A Brief Literature Review of Structuring District Heating Data based on Measured Values

Abstract

This study reviews existing literature regarding district heating (DH) data and its clustering. In district heating, heat is produced at a central plant and supplied via a pipeline network to consumers (for example: homes, businesses, and industrial facilities). Different approaches were used—some based on consumption data and others based on heat load/demand data. Common methodology was researched and checked. New methods were double checked if reused in related work or developed single purpose only. Most databases are highly susceptible to being inconsistent, incomplete (lacking attribute values), and/or noisy (containing errors or outlier values). The major obstacle to obtain knowledge is poor data. It is necessary, therefore, to ensure that the knowledge discovered from the databases is, in fact, reliable. The PRISMA flow chart was applied to screen over 60 articles and to perform the literature review. As a result, 12 papers were identified dealing with the structuring of district heating data—almost all use either K-means methodology directly or another methodology based on K-means. Additionally, this study identified a research gap regarding eastern Europe in the data used and descriptions of applied methods.