The activities of Data Analytics and Optimization Laboratory (DataOptima Lab) are focused on applications of internet of things (IoT), data science, and optimization methods in the energy sector. The principal research methodology of the lab is the development of hybrid data-driven/physical phenomena based models for simulation and employing stochastic algorithms for optimization of energy systems. The key framework of industrial activities of the laboratory is instead the utilization of Internet of things technology, data-driven models, and predictive control for performance optimization.

 

 

Industrial Customers/Collaborators

 

 

Academic Collaborators

News

  • New Lab Member | Sept 25th 2020

    Alireza Zarghamnezhad has joined our lab, he will work on Data Engineering and deployment of machine learning models for the management of HVAC systems

  • New Lab Member | Sept. 17th 2020

    Giulia Moret has joined our lab; Giulia's M.Sc thesis is focused on ML based indoor thermal behaviour modelling

  • New Publication | Sept 15th 2020

    Our recent article "Machine learning based disaggregation of air-conditioning loads using smart meter data" (collaboration with HVL Norway) is now online

  • New Lab Member | Sept 12th 2020

    Alessandro Benetti has joined out lab; Alessandro's activities, conducted in collaboration with multiphase flow lab, will be focused on Machine learning based estimation of flow regimes in two-phase flows

  • New Lab Member | Sept. 7th 2020

    Farzad Dardras Javan has joined our lab; the activities of Farzad's thesis are dedicated to data-driven modelling of indoor environments and TCL based demand side management

  • New Publication | Sept 15th 2020

    Our recent article entitled "MOIRAE–bottom-up MOdel to compute the energy consumption of the Italian REsidential sector: model design, validation and evaluation of electrification pathways" (collaboration with RSE) is now online

  • New Lab Member | August 1st 2020

    Shayan Milani has joined out lab; Shayan' activities, that are carried out in collaboration with multiphase flow lab, will be dedicated to the Machine learning based estimation of heat transfer in two-phase flows

  • New Lab Member | April 1st 2020

    Hamed Khatam has joined out lab; Hamed' activities, carried out in collaboration with the Idiap research institute, will be dedicated to the

Prof. Fabio Rinaldi

Associate Professor, Laboratory Director, and Head of System Optimization Section

Research Area

Multi-objective Optimization of Energy Systems, Evolutionary Optimization, Fuel Cell based Systems

fabio.rinaldi@polimi.it

Dr. Behzad Najafi

Assistant Professor(RTDa) and Head of the Data Science Section

Research Area

Data-driven Building simulation, Energy Data Analytics, Physical modelling/long-term performance optimization of energy systems

behzad.najafi@polimi.it

Prof. Reza Arghandeh

Full Professor in Machine Learning and Big Data at HVL Norway and Main External Research Collaborator

Research Area

Distributed Control, Data Analytics, Cyber-Physical Systems Resilience

Reza.Arghandeh.Jouneghani@hvl.no


Marco Tognoli

PhD Student

Research Area

Machine learning based modelling of district heating systes - Dynamic simulation of industrial boilers, model predictive control

marco.tognoli@mail.polimi.it

Debayan Paul

Research Assistant

Research Area

Physical-phenomena based modelling of irradiation and fenestration behaviour

debayan.paul@mail.polimi.it

Keivan Ardam

Senior Research Assistant (Data Science)

Research Area

Machine learning based modelling of two-phase flows - Data-driven Building Simulation

keivan.ardam@mail.polimi.it

Hamed Khatam

Research Assistant

Research Area

Machine learning based estimation of occupancy status and indoor air quality

hamed.khatam@mail.polimi.it

Giulia Moret

Research Assitant

Research Area

Machine learning based indoor environment modelling

giulia.moret@polimi.it

Farzad Dadras Javan

Research Assistant

Research Area

Machine learning based simulation of indoor thermal behaviour - TCL based demand side management

farzad.dadrdas@mail.polimi.it

Alessandro Benetti

Research Assitant

Research Area

Machine learning based estimation of flow regimes in two-phase flows

alessandro1.benetti@mail.polimi.it 

Shayan Milani

Research Assistant

Research Area

Machine learning based modelling of two-phase flows

shayan.milani@mail.polimi.it

Alireza Zarghamnezhad

Research Assitant (Data Engineering)

