电梯交通流量特性分析及预测方法研究

发布时间：2018-05-25 19:32

本文选题：电梯交通流量 + 周期性　；参考：《天津大学》2013年博士论文

【摘要】：在高层和中高层建筑中，为了提高电梯运行的效率，降低运行能耗，电梯群控系统的调度问题已经引起极大重视。提高电梯运行效率的关键因素之一是事先获取详细的电梯交通流量信息。电梯交通流作为电梯群控系统的调度对象，对群控系统的节能降耗起决定性的影响，电梯交通流研究的主要内容是特性分析和流量预测。本文全面、系统地对三种流量特性进行了详细的定性分析和定量描述，同时从三种流量特性出发分别建立了相应特性的流量预测模型，为电梯群控调度提供了理论依据。主要研究内容及创新点是：（1）首先对电梯交通流的典型特性即周期性和随机性，进行了定性分析和定量描述。其次，重点研究电梯交通流中存在混沌性。通过重构电梯交通流量序列的吸引子进行定性分析；计算数据序列的关联维数和最大李亚谱诺夫指数进行定量分析。结果表明：上、下高峰模式下多种交通流时间序列都呈现出了相似的低维混沌行为特征，即存在混沌现象。本结论将有助于电梯群控系统根据流量混沌特性进行流量预测或调整群控调度策略，从而提高各项性能指标。（2）对办公大楼、教学楼及住宅楼等进行电梯交通流量分析表明，该类大楼内的交通流量具有明显的周期性，并且交通流量时间序列中存在线性和非线性特征数据。提出采用ARIMA模型与GP模型相结合的混合预测方法实现了电梯交通流量的混合预测。仿真结果表明，文中所提混合预测模型较其它单一预测模型都具有较好的预测精度。（3）对医院就诊大楼、机场大楼以及宾馆等建筑物的电梯交通流量分析表明，这类大楼的交通流量具有明显的随机性。提出了采用模糊加权马尔可夫链模型对电梯交通流量交通状态进行预测。预测结果与实际电梯交通进行比较，，结果表明所提方法的有效性和可行性。（4）对于不具有周期性且随机性又不明显的不规则流量，已分析其具有的混沌特性。针对具有混沌特性的交通流量，提出了基于支持向量机的电梯交通流量混沌时间序列预测方法，并采用PSO寻优的方法确定了模型的最优参数。分析结果表明，基于SVM的预测模型，相比于其他两种单一的线性模型及非线性模型，更适合于电梯交通流这样的小样本预测，预测精度较高。
[Abstract]:In order to improve the efficiency of elevator operation and reduce the running energy consumption, the dispatching problem of elevator group control system has attracted great attention in high-rise and medium-high-rise buildings. One of the key factors to improve elevator operation efficiency is to obtain detailed elevator traffic flow information in advance. As the dispatching object of elevator group control system, elevator traffic flow plays a decisive role in energy saving and consumption reduction of group control system. The main content of elevator traffic flow research is characteristic analysis and flow prediction. In this paper, the detailed qualitative analysis and quantitative description of the three flow characteristics are carried out in detail, and the corresponding flow prediction models are established from the three flow characteristics respectively, which provides a theoretical basis for elevator group control scheduling. The main research contents and innovations are as follows: Firstly, the typical characteristics of elevator traffic flow, namely periodicity and randomness, are qualitatively analyzed and quantitatively described. Secondly, the chaos in elevator traffic flow is studied. Qualitative analysis is carried out by reconstructing the attractor of elevator traffic flow sequence and quantitative analysis is carried out by calculating the correlation dimension of the data sequence and the maximum Li Ya spectral nodal exponent. The results show that many traffic flow time series show similar low dimensional chaotic behavior in the upper and lower peak modes, that is, the existence of chaos. This conclusion will be helpful for elevator group control system to predict the flow according to the chaotic characteristics of the flow or adjust the group control scheduling strategy so as to improve the performance of the system. 2) the analysis of elevator traffic flow in office building, teaching building and residential building shows that the traffic flow in this kind of building has obvious periodicity, and there are linear and nonlinear characteristic data in the time series of traffic flow. A hybrid forecasting method based on ARIMA model and GP model is proposed to realize the mixed prediction of elevator traffic flow. The simulation results show that the proposed hybrid prediction model has better prediction accuracy than other single prediction models. 3) the analysis of elevator traffic flow in hospital, airport building and hotel shows that the traffic flow of this kind of building has obvious randomness. A fuzzy weighted Markov chain model is proposed to predict the traffic state of elevator traffic flow. The prediction results are compared with the actual elevator traffic, and the results show that the proposed method is effective and feasible. (4) the chaotic characteristics of irregular flow with no periodicity and randomness are analyzed. For the chaotic traffic flow, a chaotic time series prediction method based on support vector machine (SVM) is proposed, and the optimal parameters of the model are determined by PSO optimization method. The analysis results show that the prediction model based on SVM is more suitable for small sample prediction of elevator traffic flow than the other two single linear models and nonlinear models.
【学位授予单位】：天津大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：TU857

【参考文献】