基于非参数法的人造板及制品中甲醛释放量风险分析研究  

On the hazard analysis of formaldehyde emission content from the wood-based panel and products by the non-parametric method

作　　者：龚方 吴国灿[2] 杜恒清 周保华 林雨霏 于仕超 谭乐义 

机构地区：[1]山东出入境检验检疫技术中心,山东青岛266002 [2]北京师范大学数学科学学院,北京100875

出　　处：《安全与环境学报》2011年第3期251-254,共4页Journal of Safety and Environment

基　　金：国家质检总局科技计划项目(2009IK076)

摘　　要：木制品中有害物质风险分析涉及多指标、多水平,必须采用多元统计分析方法从整体上对多指标样品进行分析。为了更有效地估计木制品中有害物质含量的代表值,以山东口岸人造板及其制品连续3a的甲醛释放量检测数据为样本,使用Bootstrap方法对木制品中有害物质质量浓度控制上限(UCL)进行计算、评估、比较及验证。结果表明,山东口岸人造板及其制品在2007年、2008年和2009年的UCL值分别为2.508 6 mg/L、1.916 5 mg/L、1.300 9 mg/L,反映出木制品中甲醛释放量风险逐年降低的趋势。The present paper is inclined to present a more effective method for estimating the representative values of the harmful substances in wood-based products. In developing the above said method, we have taken the formaldehyde emission testing data of wood-based panel and products in Shandong ports as our case study samples for a period of three consecutive years. As the first step, we have calculated, evaluated and validated the upper control limits （UCL） of the harmful substances by using the Bootstrap method, which is a commonly used nonparametric statistical approach. It can be directly used to estimate the distribution or characteristics of randora variables through data resample without assumption the distribution of variable. In the process of managing the harmful substances of such products, it is first of all necessary to do the effective description of the overall quality risk-assessment of the above mentioned harmful substances. For the time being, risk analysis of wood product harmful substances involves several indicators. Therefore, multivariate statistical analysis method should be used as the samples for such risk evaluation. And, the estimation of representative values of the harmful substances in wood products serves one of the most basic working mechanisms. At the same time, according the calculation and analysis we have done, the UCL value of year 2007, 2008 and 2009 were 2.508 6 mg/L, 1.916 5 mg/L, and 1.300 9 mg/L, which can respectively reflect the annually decreasing risk of formaldehyde emission in the wood-based panel and products. By means of its resampiing simulation mechanism,the bootstrap method can help to resolve the problems arising from the multimodality of data distribution and limitation of sampling. This method has the prior advantage of giving a robust description of the central tendency and the variation of data profiles. In addition, we have made a detailed discussion with the bootstrap mean and interval. Thus, we can say that the bootstrap method has offered a new risk an

关 键 词：环境工程学 甲醛 人造板 BOOTSTRAP 非参数方法 风险分析 

分 类 号：X820.4[环境科学与工程—环境工程]