Research Area

Machine learning based modelling of indoor environment's thermal behaviour

alireza.zarghamnezhad@outlook.com

Giordano Bruno Zannini

Research Assistant

Research Area

Machine learning based estimation of NoX Emissions

giordano.zannini@mailpolimi.it

Alumni


Pegah Mottaghizadeh

Current Position

PhD Student at University of California Irvine

M.Sc. Thesis Title

Process system modeling of reversible solid oxide cell (rSOC) energy storage system

Mehrdad Biglarbeigian

Current Position

Senior Reliability Enigneer at Tesla

M.Sc. Thesis Title

Design, implementation and evaluation of cooperative methods for dual demand side management

Paolo Bonomi

Current Position

Associate in Intelligent Automation - PWC

M.Sc. Thesis Title

Machine learning based fault diagnosis and performance estimation of automotive PEM fuel cells through optimal EIS tests

Alireza Haghighat Mamaghani

Current Position

Postdoctoral Researcher at University of Alberta

M.Sc. Thesis Title

Simulation, optimization and long term performance analyses of an HT-PEM fuel cell based micro CHP plant

Sadaf Moaveninejad

Current Position

Postdoctoral Researcher at Free University of Bozen Bolzano

M.Sc. Thesis Title

Book Chapter: Data Analytics for Energy Disaggregation: Methods and Applications

Pedro Obando Vega

Current Position

PhD Student at TU Darmstadt

M.Sc. Thesis Title

Numerical simulation of an incineration power plant employing OpenFoam

Andrej Hanušovský

Current Position

Data Scientist at Makers S.r.O

M.Sc. Thesis Title

Reproducible machine-learning physical-based models for pressure drop estimation in two phase diabatic and adiabatic flows

Farshad Bolourchifard

Current Position

AI Energy System Monitoring at BrainBox AI

M.Sc. Thesis Title

Application of deep learning in thermal load forecasting and data-driven supply optimization of a district heating network

Monica Depalo

Current Position

Business Analyst at A2A Energia

M.Sc. Thesis Title

Machine learning based estimation of commercial buildings characteristics : determining the most influential temporal features

Arun Shaju

Current Position

Sustaiability Consultant at Ela Green Buildings

M.Sc. Thesis Title

Machine learning based building characteristics and performance estimation through analyzing consumption profiles

Ratomir Dimikj

Current Position

Product Engineer at LeafTech GMBH

M.Sc. Thesis Title

Incremental machine-learning based load prediction of a university campus aiming at performance improvement on day-ahead market and reduction of losses

Nicolas Fernando Marrugo

Current Position

Data Scientist at SIRAM

M.Sc. Thesis Title

Deep learning based occupancy prediction and HVAC behavior modeling for improving energy efficiency of commercial buildings

Enoch Nuamah Appiah

Current Position

Research Assistant at DataOptimaLab

M.Sc. Thesis Title

Development of optimal machine learning based pipelines for predicting the dynamic thermal behavior of indoor environments

Danish Ahmad Mir

Current Position

Research Assistant at DataOptimaLab

M.Sc. Thesis Title

Development of optimal machine learning based pipelines for predicting the dynamic thermal behavior of indoor environments

Michela Silva

Current Position

Proposal Engineer at Siram

M.Sc. Thesis Title

Machine Learning Based Consumption Prediction and Hourly Optimization of Heating System for a Hospital Complex

Luca Di Narzo

Current Position

Energy Analyst at SIRAM

M.Sc. Thesis Title

Machine learning based estimation of air-conditioning loads through temporal features extraction using smart meter data

Farzad Moghaddampour

Current Position

Digitalization Engineer at Electrolux

M.Sc. Thesis Title

Feasibility analysis of renewable energy systems for rural electrification in different climatic zones of Peru

Lorenzo Benevento

Current Position

Project Manager at Siemens

M.Sc. Thesis Title

Energy auditing and proposing energy saving measures for an Italian SME through building energy simulation

