선박조종시뮬레이션의 최저 시행에 관한 연구

Abstract

In the year of 2010 the Korean government introduced an enforced maritime traffic safety assessment act for the purpose of enhancing traffic safety. The act was mandated by the law to design a new port or modify an exiting one. According to Korea Maritime Safety Act, the assessment of propriety of marine traffic system comprises safety of channel transit and berthing/unberthing maneuver, safety of mooring, and safety of marine traffic flow. The safety of channel transit and berthing/unberthing maneuver can be evaluated only by ship-handling simulation which is carried out by sea pilots working with the port concerned. The vessel's proximity measure is an important factor for evaluating traffic safety. The proximity measure is composed of vessel's closest distance to channel boundary and probability of grounding/collision. Therefore, the probability of grounding cannot be ignored.

According to central limit theorem, a sample has a normal distribution on condition that its size is more than 30. However, more than 30 simulation runs lead to an increase in the assessment period which results in difficulty in employing sea pilots. Hence, this paper aims to analyze the minimum sample size for evaluating vessel's proximity.

In this research, mean and standard deviation of ten cases are obtained from the latest maritime traffic safety assessment. The probability of grounding is within 10-4. Then each case generates twenty random sample sets, each set constitutes the sample 3, 4, 5, 6, 7, 9 and 11，at the same time it calculates the h value and confidence interval of each sample sets. Then the box-plots which consists of twenty sample boxes is drawn, and the mean line and confidence interval are also shown on the box-plots. In the box-plots, the X-axis refers to the sample set, while the Y-axis is CPA to channel boundary. Based on the size of confidence interval, the change of confidence interval span, the relative position of mean line and box figures, it can be indicated that the minimum number of simulation should be more than 5.

After accumulating the mean and standard deviation curves of confidence interval, when the size of simulation runs is larger than 5, a much smaller mean and standard deviation of confidence interval than those in 3 can be obtained.

In each case of this study, the 20 sample sets of data are obtained randomly by the parameter mean and standard deviation. If the actual data is used, it is difficult to ensure the 20 sample sets of actual data have the same mean and standard deviation. As such, it is meaningless to use the actual data to analyse. Thus this study uses the random data instead of actual data to analyse.

In conclusion, this paper proposes a minimum sample size of 5, that is, the simulation should be carried out more than five times. It is recommended that actual data should be used to carry out analysis and other tests other than KS test should be applied to goodness of fit for the sample distribution.