polito.it
Politecnico di Torino (logo)

Automated Calibration Algorithm Applied To Electric Vehicle Heat Generation Logic Control Using Brent's Minimization Method

Dante Zollo

Automated Calibration Algorithm Applied To Electric Vehicle Heat Generation Logic Control Using Brent's Minimization Method.

Rel. Diego Regruto Tomalino. Politecnico di Torino, Corso di laurea magistrale in Automotive Engineering (Ingegneria Dell'Autoveicolo), 2025

[img] PDF (Tesi_di_laurea) - Tesi
Accesso riservato a: Solo utenti staff fino al 17 Ottobre 2026 (data di embargo).
Licenza: Creative Commons Attribution Non-commercial No Derivatives.

Download (4MB)
Abstract:

In this thesis, an automated calibration method for control algorithms, based on look-up tables, is proposed. The method has been applied to the algorithm strategy for battery heating controlling the thermal energy generated by the electrified powertrain. Look-Up tables are widely used in powertrain control systems, both for electric powertrains and for internal combustion engine powertrains, as they enable real-time computations based on precomputed results. This makes them ideal for online applications due to their low computational cost. The main issue of using look-up tables comes from their calibration. At present, this calibration requires expert engineers that have to tune these tables through a highly time consuming and resource intensive process which involves extensive testing and iterations. Furthermore, this manual process usually leads to suboptimal performances . This thesis proposes an algorithm capable of automating the calibration process through optimization, making it less human-dependable and enhancing the final performance of the tuned table. The algorithm automatically selects all the possible combinations of table entrees optimizing, in this way, each bin of the table at a time. A comparison between two optimization techniques, Golden Section Search and Brent’s Minimization Method, will be carried out showing eventually which are the advantages and also the limitations of the proposed method. The results confirmed that Brent's Minimization Method has a faster convergence than Goden Section Search, thus it is able to calibrate the lookup table faster. Moreover some modifications to Brent's minimization method will be described; these make the user select the error tolerance for the optimization and make it find the solution faster. However, the method presents a problem related to the non unimodality of the objective function in some bins; in conclusion a solution to this problem is addressed and its results are discussed.

Relatori: Diego Regruto Tomalino
Anno accademico: 2025/26
Tipo di pubblicazione: Elettronica
Numero di pagine: 93
Soggetti:
Corso di laurea: Corso di laurea magistrale in Automotive Engineering (Ingegneria Dell'Autoveicolo)
Classe di laurea: Nuovo ordinamento > Laurea magistrale > LM-33 - INGEGNERIA MECCANICA
Ente in cotutela: UNIVERSITY OF WINDSOR (CANADA)
Aziende collaboratrici: University of Windsor
URI: http://webthesis.biblio.polito.it/id/eprint/37434
Modifica (riservato agli operatori) Modifica (riservato agli operatori)