Heating Demand Prediction Using Sensory and Event data
Summary of the entity:
The main areas of specialization of the RTU ITI Digital Innovation Hub are enterprise integration including advanced enterprise applications and business process improvement, development of scalable cloud solutions for big data processing, data mining and machine learning including applications in cybersecurity, IoT, intelligent transportation systems and biotechnology, modelling and optimization of complex systems, development of digital twins as well as digital transformation of organization by using advanced information technologies.
ITI DIH provides companies with cloud-hosted test and development environments for prototyping and implementing advanced ICT solutions.
Summary of the challenge:
Stakeholder: Municipality, energy companies, public venues, citizens
Heating is one of the most crucial services in cold weather countries. Municipalities are responsible for this service and contracts heating companies to provide the service. In many municipalities the heating systems are highly heterogeneous including both central heating and distributed facilities and clients also have varying abilities to control ambient conditions at their buildings. That is particularly relevant to municipalities and public venues experiencing significant variations in population and visitors due to seasonal or other effects. There are significant losses of heat in the heating network and many operational decisions are made in non-automated manner without knowing the external context. That results in inefficient decisions, inefficient energy consumption and increased emissions
The challenge is to collect data from the heating production facilities and the heating network as well as context data and to use this information to improve the accuracy of the heating demand. More importantly, demand for individual venues should be predicted using the calendar information on forthcoming events. This information is used to plan heat production and distribution with possibilities for optimization of parameters settings of the heating system in runtime using technologies like digital twin. The visualization of the predictions should be provided, which would help to detect problems and anomalies in a timely manner. The operational data also should be accumulated to support tactical and strategic decision-making concerning maintenance and construction (e.g. what is the optimal diameter of pipes).
External Data Provider:
Sensory data from heat production and distribution facilities: heat meters (time, energy, total flow, flow temperature, reverse flow temperature, temperature difference, operating time, flow, consumption), pressure sensors (time, pressure in both directions)
Sensory data from heating network clients: local heating systems (time, external temperature, operating mode, flow, desired flow temperature, return temperature, return temperature limit, comfort temperature, economy temperature, flow temperature for -30C, flow temperature for -15C, flow temperature for -5C, flow temperature for 0C, flow temperature for +5C, flow temperature for +15C, min flow temperature, max flow temperature, turn off temperature, week plan,
Open source datasets:
Open Source: Meteorological data and forecasts https://www.meteo.lv/meteorologija-datu-meklesana/
Open Source: Natural gas tariffs https://lg.lv/en/for-home/tariffs-and-calculator
Open Source: Calendar of public events, applicable regulations/restrictions, regular events https://www.kuldiga.lv/kalendars
Open Source: spatial web services from municipality geoportal https://gis.kuldiga.lv/about
At least one of the following should be addressed:
- Forecast of total heating consumption
- Heating demand for individual points in the distribution network and the dynamic plan for heating production
- Increased customer satisfaction concerning ambient conditions and public venues
- Input data for strategic planning of the heating network and its redesign