Font Size: a A A

Experimental Study On Laboratory Simulation Of Larix Gmelinii Forest Spotting Ignition

Posted on:2020-12-01Degree:MasterType:Thesis
Country:ChinaCandidate:J B NingFull Text:PDF
GTID:2393330578976189Subject:Forest fire prevention
Abstract/Summary:PDF Full Text Request
Spot fire is a kind of special forest fire behavior,which is easily affected by many factors and has great randomness.Spot fire will cause great difficulty to fire fighting and serious threat to the safety of fire fighters.The phenomenon of spot fire can be generally divided into three sequential processes:firebrand generation,firebrand transportation and landing ignition of fuels.This study started from the process of spotting ignition,taking Larix gmelinii as research object which is typical coniferous trees of Daxing'an Mountains.Using cone,1-hour and 10-hour time lag twig,which are naturally present in the forest,as firebrand.Fuel beds with different compression ratios and moisture content of needle were constructed.5 400 ignition experiments were conducted under different wind speed(three kinds of firebrand for each 1 800 times).After variables were selected,a logistic model was established for predicting ignition probability of spot firebrand.The variables contained environmental factors,fuelbed and firebrand characteristics.The results are as follows:(1)The ignition capability of 3 firebrands is in the order of cone,10-hour Lime lag twig and 1-hour time lag twig;(2)It is found that wind speed is positively correlated with ignition probability within designed wind speed;fuel moisture content is negatively correlated with ignition probability,when fuel moisture content is 40%,tirebrand as cone and 10 hours time leg twig ignited times are more than 30%result in firebrand dry weight;there is no significant correlation between packing ratio and ignition probability result in the affected by different variables;(3)The final model of cone ignition probability isXW:wind speed(m·s-1);XFD:firebrand dry weight(g);XBM:fuel moisture content,the same below,and model threshold is 0.205,predicting ignition accuracy rate is 83.1%,predicting non-ignition accuracy rate is 68.6%,total predicted accuracy rate is 72.10%;the final model of 1 hour time leg twig ignition probability ismodel threshold is 0.114,predicting ignition accuracy rate is 79.5%,prcdicting non-ignition accuracy rate is 78.6%,total predicted accuracy rate is 78.5%:the final model of 10 hours time leg twig ignition probability is P10h=1/1+e-(-5.919+1.575X(?)-2.427XBM+14.495X(?))XFM is firebrand moisture content,and model threshold is 0.193,predicting ignition accuracy rate is 81.5%,predicting non-ignition accuracy rate is 79%.total predicted accuracy rate is 79.5%.
Keywords/Search Tags:Spot fire, firebrand, Larix gmelinii, laboratory simulation, ignition probability, Logistic
PDF Full Text Request
Related items