Techno-Economical Analysis of Modified District Cooling System

This article proposes a modified district cooling system (MDCS) for the building clusters with a distributed cooling power plant installed at each of the building locations. This work focuses on the cooling power loss minimization in the chilled water pipe network along with the minimum cost of the water pipe and supply water pump. It forecast the peak cooling power requirement of a day for the different types of buildings by using a Long Short-Term Memory (LSTM) network. The network is trained with the previous record of the building's peak cooling power requirement pattern. The forecasted load decides the building to be a seller or buyer node on the basis of the cooling plant capacity installed at the location. An optimum water piping network layout between the seller and buyer node is obtained by following the Grey Wolf Optimization (GWO) algorithm. As a result, a different case study has been done which shows the impact of load forecasting to decide the number and length of pipes in the water distribution network. The article also shows that any random load increment affects the number of pipe connections which increases the cost of the water pipes.