|
Predict the effective coverage area of biological warfare agents and the degree of pollution is the key link , which accurately predict pollution concentrations warfare agents and maintain the vitality of time will provide an important basis for large - scale on-site decontamination task , help achieve timely implementation of effective biological protection reaction, the appropriate response and accurate protection targets. However, for a long time , the conventional site disposal method empiric inference is based , low accuracy , a great distance with the actual situation , although can be used as the reference data , but may also give the elaboration of specific measures not appropriate consequences , may disposal ineffective or overreaction phenomenon . This study, Bacillus subtilis test bacteria , the use of biological information technology , electron microscopy , aerosol technology , viable count analysis and computer neural network tool for intelligence analysis and other methods , the surface to simulate different environmental conditions Bacillus aerosol Population fluctuation basis for building fast, accurate prediction of environmental surface Bacillus aerosols the residual resistance utility model , the results are as follows : 1 . bioinformatics homology analysis , transmission electron microscopy ultrastructural observation , as well as heat , UV and the available chlorine resistance determination analysis confirmed anthrax wax -like spores , Bacillus subtilis spores compared to similar structure and size , similar biological genetic homology of heat , UVC, and available chlorine resistance is basically the same . Alternative anthrax spores as test bacteria Bacillus and Bacillus subtilis spores wax-like . Under in simulated natural surface environment (blade , stone , tile , and a piece of cloth ) , aerosol retention resistance changes with changes in temperature, humidity , UVC irradiation intensity and time of exposure showed obvious demise trend , wherein UVC the exposure is the most obvious weakening of Bacillus resistance , the blades in Bacillus aerosol stranded resistance slightly stronger than the other surfaces . 3 . Research Matlab6.1 - based neural network technology platform Bacillus aerosol Delay stress prediction model research . According to the research purpose , the smooth curve characteristics of of analog environmental conditions and data training set of five input neurons , eight hidden layer nodes and one output neuron . The the 'tansig', 'purelin' transfer function , trainlm training function . Network 100 iterations . The review of the model prediction efficiency reached 95% , 85% the forward projections efficiency reaches .
|