1. About the project
Project “Methodological Framework for Efficient Energy Management by Intelligent Data Analytics” is financed by Croatian Science Foundation (IP-2016-06-8350). The project’s duration is four years (January 1, 2017 – December 31, 2020).
The project deals with developing a methodological framework for efficient energy management by intelligent data analytics with the focus on machine learning methods and simulation modelling. The research aims to scientifically contribute the realization of European Commission directives about reducing greenhouse gas emissions, increasing energy efficiency and using 20% of energy consumption from renewable energy resources until 2020.
In Croatia and in other EU countries there are Strategies of energy development as well as National plans of energy efficiency, which quantify and control the objectives of reducing immediate energy consumption. However, the data on energy efficiency have not been scientifically analyzed enough for the purpose of efficient management of energy consumption and cost reduction, while there is a lack of research that use machine learning methods to more precisely detect interdependence among variables, prediction of payback period and other analytics. The purpose of this project is to conduct an intelligent data analysis on public buildings energy efficiency, and to suggest methods and models that will enable better planning of national energy policy and energy cost in public sector buildings.
The project suggests a methodological framework based on machine learning methods such as neural networks, decision trees, cluster analysis, association rules, and other methods that could be used for intelligent efficient management of energy consumption and energy supply cost. The methods will be tested on data that describe energetic characteristics of buildings, on the data used in the process of planning and implementing measures for improving energetic characteristics of buildings, and on the data describing their energy consumption. In relation to the energy supply, the project will be focused on data describing the supply chain of natural gas, as one of the major energy source in the public sector, aiming to find possible improvements in its efficiency.
The assumption is that a combined usage of various machine learning methods as well as simulation modelling can lead to lower energy consumption, higher efficiency of energy supply chain management, lower energy costs, more accurate evaluation of investment payback period, and better environment protection by lowering emission of harmful gasses. In order to develop such a methodological framework, it is necessary to investigate which methods of intelligent data analytics produce successful models for explaining and predicting energy consumption and suply, and how to integrate them, which is the topic of this project.
Project manager and Principal investigator was prof. Marijana Zekić-Sušac, PhD, and the members of the research team were prof. Josip Mesarić, PhD, prof. Rudolf Scitovski, PhD, prof. Domagoj Sajter, PhD, prof. Davor Dujak, PhD, prof. Hrvoje Krstić, PhD, Zlatko Tonković, PhD, Dario Šebalj, PhD, Adela Has, Saša Mitrović, Marinela Knežević, Sanja Scitovski and Kristina Hodak.
For more project details, visit: http://merida.efos.hr/
2. Which OR area is covered in your research, and how do you use examples/data from the business sector? Provide some examples of your work or studies. What software tools do you use?
For the purpose of the research, various OR areas were covered, such as: Business Analytics, Data Mining, Forecasting, Mathematical Modeling, Project Management, Simulation, Supply Chain Management.
Data needed for the project research are collected from the state institutions, primarily:
- Information system for energy management (ISGE),
- National Energy Efficiency Authority, from their System for measuring and verifying energy savings (SMIV),
- Environmental Protection and Energy Efficiency Fund (other relevant data on measures and supports from)
- Public firm from the energy sector (natural gas) – Hep Plin d.d.
The data from ISGE system include characteristics of Croatian public buildings from the energy certificate supplement reports, and also dynamic data on real energy consumption. The data from SMIV system include planned and implemented measures for increasing efficiency in public buildings, estimated savings, estimated reduction of CO2 emission, and estimated costs of implemented measures. In addition, other potentially relevant variables are used: meteorological data, season as a fuzzy variable, and others.
Software tools that have been used: Arena Simulation (for the development of the natural gas supply chain simulation model), R and Statistica (for machine learning), MatLab (for clustering), MS Visio (for supply chain mapping).
Examples:
- Predicting Energy Cost of Public Buildings by Artificial Neural Networks, CART, and Random Forest
The paper deals with modeling the cost of energy consumed in public buildings by leveraging three machine learning methods: artificial neural networks, CART, and random forest regression trees. Energy consumption is one of the major issues in global and national policies, therefore scientific efforts in creating prediction models of energy consumption and cost are highly important. One of the largest energy consumers in every state is its public sector, consisting of educational, health, public administration, military, and other types of public buildings. A real data from Croatian public sector is used in this paper including a large number of constructional, energetic, occupational, climate and other attributes. The algorithms for data pre-processing and modeling by optimizing parameters are suggested. Three strategies were tested: (1) with all available variables, (2) with a filter- based variable selection, and (3) with a wrapper- based variable selection which integrates Boruta algorithm and random forest. Prediction models of energy cost are created using two approaches: (a) comparative usage of artificial neural networks and two types of regression trees, CART and random forest, and (b) integration of RF-Boruta variable selection and machine learning methods for prediction.
- Machine learning based system for managing energy efficiency of public sector as an approach towards smart cities
Energy efficiency of public sector is an important issue in the context of smart cities due to the fact that buildings are the largest energy consumers, especially public buildings such as educational, health, government and other public institutions that have a large usage frequency. However, recent developments of machine learning within Big Data environment have not been exploited enough in this domain. This paper aims to answer the question of how to incorporate Big Data platform and machine learning into an intelligent system for managing energy efficiency of public sector as a substantial part of the smart city concept. Deep neural networks, Rpart regression tree and Random forest with variable reduction procedures were used to create prediction models of specific energy consumption of Croatian public sector buildings. The most accurate model was produced by Random forest method, and a comparison of important predictors extracted by all three methods has been conducted. The models could be implemented in the suggested intelligent system named MERIDA which integrates Big Data collection and predictive models of energy consumption for each energy source in public buildings, and enables their synergy into a managing platform for improving energy efficiency of the public sector within Big Data environment. The paper also discusses technological requirements for developing such a platform that could be used by public administration to plan reconstruction measures of public buildings, to reduce energy consumption and cost, as well as to connect such smart public buildings as part of smart cities. Such digital transformation of energy management can increase energy efficiency of public administration, its higher quality of service and healthier environment.
- Simulation model of natural gas supply chain in a function of costs optimization: the case of Croatia
This paper deals with the problem of balancing the transmission system in Croatia. The natural gas market in the Republic of Croatia functions on a balancing principle. This means that the total amount of natural gas injected into the transmission system (the amount which the suppliers nominated) must also be withdrawn from it. Since that it is impossible to precisely predict future consumption, certain deviations between the nominated amount and the amount consumed are possible. The system is then put into an imbalance and the price of its rebalancing is paid by the suppliers. The basic purpose of this paper is to create a simulation model for testing the potential technical solution which would compensate for the prediction errors. The suggested technical solution presents a special electromotor valve which would control the accumulation of the distribution system. The input data for the model was received from the supplier/distributer of natural gas in Croatia and represents the values of consumption and nominations on an hourly level for the entire year of 2017. The results of the simulation experiment have shown that the introduction of the suggested solution would reduce the positive rebalancing energy by 32%, and the negative one by 34%.
3. Using research results in teaching
On the graduate level of the study, students can learn about various OR areas covered in the project. For example, there are courses such as: Decision support systems, Business process management, Business simulations, Supply chain management… The students were also participating in this project. One student used the data collected for writing of master thesis. Research results related to the natural gas supply chain can be used in the course Supply chain management for bullwhip effect explanation or energy supply chain mapping.
