A robust fire detection algorithm for temperature and optical smoke density using fuzzy logic

Multi sensor based fire detection (MSbFD) systems are one of the important current developments in automatic fire detection technology. The two main objectives of this progress are the still unacceptable false alarm behaviour and improvements in the fire detection capabilities (i.e. shorter detection times) of fire detection systems. The use of more than one sensor in a fire detector gives an improved image of the environment monitored and hence allows a safer alarm decision. Multi sensor technology allows but does imply the enhancement of fire detection systems in the desired directions. The crucial point is the evaluation and interpretation of the signals produced by the monitored phenomena. This signal processing (detection algorithm) mostly determines the detectors capabilities. Due to the availability of microcontrollers applicable to fire detector technology with its severe technical constraints (i.e. power consumption) modern signal processing techniques (neural networks, fuzzy logic) can be used. The paper presents a MSbFD algorithm using two fire parameters (temperature and optical smoke density). These two sensors were chosen since ionization systems may become increasingly difficult to apply because of the environmental regulations being imposed on them. The evaluation and processing of the sensor signals is carried out by the use of fuzzy logic. The concept of the algorithm is outlined and its performance and robustness in the fire and the non-fire case is shown by simulation results.