南水北调与水利科技
南水北調與水利科技
남수북조여수리과기
SOUTH-TO-NORTH WATER
2014年
1期
145-148,153
,共5页
大坝%变形%预测%灰色模型%马尔科夫链%自适应MGM-MC模型%自适应MGM(1, n)模型
大壩%變形%預測%灰色模型%馬爾科伕鏈%自適應MGM-MC模型%自適應MGM(1, n)模型
대패%변형%예측%회색모형%마이과부련%자괄응MGM-MC모형%자괄응MGM(1, n)모형
dam%deformation%prediction%Grey model%Markov chain%self-adaptive MGM-MC model%self-adaptive MGM( 1,n) model
大坝的变形受到多种因素的影响,对于一些没有长期、连续、可靠监测数据的工程,采用传统的多变量灰色模型 MGM(1, n)进行大坝变形预测时,往往随着预测时间的推移,预测精度降低。采用自适应 MGM(1, n)模型,根据有限的监测资料,综合考虑各个变量之间的相互影响,通过置入最新信息取代最老的信息,来反映坝体变形过程中的随机因素或扰动对系统的影响。以此为基础,利用马尔科夫链确定位移时序的状态转移概率矩阵,通过对实测值、拟合值以及所处状态的分析,对大坝变形进行更高精度的预测。实例表明,和传统多变量灰色模型 MGM (1, n)以及自适应 MGM(1, n)模型相比,自适应 MGM(1, n)-马尔科夫链模型( M GM-MC 模型)具有更高的精度。
大壩的變形受到多種因素的影響,對于一些沒有長期、連續、可靠鑑測數據的工程,採用傳統的多變量灰色模型 MGM(1, n)進行大壩變形預測時,往往隨著預測時間的推移,預測精度降低。採用自適應 MGM(1, n)模型,根據有限的鑑測資料,綜閤攷慮各箇變量之間的相互影響,通過置入最新信息取代最老的信息,來反映壩體變形過程中的隨機因素或擾動對繫統的影響。以此為基礎,利用馬爾科伕鏈確定位移時序的狀態轉移概率矩陣,通過對實測值、擬閤值以及所處狀態的分析,對大壩變形進行更高精度的預測。實例錶明,和傳統多變量灰色模型 MGM (1, n)以及自適應 MGM(1, n)模型相比,自適應 MGM(1, n)-馬爾科伕鏈模型( M GM-MC 模型)具有更高的精度。
대패적변형수도다충인소적영향,대우일사몰유장기、련속、가고감측수거적공정,채용전통적다변량회색모형 MGM(1, n)진행대패변형예측시,왕왕수착예측시간적추이,예측정도강저。채용자괄응 MGM(1, n)모형,근거유한적감측자료,종합고필각개변량지간적상호영향,통과치입최신신식취대최로적신식,래반영패체변형과정중적수궤인소혹우동대계통적영향。이차위기출,이용마이과부련학정위이시서적상태전이개솔구진,통과대실측치、의합치이급소처상태적분석,대대패변형진행경고정도적예측。실례표명,화전통다변량회색모형 MGM (1, n)이급자괄응 MGM(1, n)모형상비,자괄응 MGM(1, n)-마이과부련모형( M GM-MC 모형)구유경고적정도。
Dam deformation is influenced by a lot of fact ors. For those projects without long-term, cont inuous, and reliable moni-toring data, the predict ion precision of dam deformation based on the t raditional MGM( 1, n) model decreases with time. In this paper, the self-adapt ive MGM( 1, n) model was applied. T he proposed model characterizes the interact ion between each variable, and replaces the oldest informat ion with newinformat ion, which can reflect the effect s of random fact ors or perturbation on dam deformation. On the basis, the state t ransition probability matrix of the t ime series was det ermined by Markov chain, and the monitoring data and forecast data were analyzed to predict the dam def ormat ion wit h a higher precision. Compared with the tra-ditional MGM( 1, n) model and self-adaptive MGM( 1, n) model, the MGM-MC model has higher precision.
