| The accidents caused by hazardous goods transportation not only lead to serious casualties and property losses, but also cause serious environmental pollution, ecological destruction and a series of environmental and ecological problems. In this paper, in order to improve the security of hazardous goods transportation, the monitoring systems and methods of the hazardous goods smart tank are studied.The monitoring system of dangerous goods is consist of the preceding four parts, i.e. the pointer instrument digital interface, the sensor interface system, the power management system and the data transmission system. The dangerous goods tank has a large number of pointer instruments. In order to obtain the value of the pointer instrument, the pointer instrument digital interface is designed. To provide each sensor with signal conditioning circuit and power consumption, the sensor interface system is designed. To ensure the monitoring equipment working properly, the power management system is designed. To ensure the effective connect of the background part and the terminal part and achieve low level information fusion, the data transmission system is designed.To limit the power into the working condition and improve the security of monitoring, A low power gas sensor is designed. The hollow microspheres zinc oxide materials are prepared by hard template. The material is coated onto gold electrodes and a gas sensor is manufac-tured. When the concentration ethanol is 100 ppm, the response time is about 6 s while the recovery is about 94 s. The three sensors exhibit an excellent reproducibility of the sensitivity indicating a good sensor to sensor consistency. The sensor displays very good repeatability during the six testing cycles. Compared to commercial ethanol gas sensors, the developed gas sensor has prominent advantages in response time, gas response rate and power consumption.To make an accurate and fast prediction of the leakage in a closed area, the gas sensor transient response model is built. Through exponential model, the approximate transient response is built. The recursive difference model is translated from the exponential model and the parameters is estimated by the least square method. A computational fluid dynamics simulation is adopt to show the filling process. The relative error of the steady state is-0.248%. A practical gas calibrating and measuring system is set up to collect the gas concentration. The relative error of the steady state is-0.57%. The prediction of the finial concentration is taken and the predicted concentration save 220 and 298 seconds on average for 80% and 90% of the final concentrations, respectively.To monitor the whole process, the method that to decide the tank state is presented. The method is based on the finite state automata. The tank states serve as the node and the outputs of sensors are the condition of transitions between states. The operation process is divided into two parts, i.e. the initial state and the state transition. The output of the initial state is the state in the finite state automata and the state transition part determine the next state. Comparing with the four common classification methods, i.e. C4.5, support vector machine (SVM), Bagging and Naive Bayes, the proposed method has higher classification accuracy. The proposed method has the biggest average area under the curve (0.9962).To make full use of the complementary and redundancy of multiple sensors, the number of alarm sensor selection method is proposed. The method is designed for minimizing the equivalent false alarm rate and the optimal number of alarm sensors and the minimum equiva-lent false alarm rate are obtained. An experiment set-up is developed to verify the proposed method. Comparing with the common classification methods, the proposed method has the biggest area under the curve (0.9631), indicating the proposed method has the highest moni-toring reliability. |
PDF Full Text Request