【摘要】：目前國內(nèi)外在橋梁長期監(jiān)測技術(shù)的研究領(lǐng)域主要集中于大型橋梁。由于中小橋梁數(shù)量更多,垮塌事故主要發(fā)生在中小橋梁,因此,應(yīng)加大中小橋梁的長期監(jiān)測系統(tǒng)的研究。本文依托《基于物聯(lián)網(wǎng)的中小橋梁長期安全研究》(國家973科研項(xiàng)目課題),采用基于時(shí)間序列的數(shù)據(jù)挖掘技術(shù),對中小橋梁監(jiān)測數(shù)據(jù)進(jìn)行挖掘和分析,為中小橋梁監(jiān)測數(shù)據(jù)處理提供了新的思路。取得主要工作成果有:①介紹了時(shí)間序列分析方法的基本理論,分析了時(shí)間序列中的各個(gè)數(shù)據(jù)的影響因素,為建立數(shù)學(xué)模型提供理論基礎(chǔ)。②根據(jù)橋梁長期監(jiān)測系統(tǒng)采集到的數(shù)據(jù)特點(diǎn),論證了以時(shí)間序列作為理論基礎(chǔ)建立ARMA模型對數(shù)據(jù)進(jìn)行分析預(yù)測的可行性。并對ARMA模型的建立過程進(jìn)行了詳細(xì)論述,其中包括數(shù)據(jù)的最初采集,數(shù)據(jù)平穩(wěn)化、標(biāo)準(zhǔn)化處理,模型參數(shù)的估計(jì),模型階數(shù)的確定等問題。通過比選,ARMA模型參數(shù)的估計(jì)選擇了最小二乘估計(jì)法,模型階數(shù)的確定選擇了AIC準(zhǔn)則。③以一座中小橋——偏巖子橋?yàn)橐劳?將基于時(shí)間序列理論的預(yù)測理論應(yīng)用于數(shù)據(jù)挖掘中。利用MATLAB軟件對該橋的裂縫測點(diǎn)、撓度測點(diǎn)、傾斜度測點(diǎn)、應(yīng)變測點(diǎn)的監(jiān)測數(shù)據(jù),建立ARMA模型,調(diào)整到合適的模型階數(shù)及模型參數(shù)后,進(jìn)行每次向外延伸5步的短期預(yù)測,對預(yù)測值與實(shí)際監(jiān)測值進(jìn)行比較,從而驗(yàn)證基于ARMA模型的中小橋梁監(jiān)測數(shù)據(jù)挖掘方法的可行性與有效性。
[Abstract]:At present, the research field of long-term bridge monitoring technology at home and abroad mainly concentrates on large-scale bridges. Because of the large number of small and medium-sized bridges, the collapse accidents mainly occur in small and medium-sized bridges, therefore, the long-term monitoring system of small and medium-sized bridges should be studied. Based on the long term Safety Research of small and Medium-sized Bridges based on the Internet of things (the national 973 scientific research project), this paper uses the data mining technology based on time series to mine and analyze the monitoring data of small and medium-sized bridges. It provides a new idea for monitoring data processing of medium and small bridges. The main achievements are: 1 introduces the basic theory of time series analysis method, analyzes the influence factors of each data in time series, In order to provide theoretical basis for the establishment of mathematical models, according to the characteristics of the data collected by the bridge long-term monitoring system, the feasibility of establishing the ARMA model to analyze and predict the data based on the time series is demonstrated. The establishment process of ARMA model is discussed in detail, including the initial collection of data, the stabilization of data, the standardization of processing, the estimation of model parameters, the determination of model order and so on. The least square estimation method is used to estimate the parameters of the model. The order of the model is determined by AIC criterion .3. The prediction theory based on the time series theory is applied to the data mining based on a medium and small bridge. By using MATLAB software, the monitoring data of crack, deflection, inclination and strain measurement points of the bridge are measured, and the ARMA model is established. After adjusting to the appropriate model order and model parameters, the short term prediction of 5 steps extending out each time is carried out. By comparing the predicted values with the actual monitoring values, the feasibility and effectiveness of the monitoring data mining method for small and medium-sized bridges based on ARMA model are verified.
【學(xué)位授予單位】：重慶交通大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2015
【分類號】：U446

【參考文獻(xiàn)】

相關(guān)期刊論文 前1條

1 梁宗保;陳偉民;符欲梅;胡順仁;朱永;;混凝土橋梁結(jié)構(gòu)應(yīng)變監(jiān)測的溫度效應(yīng)分離方法研究[J];混凝土;2005年12期

，

本文編號：2165415