Lidia Premoli vilà

Current Position

Intern at RSE

M.Sc. Thesis Title

Bottom-up modelling of Italian residential sector: Development, validation, and application in evaluation of decarbonisation scenarios

Darko Micev

Current Position

Research Assistant at DataOptimaLab

M.Sc. Thesis Title

Incremental machine-learning based load prediction of a university campus aiming at performance improvement on day-ahead market and reduction of losses

Farshad hasanabadi

Current Position

R&D Intern at EURAC Research

M.Sc. Thesis Title

A Machine Learning based Approach for PV Self-consumption Enhancement in an NZEB Using a Geothermal Heat Pump Driven Heating System

Giovanni D. Temporelli

Current Position

Intern at Edilclima s.r.l

M.Sc. Thesis Title

Data-driven dynamic modelling and implementation of an improved control strategy for a geothermal heat pump based heating system in a nearly zero energy building

Matteo Magni

Current Position

Technical Expert at Energy Power Technology SRL

M.Sc. Thesis Title

Modelling a steam reforming reactor utilized in PEM fuel cell based micro-generation System

Amin Solouki

Current Position

Research Assistant at Ecole Polytechnique de Montreal

M.Sc. Thesis Title

Modeling infrastructure and implementation of state-of-the-art heat transfer models for plate fin heat exchangers

Dario Bertani

Current Position

Researcher in RSE SpA

M.Sc. Thesis Title

Analysis of hybird off grid renewable electrification system for the case study of Val Codera

Veronica Galli

Current Position

Procurement Engineer at Mandelli

M.Sc. Thesis Title

energy and exergy analysis of a waste to energy plant and its R1 energy recovery efficiency evaluation

Book Chapters

Behzad Najafi, Sadaf Moaveninejad, Fabio Rinaldi,

Data Analytics for Energy Disaggregation: Methods and Applications

Chapter 17 of Big Data Application in Power Systems, Elsevier Science 2018, Pages 377–408

Link

Journal Articles

B. Najafi, L. Di Narzo, F. Rinaldi, R. Arghandeh

Machine learning based disaggregation of air-conditioning loads using smart meter data

IET Generation, Transmission & Distribution, DOI: 10.1049/iet-gtd.2020.0698

Link

F. Rinaldi, F. Moghaddampoor, B. Najafi, R. Marchesi

Economic feasibility analysis and optimization of hybrid renewable energy systems for rural electrification in Peru

Clean Technologies and Environmental Policy, 1-18

Link

B. Najafi, P. Bonomi, A. Casalegno, F. Rinaldi, A. Baricci

Rapid Fault Diagnosis of PEM Fuel Cells through Optimal Electrochemical Impedance Spectroscopy Tests

Energies 2020, 13(14), 3643; https://doi.org/10.3390/en13143643

Link

G. Besagni, M. Borgarello, L.P. Vilà, B. Najafi, F. Rinaldi

MOIRAE–bottom-up MOdel to compute the energy consumption of the Italian REsidential sector: model design, validation and evaluation of electrification pathways

Energy, 211, 2020, 11867

Link

M Manivannan, B Najafi, F Rinaldi

Machine Learning-Based Short-Term Prediction of Air-Conditioning Load through Smart Meter Analytics

Energies 10 (11), 1905

Link

AH Mamaghani, B Najafi, A Casalegno, F Rinaldi

Optimization of an HT-PEM Fuel Cell based Residential Micro Combined Heat and Power System: A Multi-Objective Approach

Journal of Cleaner Production 180, 2018, 126-138

Link

M Tognoli, B Najafi, F Rinaldi

Dynamic modelling and Optimal Sizing of Industrial Fire-tube Boilers for Various Demand Profiles

Applied Thermal Engineering,132,2018, 341-351

Link

AH Mamaghani, B Najafi, A Casalegno, F Rinaldi

Predictive modelling and adaptive long-term performance optimization of an HT-PEM fuel cell based micro combined heat and power (CHP) plant

