Bioingeniería
Permanent URI for this collection
Browse
Browsing Bioingeniería by Author "Amado Ospina, Santiago"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
Item Modelo matemático de regresión lineal múltiple para terapia de hipertermia con nanopartículas y calentamiento inducido por láser(Universidad Santiago de Cali, 2021) Amado Ospina, Santiago; Valdés Cabrera, Daniel Mauricio; Vargas, Jhon Edwar (Director)Nowadays, there is interest in the study of hyperthermia therapy for the treatment of cancer. This therapy consists on increasing body temperature to between 37 °C and 43 °C, which can be achieved by inducing electromagnetic waves like radiofrequency, microwave and/or laser and ultrasound waves. Additionally, the use of nanoparticles loaded in the tumor allows a focalized treatment reducing damage to surrounding healthy tissues. For the implementation of this treatment it is necessary to perform in vitro experiments and mathematical model simulations to obtain a better understanding of the therapy. There are different mathematical models which represent the dynamics of this therapy. However, these models are complex and, in many cases, require a high computational cost. This limits the optimization processes since they would take a long time to be solved. In this sense, this work proposes a model with a reduced computational execution time using the multiple linear regression technique. The training and validation of the regression model was performed on database of 200 simulations produced by a mathematical model developed in Comsol Multiphysics® by Python. Experimental measurements were taken in the hyperthermia process with heat induction by near infrared during 100s in pure distilled water and two different ratios of iron oxide nanoparticles dissolved in distilled water, i.e., 0.025% and 0.050% by weight. Once the model is obtained, a parameter estimation process is performed by means of the Gauss-Newton (GN) and Levenberg-Marquardt (LM) algorithms. The results indicate that the temperature field predicted by the multiple linear regression model is close to the output values of the full mathematical model and the experimental measurements. The correlation of the model proposed was 0.99 with respect