• Wind and Structures
• /
• v.5 no.1
• /
• pp.1-14
• /
• 2002

Artificial Neural Networks (ANN) have the capability to develop functional relationships between input-output patterns obtained from any source. Thus ANN can be conveniently used to develop a generalised relationship from limited and sometimes inconsistent data, and can therefore also be applied to tackle the data obtained from wind tunnel tests on building models with large number of variables. In this paper ANN model has been developed for predicting wind induced pressures in various zones of a Gable Building from limited test data. The procedure is also extended to a case wherein interference effects on a gable roof building by a similar building are studied. It is found that the Artificial Neural Network modelling is seen to predict successfully, the pressure coefficients for any roof slope that has not been covered by the experimental study. It is seen that ANN modelling can lead to a reduction of the wind tunnel testing effort for interference studies to almost half.

• The Korean Journal of Emergency Medical Services
• /
• v.9 no.2
• /
• pp.55-78
• /
• 2005

The purpose of this study was to provide basic data for the improvement of the guidelines and training programs regarding the cardiopulmonary resuscitation performance of bystanders who can respond to the incidents in earlier times as the first responder of the cardiac arrest incident, by reviewing the performance of basic CPR and the influencing factors after providing 70 students of Department of Emergency Medical Technology with the CPR training. For the purpose of the study, the collected data were computerized and analyzed by SPSS-WIN program(ver. 10.1). The results for this study were as follows The duration of session between the groups in the BLS CPR were 3 minutes and 36 seconds, 2 minutes and 32 seconds respectively. The average compression number per minute were 24.3 times and 33,2 times respectively(p=.000), and the average compression rate per minute were 112 times and 122 times respectively(p=.000). The average ventilation number per minute were 3.54 times and 5.1 times respectively(p=.000). The errors in compression "Too shallow" were 20.73 times(34.6%) and 23,23 times(38,7%) out of 60 times in 4 cycles with the standard of 38 nun. In CPR performance results according to gender in the first episode, males showed better results in compression depth as 41.5 mm comparing to females average 38.2 mm(p=.015). When ventilation results were compared according to the use of FS, the average ventilation number per minute, total ventilation per minute and the average volume per episode were significantly higher when FS was not used(<.040), There was no significant difference in ventilation accuracy between two groups. According to the results, we need to improve and distribute portable barrier devices, and to be familiar with those devices. We need to enforce ventilations as well as to include compressions so that faster and more accurate CPR can be performed. Additionally, we need to exclude ventilation only cases, minimize the interference time of chest compression due to inaccurate ventilation, simplify or minimize the complicatedness of CPR performance and responding time related to breathing, provide first responders with various training programs such as initial assessment and ventilations only, or initial assessment and chest compression-only CPR and than provide advanced training with AHA BLS education including CPR for more than two people according to CPR skills and target characteristics.