Bibliographic citations
Chambi, N., (2023). Diseño de un sistema de compensación de error de pesaje de faja de llenado de mineral usando redes neuronales [Tesis, Universidad Nacional de Ingeniería]. http://hdl.handle.net/20.500.14076/26611
Chambi, N., Diseño de un sistema de compensación de error de pesaje de faja de llenado de mineral usando redes neuronales [Tesis]. PE: Universidad Nacional de Ingeniería; 2023. http://hdl.handle.net/20.500.14076/26611
@mastersthesis{renati/712685,
title = "Diseño de un sistema de compensación de error de pesaje de faja de llenado de mineral usando redes neuronales",
author = "Chambi Quiroz, Nelson Esteban",
publisher = "Universidad Nacional de Ingeniería",
year = "2023"
}
In the present thesis a design of an error compensation system is developed using Neural Networks to compensate for the weighing error of the mineral filling belt, which is above 10%. For the development of the error compensation system method, the Neural Networks method will be used, where it will be possible to calculate the synaptic weights, bias and the architecture of the Neural Network, these values are calculated with the back propagation algorithm during the training of the Neural Networks, as well as the parameters of the Network and then they will be validated to corroborate the error compensation. For the training of the Neural Network, the mathematical model, input and output values and parameters of the Neural Network are proposed to later be trained and analyze the error compensation for each training in order to select the optimal Network with synaptic weights and bias. To achieve this, field data was obtained that was collected from a mineral filling belt, which used 70% of the data for training and 30% for validation in addition to this the data was taken randomly and to obtain the results the MATLAB software is used. To compensate for the error, a Neural Network with 3 inputs, 1 output and 2 hidden layers with 5 neurons in each hidden layer was designed, this Neural Network will be in charge of the process of compensating for the weighing error. Likewise this proposed model reduces the error to 1.65% to achieve this the hidden layers and neurons for each layer are optimized. The optimization of the design of the Neural Network manages to obtain good results that showed that the designed network made the error compensation and therefore manages to reduce the margin of error in the final weight of the mineral from 10% to 1.65%.
This item is licensed under a Creative Commons License