改进Elman神经网络在短期热负荷预测中的应用

热电厂的短期热负荷预测在城市集中供暖中起着至关重要的作用,直接影响热电厂的经济效益和热能利用率。电厂的短期热负荷一般采用神经网络预测模型进行预测,而BP神经网络应用最为广泛。Elman神经网络算法在BP神经网络基础上加入了承接层,作为一步延时算子,实现记忆能力,使系统具备适应时变能力,增强系统全局稳定性。但Elman神经网络算法模型的构造依然需要大量样本的支撑,而且输入层的变量多,导致预测时间依然很长,收敛速度慢。该文在Elman神经网络预测前,进行了相关系数预处理和对样本中异常值的平均化预处理,通过数据归一化运算,使Elman神经网络输入层变量大幅减少。仿真实验表明,改进的Elman神经网络算法使预测模型快速寻优,减少预测时间的同时明显提高预测精度。     

压裂水平井产能预测方法研究进展

水平井压裂技术在低渗透及非常规储层中得到了广泛应用,压裂后水平井的产能预测关系到油田开发方案的制定,因此,国内外学者对压后渗流模型的建立和求解方法做出了不懈的努力。本文详细回顾了国内外学者所建压裂水平井产能预测模型及求解方法,指出了不同模型和求解方法的优缺点,并展望了压裂水平井产能模型的发展方向。     

Reference governors and maximal output admissible sets for linear periodic systems

ABSTRACT

In this paper, we consider the problem of constraint management in Linear Periodic (LP) systems using Reference Governors (RG). First, we introduce the periodic-invariant maximal output admissible sets for LP systems. We extend the earlier results in the literature to Lyapunov stable LP systems with output constraints, which arise in RG applications. We show that, while the invariant sets for these systems may not be finitely determined, a finitely-determined inner approximation, which is periodically invariant, can be obtained by constraint tightening. We then analyze the geometric and algebraic relationship between these sets and show that these sets are related via simple transformations, implying that it suffices to compute only one of them for real-time applications. This greatly reduces the memory burden of RG (or other similar constraint management strategies), at the expense of an increase in processing requirements. We present a thorough analysis of this trade-off. In the second part of this paper, we present two RG formulations, and discuss their properties and algorithms for their computation. Numerical simulations demonstrate the efficacy of the approach.     

煤系地层致密砂岩气甜点区地震逐级预测——以鄂尔多斯盆地东南缘下二叠统山西组2~3亚段为例

鄂尔多斯盆地东南部下二叠统山西组2~3亚段(以下简称山2~3亚段)为该盆地重要的天然气勘探目的层,但该亚段储层薄、厚度变化快、非均质性强,储层预测和勘探目标优选难度大。为了准确预测该亚段煤系地层致密砂岩气甜点区、提高天然气勘探成功率,针对该套储层的特征和预测难点,提出了90°相移技术识别河道外形、模型约束波阻抗反演刻画砂体厚度和子波衰减梯度属性识别含气砂体的地震逐级预测技术。研究结果表明:①山2~3亚段上覆5号煤地震强反射层,下伏储层地震反射能量弱,加之为稀疏二维地震测网、井控程度低,致使致密砂岩气甜点区预测难度大;②所提出的技术方法通过地震逐级预测约束,可以有效地刻画河道砂体分布并识别有效含气储层,提高了对勘探开发目标预测的精度;③基于该技术方法指导部署的勘探开发目标实钻效果好,地震预测结果横向分辨率高,真实地反映了河道及河道砂体的变化特征。结论认为,采用该方法可以有效地解决二维地震勘探区煤系地层强非均质性、薄储层致密砂岩气甜点区预测的地质难题。     

Robust predictive extended state observer for a class of nonlinear systems with time-varying input delay

ABSTRACT

This paper deals with asymptotic stabilisation of a class of nonlinear input-delayed systems via dynamic output feedback in the presence of disturbances. The proposed strategy has the structure of an observer-based control law, in which the observer estimates and predicts both the plant state and the external disturbance. A nominal delay value is assumed to be known and stability conditions in terms of linear matrix inequalities are derived for fast-varying delay uncertainties. Asymptotic stability is achieved if the disturbance or the time delay is constant. The controller design problem is also addressed and a numerical example with an unstable system is provided to illustrate the usefulness of the proposed strategy.     

Adaptive neural network command filtered backstepping control of pure‐feedback systems in presence of full state constraints

An adaptive neural network (NN) command filtered backstepping control is proposed for the pure‐feedback system subjected to time‐varying output/stated constraints. By introducing a one‐to‐one nonlinear mapping, the obstacle caused by full stated constraints is conquered. The adaptive control law is constructed by command filtered backstepping technology and radial basis function NNs, where only one learning parameter needs to be updated online. The stability analysis via nonlinear small‐gain theorem shows that all the signals in closed‐loop system are semiglobal uniformly ultimately bounded. The simulation examples demonstrate the effectiveness of the proposed control scheme.     

Looking into the TESSERACT: Time-drifts in event streams using series of evolving rolling averages of completion times

Business processes are dynamic and change due to diverse factors. While existing approaches aim to detect drifts in the process structure, Tesseract looks for temporal drifts in activity interim times. This orthogonal view on the process extends the traditional data cube of events – case id, activities and timestamps – by a fourth dimension and improves the operational support by a visualization of temporal drifts in real-time.Insights about temporal deviations lead to an augmented awareness of imminent failures or improved service times. The detection of related structural concept drifts can be improved by early warning, as operation times of critical parts often increase before they catastrophically fail.     

基于聚类的数据加权优化在犯罪预测中的应用

近年来，我国传统暴力犯罪与成年人犯罪呈下降态势，但是，犯罪案由层出不穷。为有效提升公安实践工作中犯罪预测能力，打击各类违法犯罪事件,本文针对犯罪数据，提出一种新型犯罪预测模型。利用密度聚类分析方法将犯罪数据分类，然后进行数据降维提取关键属性生成特征数据，继而对特征数据进行加权优化并采用机器学习的方式对特征数据进行学习，从而预测犯罪案由。实验结果表明，与传统方法相比，本文方法具有更好的预测效果，为公安实践工作中类似案件的侦破和预防，提供新的路径支撑。