Research On Fire Monitoring System Based On Fuzzy-rough Set

Now the fire prevention safety problem in buildings is being taken seriously and the indoor fire monitoring work plays a crucial role in the whole process. A new trend in fire monitoring work involves utilizing the multi-sensor data acquisition and analysis techniques, which make the intelligentization and networking of the fire monitoring system possible. Hence, it is practically significant to develop a fire monitoring system which is economically applicable and precisely effective.The overall purpose of the fire monitoring system is to monitor the indoor environmental change and the focus should be attached to its real-time performance and preciseness. Only in this way can the safety of life and property be ensured. Currently most of the fire monitoring systems use single smoke detection method which could easily give rise to the misinformation and the omissive information incidents. To solve these problems and to enhance the real-time performance and dependability of the monitoring work, the monitoring system talked about in this thesis makes a composite and reliable monitoring of the indoor environment by adopting the temperature sensor DS18B20, smoke transducer MQ-2 and IR flame sensor. The low-power STC89C52 SCM is acting as the controller and in the wireless communication module, the high cost-effective, low-power nRF24L01 chip is responsible for the data transmission. Via nRF24L01, the collected data is sent from the Internet nodes to the upper computer. Meanwhile, the rule extraction is done by the fuzzy-rough sets and the problem that the threshold value of the initial data sent by the sensor’s node cannot be referred to could be addressed.The main concern of this thesis includes the followings: on the one hand, by analyzing the technical difficulties and the existent problem of the current fire monitoring system, a fire monitoring model is set up with fuzzy-rough sets as its theoretical basis. This new model would conduct a rule extraction of the sample data collected in fire monitoring system, regulate the threshold value of the data sent by the sensors’ nodes, improve the preciseness of the monitoring data, upgrade the focus intensity of fire monitoring and make the data converging to the upper computer better reflect the actual situation. On the other hand, a hardware structural design is made which encompasses the collection and sending of sensor nodes and converging nodes of the data, the analysis of network communication protocol design and network routing protocol design as well as the flow chart depicting the work done by each module.In the very last part, by testing and verifying the fire monitoring system network, the testing result indicates that each modules work properly and the protocol design features are reasonable which is also reliability. Several fire simulation tests are practiced in specific experimental environment and the results demonstrate that the system can satisfactorily monitor the three environmental variables, namely, temperature, smoke and flame. Besides, the fire monitoring model investigates the rules of the collected data and evaluates the correlation between the rules and the system so as to perfect the current research.