Politecnico di Torino (logo)

Virtual Sensing for the Estimation of the State of Health of Batteries

Davide Faverato

Virtual Sensing for the Estimation of the State of Health of Batteries.

Rel. Alessandro Rizzo, Stefano Alberto Malan, Giovanni Guida. Politecnico di Torino, Corso di laurea magistrale in Mechatronic Engineering (Ingegneria Meccatronica), 2020


Nowadays an increasingly efficient and optimal use of electrical energy storage systems plays an important role in the main international scientific and industrial lines of action: there is a need to gradually move away from the use of fossil fuels in order to ensure an improvement in the life quality, with a consequent increasing demand for self-sufficient and energy-efficient systems. Despite extensive literature and a large number of patents filed, the number of products on the market adopting such technologies is limited. BAT-MAN project aims to bring significant innovations in the development of applications based on a new estimation and diagnostic techniques, not only in the field of storage systems. The objective concerns the realization of a low-cost product idea, measuring the SoC (State of Charge) and the SoH (State of Health) of a storage system or battery, which is capable of generate appropriate alarms for the recognition of critical states in order to allow operations to restore and/or replace the battery itself and informe the user or another system of the current state of charge and the general state of health of the battery. This thesis aims to formalize the ERMES (Extendible Range MultiModal Estimator Sensing) algorithm designed by Brain Technologies whose innovative value is to be identified in the methodologies and technologies to be applied to the problem of the diagnosis of an accumulation system, and in particular to the problem related to the estimation of the SoH. Before evaluating the functionalities of the algorithm produced by Brain Technologies it was necessary to identify the methodology to be adopted to formalize this algorithm. The methodologies available are basically two: the first provides a more mathematical and methodological approach to the problem while the second one is more experimental through a DoE (Design of Experiment) approach. After a depth evaluation it was decided to proceed by adopting, because of its advantages, the second methodology available. The main contributions of this Thesis are: - The identification of the main parameters for DoE analysis on the algorithm. It is necessary to identify which parameters, such as the current profile and its harmonic content, the simulation time, the reset time or the characterization of the residual error, have the greatest influence on the performances. - The generation of specific simulation scenarios to identify and quantify the benefits of this new battery state of health sensing technique. - A Sensitivity Analysis to identify the effects of any exogenous disturbances on the performance. Exogenous disturbances consist of incorrect modelling of battery model parameters such as nonlinear resistance, OCV (Open Circuit Voltage) or internal dynamics. In an ideal scenario it is always possible to create a mathematical model that perfectly describes the system under examination, but when this is not the case it is necessary to evaluate the effects of the lacks of knowledge on performance to verify the robustness of the algorithm. - A final comparison between the proposed algorithm and a well known and commonly adopted estimation technique such as the Extended Kalman Filter with augmented states. Starting from the single EKF adopted in the algorithm an augmented states EKF is generated and the performances in different scenarios are evaluated in order to deduce advantages and possible disadvantages of the new proposed estimation technique.

Relators: Alessandro Rizzo, Stefano Alberto Malan, Giovanni Guida
Academic year: 2020/21
Publication type: Electronic
Number of Pages: 249
Additional Information: Tesi secretata. Fulltext non presente
Corso di laurea: Corso di laurea magistrale in Mechatronic Engineering (Ingegneria Meccatronica)
Classe di laurea: New organization > Master science > LM-25 - AUTOMATION ENGINEERING
Aziende collaboratrici: Brain technologies
URI: http://webthesis.biblio.polito.it/id/eprint/16052
Modify record (reserved for operators) Modify record (reserved for operators)