PREDICTION DEVICE POWER CONSUMPTION BASED multilayer neural networks

摘要

The invention relates to apparatus for predicting energy consumption / power consumption of a physical object based on a multilayer neural network. The device comprises a multilayer neural network including a first input layer neurons is determined by the amount of input data, a second hidden layer neuron number which is selected experimentally, the third output layer comprising one neuron. approximation unit has one input and one output, which is connected to one of the neurons in the input layer of the multilayer neural network to take account of the dynamics of growth / reduce power / electricity consumer. Output neuron of the third layer is connected to one of the adder inputs, a second input of the adder coupled to an output approximating unit. The output of adder has no connections to the inputs of the first and second layers of neurons. The invention relates to intelligent devices power / energy consumption prediction and can be used to determine the short and medium volumes of energy / electricity. Approximating unit composed of a multilayer neural network provides an approximation of the input data on energy consumption / power consumption and generates a value obtained by the difference of the actual values ​​and approximated energy / power consumption, which can significantly narrow the interval data normalization, increasing the accuracy of prediction of energy / electricity. Block approximation and the neural network can be implemented on modern element base - block ATtiny approximation on the microcontroller 12, a neural network on the microcontroller ATMega32. 2 yl., 1 tab.