Applied energy 192, 519-529

Link

AH Mamaghani, SAA Escandon, B Najafi, A Shirazi, F Rinaldi

Techno-economic feasibility of photovoltaic, wind, diesel and hybrid electrification systems for off-grid rural electrification in Colombia

Renewable Energy 97, 293-305

Link

M Aminyavari, AH Mamaghani, A Shirazi, B Najafi, F Rinaldi

Exergetic, economic, and environmental evaluations and multi-objective optimization of an internal-reforming SOFC-gas turbine cycle coupled with a Rankine cycle

Applied Thermal Engineering 108, 833-846

Link

AH Mamaghani, B Najafi, A Casalegno, F Rinaldi

Long-term economic analysis and optimization of an HT-PEM fuel cell based micro combined heat and power plant

Applied Thermal Engineering 99, 1201-1211

Link

B Najafi, S De Antonellis, M Intini, M Zago, F Rinaldi, A Casalegno

A tri-generation system based on polymer electrolyte fuel cell and desiccant wheel Part A: Fuel cell system modelling and partial load analysis

Energy Conversion and Management 106, 1450-1459

Link

B Najafi, AH Mamaghani, F Rinaldi, A Casalegno

Fuel partialization and power/heat shifting strategies applied to a 30 kWel high temperature PEM fuel cell based residential micro cogeneration plant

International Journal of Hydrogen Energy 40 (41), 14224-14234

Link

B Najafi, AH Mamaghani, F Rinaldi, A Casalegno

Long-term performance analysis of an HT-PEM fuel cell based micro-CHP system: operational strategies

Applied Energy 147, 582-592

Link

AH Mamaghani, B Najafi, A Shirazi, F Rinaldi

4E analysis and multi-objective optimization of an integrated MCFC (molten carbonate fuel cell) and ORC (organic Rankine cycle) system

Energy 82, 650-663

Link

B Najafi, PO Vega, M Guilizzoni, F Rinaldi, S Arosio

Fluid selection and parametric analysis on condensation temperature and plant height for a thermogravimetric heat pump

Applied Thermal Engineering 78, 51-61

Link

AH Mamaghani, B Najafi, A Shirazi, F Rinaldi

Exergetic, economic, and environmental evaluations and multi-objective optimization of a combined molten carbonate fuel cell-gas turbine system

Applied Thermal Engineering 77, 1-11

Link

B Najafi, AH Mamaghani, A Baricci, F Rinaldi, A Casalegno

Mathematical modelling and parametric study on a 30 kWel high temperature PEM fuel cell based residential micro cogeneration plant

International Journal of Hydrogen Energy 40 (3), 1569-1583

Link

A Shirazi, B Najafi, M Aminyavari, F Rinaldi, RA Taylor

Thermal economic environmental analysis and multi-objective optimization of an ice thermal energy storage system for gas turbine cycle inlet air cooling

Energy 69, 212-226

Link

M Aminyavari, B Najafi, A Shirazi, F Rinaldi

Exergetic, economic and environmental (3E) analyses, and multi-objective optimization of a CO2/NH3 cascade refrigeration system

Applied Thermal Engineering 65 (1-2), 42-50

Link

B Najafi, A Shirazi, M Aminyavari, F Rinaldi, RA Taylor

Exergetic, economic and environmental analyses and multi-objective optimization of an SOFC-gas turbine hybrid cycle coupled with an MSF desalination system

Desalination 334 (1), 46-59

Link

A Shirazi, M Aminyavari, B Najafi, F Rinaldi, M Razaghi

Thermal/economic/environmental analysis and multi-objective optimization of an internal-reforming solid oxide fuel cell/gas turbine hybrid system

international journal of hydrogen energy 37 (24), 19111-19124

Link

Socio-economic Clustering of Resdiential Buildings through smart-meter analytics

Status:On-going

In the first phase of this project, building physics based filters are employed in order to convert the temporal characteristics of electrical consumption time-series into features. In the second phase, these features alongwith the socio-economic characteristics of residential-buildings are employed as the training dataset of machine learning algorithms. Several machine learning algorithms are employed in order to classify the residential building based on the given consumption profile and the one with the highest accuracy while predicting the test data-set is chosen.

