基于ARIMA时间序列模型的中国塑料制品产量预测研究

根据中国2000~2019年塑料制品产量的时间序列数据,研究差分自回归移动平均模型ARIMA(p,d,q)的建模和应用。利用白噪声检验和平稳性检验对原始序列进行预处理及ARIMA模型识别,在模型定阶后进行参数估计,检验模型拟合效果,并预测塑料制品的产量。结果表明:ARIMA(2,1,1)模型可以很好地描述塑料制品产量的变化趋势,使用该模型可以预测未来五年的塑料制品产量。     

A mechanism informed neural network for predicting machining deformation of annular parts

Controlling machining deformation of annular parts is crucial for ensuring the performance of high value products and equipment. For example, during manufacturing of critical parts in aircrafts and spacecrafts, accurate prediction of machining deformation is the basis for guiding the formulation of deformation control strategies. However, due to the complexity of the machining deformation of annular parts, existing methods still have limitations in accurate prediction. To this end, this paper proposes a mechanism informed neural network (MINN) to predict machining deformation of annular parts. MINN is realized by establishing the dual sub-networks structure and using enhanced loss functions with the consideration of the deformation mechanism model characteristics of annular parts. The deformation was decomposed into the axisymmetric portion and the non-axisymmetric portion according to the deformation superposition principle, and modeled separately based on the thin-shell theory and Fourier series. Experiment results showed that the proposed method could predict the machining deformation of annular parts more accurately and stably with a small amount of training data, compared with previous methods.     

基于时序数据的云网协同平台人工智能运维体系

云计算在企业应用中的拓展不但表现为平台规模的拓展，也表现为平台应用的延伸。“云网协同”和“微服务化”是当前企业云平台演进的重要趋势。随着企业信息化建设重要性的持续提升，微服务化云网协同平台的运行维护面临极大挑战。首先分析了平台运维面临的挑战，梳理了平台人工智能运维需求，提出了基于时序数据分析的平台人工智能运维技术体系，并给出了云网协同平台人工智能运维子系统参考设计。所提技术体系和系统设计具有实用性和推广性，可以作为企业云平台建设和优化的技术途径参考。     

基于CPET时序聚类的中长跑耐力运动员选拔方法

心肺运动试验（CPET）能将人体的呼吸系统、心血管系统等综合为一体，不仅能够体现受试者的有氧运动能力，评估受试者的心肺耐力，而且能以整体整合医学的视角来研究受试者对运动的应激反应。为对CPET数据进行凝聚层次聚类分析，提出一种基于时间序列形态特征的算法。选取15名业余中长跑运动员的CPET数据作为聚类对象，聚类指标选取了表征有氧能力和心肺耐量的耗氧量、二氧化碳、心率、分钟通气当量、代谢当量、生理死腔与潮气量比值、呼吸商及每搏输出量等8类指标，体现运动员摄取、利用氧的效率、肺循环以及心功能等综合状况。通过聚类分析发现受试者个体差异较大，未出现明显的“群居分布”特征，根据轮廓系数评估可剔除心肺耐量较差的测试者。实验结果表明，该算法在确保聚类准确率的同时能够降低数据压缩率，且对形态特征显著的数据集进行聚类效果更佳。     

Subspace Clustering for Situation Assessment in Aquatic Drones: A Sensitivity Analysis for State-Model Improvement

Abstract

In this paper, we propose the use of subspace clustering to detect the states of dynamical systems from sequences of observations. In particular, we generate sparse and interpretable models that relate the states of aquatic drones involved in autonomous water monitoring to the properties (e.g., statistical distribution) of data collected by drone sensors. The subspace clustering algorithm used is called SubCMedians. A quantitative experimental analysis is performed to investigate the connections between i) learning parameters and performance, ii) noise in the data and performance. The clustering obtained with this analysis outperforms those generated by previous approaches.     

