Two-step decision and optimisation model for centralised or decentralised thermal storage in DH&C systems
Description of the project
The objective of the project is to develop a decision and optimisation model for optimum dimensioning of centralised or decentralised thermal storage in DH&C systems.
A rational decision considering thermal storage within structures like district heating and cooling systems is based on the solution of a rather complex problem. The decision whether storage is needed or not, and of so, the optimised size of it, can thus be helped by a decision and optimisation model. Through the decision model an investment decision within the boundaries of a free market situation (power prices) can be improved regarding minimising the future economic uncertainties.
There are several reasons for considering or installing centralised or decentralised thermal storages in district heating or cooling systems.
These may be:
- bottlenecks in production
- reduction of peak load unit operation
- bottlenecks within the distribution network
- improved utilisation of waste heat or base load units
- time shift of heat versus electricity utilisation in case of CHP
- time variable energy prices in the actual market
The main priority is normally to produce and distribute the energy in the most cost efficient way. However, within the distribution network there can be parts where the utilised capacity is close to the maximum capacity. The lack of capacity in certain parts may have impact on the heat production and/or the temperature level of other parts of the network. By analysing the need and effect of centralised or decentralised storages these problems can be helped. The decentralised storage affects both the local capacity problem and its impact on the production. Local capacity problem may for instance occur if the expansion along a certain line exceeds the planned expansion. It may also occur in connection to heat exchanger stations along the network due to large altitude variations (e.g. DH-system in Trondheim, Norway).
Regarding heat production the system can benefit from a smoother heat production by utilising a storage. A storage can reduce the variation in the production and thus increase the utilisation time of the production units. Storage is also an alternative to local production capacity.
It may also be possible to take advantages of thermal storage in development of district heating in low-density areas.
STEP ONE MODEL
A simple method to solve the thermal storage dimension and existing problem is to look at historical data. Data for heat load and production prices (historical or scenario) are used to analyse the benefit of the thermal storage by comparing actual operation cost with the cost of running the system with a thermal storage. The cost differences can be considered to be the profit from the fictive thermal storage. The problem with this approach is that when you are analysing historical data you will always make "correct decisions" on a problem that is somewhat different from the actual problem. In reality you will have to make the decisions based on forecast of load and energy prices for different productions load and time of the day. A simple method can though be utilised to decide whether further actions towards analysing the effect of optimum storages shall be done. This method or model should be considered as a step 1 model. In certain situations a step 1 analysed result can be sufficient for a decision. The knowledge of when this situation occurs is presently not known.
STEP TWO MODEL
In a situation with only forecasts of heat load and energy prices for different productions load and time of day you will make more or less wrong decisions in the operation of the thermal storage. Depending on the quality of the forecast and distribution of the forecast error the optimum dimension of the thermal storage can be quite different from an optimum size based on historical data.
Thus, when a decision is made to investigate a storage, the size must be optimised subject to the market and load boundaries of the problem. Today there are no known methods or models available to handle this problem. (However operational optimisation model exists and they can be utilised in order to calculate operational cost for operating the system - with or without a specific storage size).
This is the background for the need of a step 2 optimisation method or model for storages.
Independent of the reason for installing a thermal storage in a network it may also be possible to utilise the thermal storage for production optimisation. This situation may, for instance, occur when certain storage is installed for capacity reasons. At off-peak load situation the storage can even be utilised to smoothen the heat production and/or increase the efficiency of cold production.
There is, however, a need to consider the storage both as a buffer and a production capacity to enable optimum operation of the whole system. The optimum operation of a storage versus other production units (on an hourly basis) is in itself a complex optimisation problem. This should not be mixed up with the optimum size and position problem, which contains both optimised operation (to calculate new operational costs) and respective additional investment costs (for the storage).
Summary of the final report of the project
The objective of this project has been to develop a decision and optimisation methodology for optimum dimensioning of centralised or decentralised thermal storage in DH&C systems.
This report presents a methodology for assisting the planning of introducing thermal storage into a DH&C plant. The methodology is divided into the solving of two sub problems; the existence problem, in this report referred to as the step one problem and the dimensioning problem referred to as the step two problem.
The report contains a study on the technical design of the storage and how the shape of the storage affects the efficiency and a survey on short term operational optimisation of DH&C plants as well as a discussion on the storage optimisation problem and how optimisation uncertainties affects the dimensioning.
The methodology for solving the existence problem, i.e. to find out whether a store should be further investigated or not, is presented. The methodology is based on historical data and the main idea behind this decision model is to study periods where the heat load is fluctuating around the maximum capacity of the base heat production. The necessity and appropriate size of a storage is evaluated by calculating the demand of energy above and below the actual production limit.
Solving the dimensioning problem is to find the optimum size of heat storage for a given district heating plant. A methodology for solving this problem is presented. Non-linear operational optimisation models from the literature survey are used to determine the optimal operation of the system (dispatch problem) and dynamic programming is employed for finding the optimal size of the storage (unit commitment).
Minimising the heat loss to the surroundings of a cylinder shaped storage implies an H/D ratio equal to 1,0. However, if the most important issue is to minimise the amount of useless volume the best H/D ratio is 2,0. However some additional height is inevitable due to the space required for diffusors and, if present, also for steam pockets. This is quite consistent with the existing storages in the Nordic countries where a majority of the stores have a H/D ratio range of 1,0 - 2,0.
A simplified method for calculating the economic benefits by using heat storage in different district heating systems is presented and demonstrated. The method is based on historical data. The calculations for some cases shows that the method is applicable as a first approach for an investment decision in heat storage.
A methodology for solving the storage dimensioning problem is presented together with a numeric example with a district heat load from a real DH system and where the heat is produced by a back pressure steam turbine CHP in combination with a oil fired heat boiler. For an annual heat load of 712 000 MWh the optimal heat storage volume is approximately 27 000 m 3 and the savings are about 3 % of the annual running costs.
Swedish National Testing and Research Institute and Sintef
Mr John Rune Nielsen
Mr Jacob Stang
SP Swedish National Testing and Research Institute, Sweden
Sintef Energy Research, Norway