Short-term/Medium-term prediction of Air-conditioning loads through energy disaggregation

Status:On-going

This Project is focused on short-term prediction of air-conditioning (AC) load of residential buildings using the data obtained from a conventional smart meter. The AC load, at each time step, is separated from smart meter aggregate consumption through energy disaggregation methodology. The obtained air-conditioning load and the corresponding historical weather data are then employed as input features for the prediction procedure. In the prediction step, different machine learning algorithms, including Artificial Neural Networks, Support Vector Machines, and Random Forests, are used in order to conduct hour-ahead and day-ahead predictions

Machine Learning based Prediction of Pressure-drop in Two Phase Flows

Status:On-going

In the firt phase of the project the pressure drop of a two-phase adiabatic flow at a wide range of operating conditions is measured. Next, physical phenomena based pre-processing procedure is carried out, in which the operating conditions of each test are converted into related dimensionless parameters. In the last phase, utilizing the obtained dataset, different machine-learning (ML) algorithms are trained in order to predict the pressure drop (target) being provided the dimensionless parameters as input features. The accuracy of the algorithms are then evaluated through cross validation and the most accurate algorithm is determined. The developed ML based model will eventually be provided public access as an open-source tool.

Data-driven Real-time Fault Diagnosis of PEM Fuel Cells Through Electrichemical Impendance Spectroscopy

Status:On-going

In this project the results of Electrochemical Impedance Spectroscopy (EIS) are utilized in order to train machine learning algorithms which are later employed for rapid and robust diagnosis of faults within PEM fuel cells. In the first phase of the project, the EIS data of the cell at various operating conditions which represent the common faults are measured. Next,a parat of the obtained data is utilized as the training dataset and machine learning (ML) algorithms are accordingly trained using the corrresponding EIS data as inputs and the type of the fault as output. Next, the trained machine learning model is tested with the remaining faulty EIS data (test Dataset) in order to evaluates its fault classification accuracy.

Data-driven Short-term/Meidum-term prediction of electrical demand and heating consumption of commercial buildings

Status:On-going

In this project, hour ahead, day ahead preditions of both heating and electrical consumption of commercial buildings are conducted employing the available dataset for a large set of commercial buildings. Measured smart meter and heating management system data alongwith corresponding climatic data obtained from external resources are utilized as the training dataset. Combinations of machine learning algorithms alongwith LSTM methodology are evalauted in order to determine the most accurate algorithm with the least possible computation cost.

Dynamic modeling and optimal sizing of fire-tube boilers for varous demand profiles

Status:On-going

In this project, detailed dynamic model of an industrial fire-tube boiler is first developed and five different geometrical configurations, each of which corresponds to a boiler model, are considered. Next, a PID controller is implemented and tuned for each configuration aiming at controlling the steam pressure, while addressing a demand with a variable flow rate. The operation of the developed boiler models, while providing four different steam demand profiles, are next simulated. The resulting cumulative average efficiency along with the cumulative pressure deviations and minimum and maximum pressure levels, which are achieved in each simulation, are then determined. The obtained results provides practical information regarding the trade-off between the size of the boiler and its corresponding performance and controllability.

Developing a Big Data Analytics Tool for Large-scale Building Consumption Prediction using Spark

Status:On-going

In this project, in the context of hadoop eco-system, building energy consumption tools, which facilitate performing the operations in parallel on multiple nodes in a fault tolerant, manner, are developed. The developed tools provide the possibility of applying time-series prediction algorithms on a large set of building in order to predict the consumption of a specific district. Hadoop Distributed Files System (HDFS) is employed in order to store data on multiple nodes while Spark is utilized in order to automizing the mapping and reducing processes on multiple nodes. State of the art Machine learning algorithms implemented in python are instead used in order to implement time-series prediction.

