Fire Alarm System and Management Level Evaluation Model Design

Xuan Tang ； Zhi Liu ； Liangliang Yang ； Xilai Chen ； Peixuan Shan

《World Scientific Research Journal》 8卷11期 (2022/11) Pp. 156-176

https://doi.org/10.6911/WSRJ.202211_8(11).0024

摘要

This paper examines the issue of fire alarm systems. With the rapid development of fire detector industry, fire detection and alarm technology is becoming more and more perfect, and the number of newly installed fire detectors in buildings is huge every year. Therefore, improving the reliability of fire detectors and reducing the failure of fire alarm systems has a major role in timely detection of fire, control of fire and protection of life and property. In response to question 1: First, data cleaning was performed on Annex 1 to remove some missing values, compare the remaining data for the same address, machine number and loop code, and check the number of the same alarm message, and finally determine the real number of fires as 440. The top three fires in the brigade area were 51 in G Brigade, 49 in M Brigade and 41 in C Brigade. The frequency of use, the number of alarms per square kilometer, the false alarm rate and the failure rate were used as evaluation indicators to build the evaluation model, and the CRITIC weighting method, the entropy weighting method and the coefficient of variation method were used to analyze the degree of correlation between the two factors, followed by the game theory comprehensive weighting method to obtain the weights of frequency of use, number of alarms per square kilometer, false alarm rate and failure rate as 0.243, 0.227, 0.229 and 0.301 respectively. Finally, the evaluation model of TOPSIS was established to score the various detectors, in which the signal valve detector has the highest score of 0.98854, followed by the intelligent photoelectric detector with a score of 0.97571. Therefore, the optimal detector is the signal valve detector. For problem two: data cleaning based on the data findings given in problem one, reselection of indicators, elimination of irrelevant data. Secondly, RUSBoost prediction model, Random Forest prediction model and Adaboost prediction model were picked to train the sample data respectively, compare and use the RUSBoost algorithm with the highest accuracy rate. Then this model was used to predict the target and evaluate the authenticity of the data in Annex 3. Some of the real fire probability prediction results: The probability of real fire in G brigade, J brigade and N brigade are 0.4167, 0.2746 and 0.3242. For question 3: Based on the above questions and Table 1, the data were first processed and the indicators were reselected to obtain the weights using the entropy weighting method. Subsequently, TOPSIS model was established to evaluate the comprehensive management level of each fire brigade, and the analysis results can be obtained: the highest rating of K fire brigade is 0.8689, and the lower ratings of F fire brigade, R fire brigade and M fire brigade are 0.3666, 0.3821 and 0.4073, and the improvement plan is proposed for this result. For problem 4: Firstly, analyze the model results of the above problem and put forward suggestions for the maintenance of each component of the fire alarm system from three aspects: reasonable selection of fire detectors, improving the authenticity of alarm signals and effectively improving the comprehensive management level of the fire brigade, taking into account the actual situation. Finally, this paper evaluates and generalizes the model. It is concluded that this model can also be generalized to other regional or national fire safety system management scoring, component reliability scoring problems.

關鍵字

TOPSIS Game Theory Integrated Empowerment Method RUSBoost Entropy Method Evaluation Model

參考文獻

Jin Chao. Selection of automatic fire alarm detectors for cigarette production joint workshop[J].BuildingElectricity,2015(12):42-45.doi:10.3969/j.issn.1003-8493.2015.12.007.

Zhao Chuan. Improvement and implementation of fire alarm system in nuclear power plants J].InstrumentUser,2022,29(3):49-52.doi:10.3969/j.issn.1671-1041.2022.03.012.

Xiao Chengjia, Li Hongwen. Proper selection and configuration of smoke detectors can reduce non-fire alarm rates [J]. Fire Protection Technology and Product Information, 1994,(10):28-31.

Guidance on actively promoting the role of freestanding smoke sensitive fire detection alarms for fire prevention and control [J]. Fire Industry,2015,0(4):50.