Similar users

Sara HADJI REZAI

Sara HADJI REZAI

Construction,  Civil Engineering,  Risk Analysis,  Building Performance,  Hazard Analysis,  Scientific Research,  Scientific Vulgarization,  Building Construction
Baptiste Gaborieau

Baptiste Gaborieau

Science Communication,  Biosciences
Alex Cauvi

Alex Cauvi

Chemistry,  HéRitage,  Material Science,  Charity,  Sci Comm,  Documentary,  Communication
vladimir bajic

vladimir bajic

Data Mining,  Bioinformatics,  Artificial Intelligence,  Knowledge Mining,  Drug Repurposing,  Diagnostic Biomarkers
Mohamad Ivan  Fanany

Mohamad Ivan Fanany

Machine Learning,  Computer Vision,  Data Science
Lisa Rivalin

Lisa Rivalin

Research Scientist - Data Science Project Manager

Disciplines: Mathematics Multidisciplinary
United States

Summary

Lisa Rivalin is a Research Scientist at ENGIE Axima, a leading French company in Climate Engineering. She is also currently working as an affiliate research scholar at Lawrence Berkeley National Laboratory (LBNL) on energy prediction and management for Smart Buildings. She is working on Model Predictive Control (MPC), which can exploit the historic measurements of a building to predict the future performance and find an optimal control strategy based on these predictions. She received a PhD in Applied Statistics and Building Energy from Paris Sciences et Lettres - PSL Research University (prepared in Mines ParisTech)

Experience

Data Science Project Manager Since June 2016

ENGIE Axima (Nantes FR)

Research Scientist - Data Science Project Manager Since June 2016

Engie Axima & Lawrence Berkeley National Laboratory ( US)

Work as an Affiliate Researcher on energy prediction and management for Smart Buildings, specifically on Model Predictive Control (MPC). Develop MPC algorithms capable of exploi-ting historical measurements of a building to predict energy performance and create optimal control strategies. Drive the creation of a Data Mining Platform for “Smart Buildings” that in-tegrates innovation and industry best practices to develop algorithms to control buildings. De-velop an in-depth understanding of Machine Learning and optimization algorithms, streamli-ning their use and integration to improve data collection.

  • Work within a team on the development and testing of the open-source platform MPCPY

  • Conduct studies on the impact of data quality (both missing and noisy data) on prediction accuracy

  • Collaborate with customers on applying research results to their use cases

  • Identify and evaluate innovative start-up companies, organize learning expeditions, and forge partnerships

  • Launch new co-working programs to facilitate collaborative, cross-departmental work in the company, increasing engagement and reducing overhead

  • Track project progress, draft regular reports to communicate timelines and resource needs with clarity and close attention to detail to ensure timely delivery of work product

  • Contribute to academic research proposal writing and expert conference organization

  • Publish several scientific papers and participate actively in conferences

Research Engineer, Ph.D. in Applied Statistics and Energy September 2012 - May 2016

Engie Axima Nantes and Mines ParisTech ( FR)

Selected for my academic and experiential success to receive corporate financial support in pursuit of my degree. De-signed a comprehensive research study of potential sources of uncertainty in the building modeling process and created a statistical tool to predict energy performance and identify parameters to monitor, given those uncertainties.

  • Completed 70% of my Ph.D. work at Engie-Axima

  • Created new statistical tools from scratch to estimate energy performance in corporate contracts, predict energy consumption of buildings before construction, assess uncertainties that were introduced because of uncertain inputs, and conduct sensitivity analyses to characterize the most influential input variables

  • Successfully used metamodels to reduce simulation time in developing new analytic tools, reducing the total compu-tation time of previous models by 95%

  • Ph.D. thesis downloaded more than 1000 times in two years, earning recognition in the field

  • Honors: Finalist to the international Engie’s Innovation Trophies

    Relevant Coursework: Uncertainty and Sensitivity Assessment Methods, Python Programming, Uncertainty Analysis with OpenTURNS.

Ingénieure de Recherche (Doctorante) April 2012 - May 2016

ENGIE Axima (Nantes FR)

Education

Ph.D. in Building Energy and Applied Statistics December 2012 - May 2016

Mines ParisTech (Paris FR)

Ph.D. in Applied Statistics and Energy September 2012 - May 2016

https://pastel.archives-ouvertes.fr/tel-01376689

Ph.D. in Applied Statistics and Energy September 2012 - May 2016

https://pastel.archives-ouvertes.fr/tel-01376689

 

Abstract : Before the construction of a building, an energy performance guarantee consists in predicting the energy required for user comfort. To do that, it is necessary to state a contractual consumption and identify the key parameters to pay special attention to. Thus, for new buildings, consumption is calculated under design phase, when several data are uncertain. Thus, the dynamic thermal simulation is carried out with hypothetical data, without having the possibility to calibrate with measures.This PhD thesis aims to develop a method of uncertainty quantification during the design step and construction process of a new building. These uncertainties are classified into three categories: those associated with the calculation methods used for building and system modeling, those related to the lack of knowledge of model parameters and those due to the real use conditions of the building (occupancy and weather).To achieve this goal, uncertainties associated with the calculation methods are addressed, to identify some practices limiting the number of errors and the associated uncertainties. Then, a methodology is defined to choose the critical parameters used for the probabilistic study and to associate them with a distribution according to the available knowledge. The central part of this PhD thesis is dedicated to an exhaustive comparison of methods to select a fast uncertainty propagation and sensitivity analysis method. Finally, after illustrating the overall contracting approach and discussing the inclusion of financial risks, the method is applied on a real case, on which an adjustment formula is added to take into account actual weather and usage.

Master’s Degree in History and Philosophy of Sciences Since 2012

Université Paris Diderot Paris VII

Main Courses: History, Sociology, and Philo-sophy of Physics

All exams passed successfully in 2012; cur-rently writing final thesis entitled “How to model knowledge” related to my Ph.D. ex-ploratory study

Master’s Degree in History and Philosophy of Sciences Since 2012

Université Paris Diderot Paris VII

Main Courses: History, Sociology, and Philo-sophy of Physics

All exams passed successfully in 2012; currently writing final thesis entitled “How to model knowledge” related to my Ph.D. exploratory study

Master’s Degree in Engineering 2009 - 2010

ensi poitiers

General Courses: Statistics, Mechanics, Fluid Mechanics, Electronics

Energy Courses: Industrial Combustion; Motors; Thermal Machines; Building Energy; Heat Transfer; Solar, Nuclear and Wind Energy

Programming Courses: Database, Scientific Programming, Fortran, Matlab

Awards

Finalist 2017 Engie Innovation Trophies 2017

https://innovation.engie.com/en/innovation-trophies/algorisk/4601

Interests

Research Data Visualization Data Science Deep Learning Statistics Uncertainty Sensitivity Complex System Modeling Physical Modeling Simulation Reliability Optimization Machine Learning Philosophy Of Science

Contact me

Report this profile