Investigation of seasonal variations of Beijing pavement condition data using unevenly spaced dynamic panel data model

Beijing municipal highway administration started to collect pavement condition data on its major expressways since 2006. It is advised in the Chinese practice standard that data collection shall be conducted on annual basis. However, pavement data are usually collected at different seasons of the year, which may cause significant seasonal variations in the observed condition. Moreover, for some reasons, data are missing for some of the pavement sections at certain years, which could bring up difficulties in performance model estimation and inference. These concerns have been simply neglected in past practice. This study proposed an unevenly spaced dynamic panel data model to investigate the seasonal patterns of a performance indicator called Ride Quality Index (RQI). A quasi-differencing approach was adopted for the estimation. Data collected from the 5th Ring Road of Beijing were used in the case study. It was found that RQI data collected during the fall season are expected to be lower than that collected during the spring or summer seasons. Findings from this research would be helpful to pavement engineers in using unevenly spaced pavement condition data for future condition estimation.     

A finite line source simulation model for geothermal systems with series- and parallel-connected boreholes and independent fluid loops

A model for the simulation of geothermal systems with parallel- and series-connected boreholes is presented. Mass and heat balance problems are formulated for each component in the system and are assembled into system-level problems. A third problem is formulated to account for heat transfer in the bore field, using the finite line source solution. This third problem is coupled to the system-level heat balance problem by an analytical solution of the heat transfer inside boreholes with multiple U-tubes. The simulation model allows for any number of independent fluid loops within the bore field or within individual boreholes and allows for combinations of specified inlet fluid temperatures and heat extraction rates in independent fluid loops. The model accounts for the axial variation of the fluid and borehole wall temperatures and heat extraction rates. The capabilities of the model are demonstrated through three example simulations.     

Mechanism of mechanical–chemical synergistic activation for preparation of mullite ceramics from high-alumina coal fly ash

High-alumina coal fly ash (HAFA) is a special solid waste due to the existence of more than 45% alumina and 35% silica, which can be applied to prepare Al-Si series ceramics if the impurities can be removed and the Al/Si mass ratio can exceed 2.55 (Mullite: 3Al₂O₃.2SiO₂). In this work, a new mechanical–chemical synergistic activation–desilication process is proposed, and the contents of different impurities can be lowered up to less than 1%, and the Al/Si mass ratio can be elevated from 1.26 to 2.71. Especially, the mechanism of this process is investigated in detail. The analysis of the mechanism shows that the decrease of Q⁴(3Al) and the increase of Q⁴(0Al), Q⁴(1Al), and Q⁴(2Al) improve amorphous silicate reactivity through synergistic activation, and the exposed amorphous Si-O-/Si-O-Si can be removed by OH- during the desilication process (desilicated ratio>55%), which help the fine mullite to exhibit excellent properties (bulk density > 2.85 g/cm³, apparent porosity < 0.5%) during the sintering process. Finally, this process not only decreases the pollution but also alleviates the shortage of Al/Si resources and promotes the clean development of coal-fired power generation.     

地磁图数据随机误差建模与补偿

地磁图是影响地磁匹配导航性能的重要因素,其中地磁图的随机误差严重影响匹配定位精度,甚至导致误匹配。 为提 高地磁匹配性能,提出了地磁图数据随机误差建模与补偿方法。 该方法在对地磁图数据随机误差特性分析的基础上,建立数据 的非平稳时间序列模型,并将此模型作为状态方程,以实时数据为量测量,采用 Kalman 滤波方法对地磁图数据进行滤波,补偿 地磁图数据的随机误差。 通过基于滤波前后地磁图数据的导航定位实验,间接证明滤波方法的有效性。 对实际地磁图数据处 理结果表明,应用该方法滤波后的地磁图数据进行导航,其定位精度提高了 54. 7%。     

基于多变量时间序列及向量自回归机器学习模型的水驱油藏产量预测方法

提出了一种基于多变量时间序列(MTS)及向量自回归(VAR)机器学习模型的水驱油藏产量预测方法,并进行了实例应用。该方法在井网分析的基础上通过MTS分析对注采井组数据进行优选,并将井组内不同采出井产油量及注入井注水量作为彼此相关的时间序列,通过建立VAR模型从多个时间序列中提取出相互作用规律,挖掘注采井间流量的依赖关系从而进行产量预测。水驱油藏历史生产数据分析结果表明,与数值模拟历史拟合结果相比,机器学习模型产量预测结果具有更高精度,同时不确定性分析提升了预测结果的安全性。通过脉冲响应分析对注入井的采油贡献量进行评价,可为注水开发方案调整提供理论指导。