Developing Optimization Algorithms over IoT based Building Management Systems for Commercial Buildings

Status:On-going

In this project, starting from related open-source platforms, an IoT based building management system is first developed which facilitates integration of various commercial building facilities. State-of-the-art Machine learning algorithms are employed to predict heating/cooling and electrical demand of the building in the next hour using it consumption history and weather predictions. Next, optimal operating conditions of the Air-conditioning units are determined and applied employing the developed IoT based BMS system. Similar methodologies are also employed for controlling Domestic hot water generation units and lighting systems.

Dynamic Modelling, Experimental Validation and Thermo-economic Analysis of Industrial Fire-tube Boilers with Stagnation Point Reverse Flow Combustor

Status:On-going

In this project, a detailed dynamic model of fire-tube boilers with Stagnation Point Reverse Flow type combustor is developed. The data obtained through an experimental testing procedure is then employed in order to validate the developed model at various operating conditions. Next, a PID controller is implemented and tuned in order to control the steam pressure while supplying steam with a variable flow rate. In the second part, a set of daily vapour request profiles is considered and the overall efficiency of the boilers with different sizes, while addressing the considered demand profiles, is obtained. The smallest boiler, which can provide the demand with an acceptable efficiency, is then determined. A similar procedure is carried out for determining the optimal size as investment venture of the boiler while considering several different daily load profiles.

StatisticalAnalysis of accelerated life test data for useful life prediction of residential boilers

Status:terminated

The useful life of residential boilers at various test-benches while undergoing different levels of increased stresses is first measured. Next, life distributions of boilers at different stress levels are determined. Finally, utilizing state-of-the-art statistical methods for analysing accelerated life test data, the characteristic useful life of boilers while operating at normal conditions is predicted. The implemented method facilitates predicting the life of products in a notably shorter time and consequently with significantly lower operating cost and test bench invesment.

Energy and Environmental Technologies for Building Systems

Program: Master of Science in Energy Engineering

The first part of this course is dedicated to fundamentals of building physics, which enable the students to calculate the buildings’ overall thermal demand, heating and cooling peak loads, and assess the energy performance of dwellings. In the second part, data-driven approaches for simulating the energetic performance of buildings are presented. The third part is focused on heating, cooling and air-conditioning technologies, including both the centralized and decentralized architectures, in order to address the calculated thermal demand. Finally, the last part of the course is devoted to the solar thermal systems, their characteristics, and their integration in buildings for supplying both the corresponding thermal demand and the required domestic hot water.

Building physics simulation and Data-driven Modelling

Python programming language alongwith Python scientific computing and data science modules are employed in this M.Sc. course for both building physics simulation and Data-driven modelling. Firstly, simple heat transfer through wall calculations and radiation heat transfer calculation are implemented utilizing basic python scripts. You can find the developed scripts together with a brief introduction to python programming language in this Github Repository. Next, Numpy and Pandas modules are employed to accomplish the task of reading from tables and conducting vectorized operations which are essential for calculating the incident solar irradiaton, heat transfer through windows and infiltration. Matplotlib modules is instead utilized in order to plot the variations of building's load with various construction characteristics. this Github Repository includes the above-mentioned implementations and an introduction to the mentioned modules. In the final step, in the context of data-driven building energy simulation, Pandas module is utilized in order to import the dataset of hourly consumption of a residential building within a year, exploring the correlation between the consumption and external climatic conditions and the seasonal parameters. Sci-kit learn module is finally utilized in order to implement Machine-learning based hour ahead prediction of residential energy consumption utilizing the features which have been found to be influential in the previous step.Here you can find the mentioned implementation.
Throughout this course the students submit weekly assignments in order to be able to evaluate their learning progress. Furthermore, At the end of the course students carry out group projects dedicated to data-driven building simulation.

Contact us

Piacenza Offices:

Polo Territoriale di Piacenza, Politecnico di Milano
Software Section: Via Scalabrini 111, 29121,Piacenza, Italy
Laboratory: Via Scalabrini 111, 29121,Piacenza, Italy
+39 0523 356811
behzad.najafi@polimi.it

Milan Offices:

Dipartimento di Energia,Politecnico di Milano
Via Lambruschini 4, 20156,Milan, Italy
+39 02 2399 8518
behzad.najafi@polimi.it
fabio.rinaldi@polimi.it