The current study aimed to design and test a prototype of a thermoelectric cooler (TEC) using thermoelectric modules (TEM) operating near room temperature. The thermoelectric cooler utilized in this investigation has a maximum cooling power of 46 W and dimensions of 40 mm × 40 mm × 3.6 mm. After a series of measurements, the device temperature decreased from an ambient temperature of 19.6 °C to 1.6 °C, with a notable coefficient of performance around 0.9, achieved through the utilization of both serial and parallel connections. The secondary objective of this research was based on an artificial neural network (ANN) approach. An ANN model was constructed using an experimental database acquired from our thermoelectric refrigeration device. The input parameters of humidity, time, performance coefficient, cooling power, and heat dissipation were introduced into the model to enhance the cold temperature as the output. Using a multilayer perceptron (MLP), the experimental dataset was used for training, testing, and validating the ANN. The precision of the model was evaluated using three established statistical metrics: mean squared error (MSE), mean absolute percentage error (MAPE), and R-squared (R2).
Read full abstract