolchain for Automated Control and Optimization (TACO). • A scalability study of Modelica-based dynamic optimization of a district heating system using direct collocation implemented in the OPTIMICA Compiler Toolkit (OCT); furthermore, the performance of the new version is compared with the older version used in [10]. In the literature, the comparison of various methods and algorithms is generally referred to as optimization benchmarking; the authors are aware that the comparison of different optimization methods and algorithms is a complicated task that necessitates many subtle considerations in order to yield a fair and unbiased evaluation [11].…”

The Impact of the Future Fluctuating Energy Consumption on a District Heating System: Case Study of Omsk and Krasnoyarsk Cities

“…Limitations of the most of the regulation methods are that they only consider dynamics to a very limited extent. However, the simulation time in [39] is considerably reduced for model predictive control of real-time applications.…”