中文摘要：

飞机熏着陆严重影响着飞行安全，控制霞着陆事件风险是该领域研究难点。提出了模糊加权Markov模型方法，进行飞机重着陆状态预测研究。首先，通过起落架接地受力分析，确定飞机接地载荷作为描述重着陆状态的特征参数，并引入模糊有序聚类算法，获得蠹着陆状态划分规则；然后，针对起落架接地载荷进行状态分级，计算载荷序列的自相关系数，获取多时滞Markov链的权值；最终，建立加权Markov模型，预测起落架接地载荷时序分布，根据载荷序列概率分布来评价重着陆等级，实现重着陆风险的控制。工程实例表明，该算法准确率高，具有较好的空间性能和时间性能。

英文摘要：

The landing performance of aircraft affects the flight safety seriously. Within the concept of Continuous Security Risk Management （CSRM） model being put forward by SEI, Risk assessment of hard landing state becomes key point in aviation safety. A Markov modeling method is proposed to predict landing state for risk management, also fuzzy clustering is introduced to improve model performance. Firstly, in consideration of the time-varying characteristic of hard landing signal, feature parameter for hard landing is obtained using normal force analysis of main gear and nose gear during landing period, a nonlinear function between landing load and impact emerge is deduced, then state definition rules for hard landing are achieved by sequential fuzzy clustering which is solved by yielding point algorithm. Secondly, after indicating the stochastics and dynamics of hard landing process with Markov chain, a five-level state space of hard landing event is described with definition rules, and weights of variable-step Markov model are calculated according to self-correlation coefficient, consequently the essence of hard landing event is constituted from variable-step Markov method. Finally, considering an array which made up of forty landing loads from an A320 - 200 type plane, transition matrix in five states is calculated for each four step, thus Markov model is established with corresponding modeling weight, model output is achieved by weighted summation, and state prediction for landing gear load is performed in time sequence, also landing state is labeled with five scale which evaluating the risk of hard landing. Engineering practice shows that fuzzy Markov method can get satisfied forecast precision and reliable forecast results, furthermore Markov method is interpretable for hard landing with respect to black box method such as artificial neural work.