The Neural System of Monitoring and Evaluating the Parameters of the Elements of an Intelligent Building

Abstract

In the intelligent buildings more and more features are dependent on the proper operation of equipment powered by electricity. Possible damage to these devices is very detrimental to the functioning of the building and must be repaired or replaced immediately. In many cases a failure prevents the building from being used. Possible equipment failures, however, can be predicted from the analysis of how they operate. Monitoring and analysis of power parameters of individual electrical devices allow to distinguish characteristic parameters of each electrical receiver. Any departure from stable working conditions may be recorded by a neuronal monitoring and evaluation system. The mechanism for such an analysis is the implementation of an adaptive prediction algorithm using artificial neural networks. This method allows adaptation of the decision mechanism to the current working conditions of controlled devices. The measured parameters are the measurements of physical quantities that illustrate the operation of the device. For example, for air handling and ventilation units, this is the electricity consumed. The advantage of power analysis is the identification of common faults and corresponding deformations of power supply parameters. Based on the previously prepared pattern, neural networks identify component damage or predict the predicted critical failure time of the component or control subsystem using MWF. Early forecasting of the failure situation contributes significantly to the comfort and security of intelligent building users.