基于模糊粗糙集和事例推理的凝汽器真空故障诊断

针对凝汽器真空故障诊断的不确定性和复杂性,提出一种基于模糊粗糙集和事例推理的凝汽器故障诊断新方法。首先,运用模糊粗糙集属性约简方法对故障特征进行约简和权重分配,不仅提取了反映故障的主要特征量,降低了特征变量之间的非线性相关性,而且避免了人的主观性对权重分配的影响。然后,在分析凝汽器真空故障特征的基础上,建立凝汽器真空故障树,以约简特征作为条件对故障树根节点进行归纳检索,有效地减少了候选事例的数量,再通过最近邻法检索故障树叶节点,对凝汽器真空故障进行智能定位。通过对汽轮机凝汽器历史故障特征数据集仿真,验证了该方法的有效性。     

A fault detection method based on sparse dynamic canonical correlation analysis

Fault detection based on canonical correlation analysis (CCA) has received increased attention due to its efficiency in exploring the relationship between input and output. However, traditional CCA may generate redundant features in both the input and output projections while maximizing the correlations. In this paper, sparse dynamic canonical correlation analysis (SDCCA) is developed for dealing with the fault detection of dynamic processes. Through posing sparsity in the extraction of features, the interpretability of canonical variates is enhanced attributed to the sparsity of canonical weights. Based on the SDCCA model, the $$ monitoring metric is established for fault detection. Moreover, the upper control limit (UCL) based on $$ monitoring metrics is determined by the kernel density estimation (KDE) method to avoid the violation of the Gaussian assumption. The superiority of the proposed SDCCA-based fault detection method is illustrated through a comparative study of the Tennessee Eastman process benchmark.     

A fuzzy diagnosis approach using dynamic fault trees

By incorporating digraph models, fault trees and fuzzy inference mechanisms in a unified framework, a novel approach for fault diagnosis is developed in this work. To relieve the on-line computation load, the fault origins considered in diagnosis are limited to the basic events in the cut sets of a given fault tree. The symptom occurrence order associated with each root cause is derived from system digraph with the qualitative simulation techniques. The implied candidate patterns are enumerated according to two proposed theorems and then encoded in the inference system with IF-THEN rules. The simulation results show that the proposed approach is not only feasible but also capable of identifying the most likely cause(s) of a hazardous event at the earliest possible time.     

基于ICA混合模型的多工况过程故障诊断方法

针对工业过程数据的多模态和非高斯特性,提出一种基于独立元混合模型（independent component analysis mixture model,ICAMM）的多工况过程故障诊断方法。该方法将独立元分析与贝叶斯估计结合,同时完成各个工况的数据聚类和模型参数求取,并建立基于贝叶斯框架下的集成监控统计量实时监控过程变化。在检测到故障后,针对传统的变量贡献图方法无法表征变量之间信息传递关系的缺点,提出基于信息传递贡献图的故障识别方法。该方法首先计算各变量对独立元混合模型统计量的贡献度,进一步通过最近邻传递熵描述故障变量之间的传递性,挖掘故障变量之间的因果关系,从而确定故障源变量和故障传播过程。最后对一个数值系统和连续搅拌反应釜（CSTR）过程进行仿真研究,结果验证了本文所提出方法的有效性。     

基于粗糙集理论的故障树重要度分析

在工程风险评价的失效可能性计算中,故障树重要度分析是一项重要技术,它经常遇到许多不确定、不完整的因素.采用粗糙集理论对这些不确定、不完整数据进行处理与分析,并通过埋地燃气管道的外腐蚀故障树重要度分析的实例,阐述了粗糙集理论的概念、原理、方法和应用.结果表明,粗糙集理论是一种完全基于客观的综合重要度分析方法,在工程上具有实用意义.     

一种基于KICA-GMM的过程故障检测方法

核独立元分析(kernel independent component analysis,KICA)故障检测方法的故障检测时间易受独立元顺序和主导独立元数目经验选取的影响,针对这个问题,提出基于KICA和高斯混合模型(Gaussian mixture model,GMM)的故障检测方法。采用KICA从正常工况测量数据中提取独立元,用GMM拟合各独立元的概率密度函数,建立基于GMM的监控量及其控制限;计算各独立元的监控量均值,以此判断其非高斯性强弱,对每个强非高斯独立元进行单独监控,对弱非高斯部分采用主元分析法进行监控。在Tennessee Eastman过程上的仿真结果说明,相比于KICA故障检测方法,所提方法不需要排序独立元和选取主导独立元数目,避免了其对故障检测时间的影响,能够有效利用过程信息,缩短故障检测的延迟时间。     

A process fault diagnosis method using multi-time scale dynamic feature extraction based on convolutional neural network

Unlike many other techniques used in process control, which are widely applied in practice and play significant roles, abnormal situation management (ASM) still relies heavily on human experience, not least because the problem of fault detection and diagnosis (FDD) has not been well addressed. In this paper, a process fault diagnosis method using multi-time scale dynamic feature extraction based on convolutional neural network (CNN) consisting of similarity measurement, variable ranking, and multi-time scale dynamic feature extraction is proposed. The CNN-based model containing the fixed multiple sampling (FMS) layer can extract dynamic characteristics of process data at different time scales. The benchmark Tennessee Eastman (TE) process is used to verify the performance of the proposed method.     

Chemical process fault diagnosis based on a combined deep learning method

The study on fault detection and diagnosis (FDD) of chemical processes has always been the top priority of the chemical process safety. In this paper, a fault diagnosis method combining the deep convolutional with the recurrent neural network (DCRNN) is proposed. In this method, the data from chemical processes are input to the deep convolutional neural network (DCNN) to extract features in spatial domains, and then, the features are fused into the bidirectional recurrent neural network (BRNN). Due to the powerful capabilities of DCNN to extract features in spatial domains and the sensitivity to time series of RNN, the combined method can adaptively learn the dynamic information of the raw data in both spatial and temporal domains and has unique advantages in multivariate chemical processes. The application of the DCRNN model in the Tennessee Eastman (TE) process demonstrates the high accuracy of this proposal in identifying abnormal conditions for the chemical process, compared with the traditional fault identification algorithms of deep learning.     

Fast distributed MPC based on active set method

Modern chemical plants are characterized by their large-scale, strong interactions and the presence of constraints. With its ability to systematically handle these issues, distributed model predictive control (DMPC) is a promising approach for the control of such systems. However, the problem of how to efficiently solve the resulting distributed optimization problem is still an open question. This paper develops a novel fast DMPC approach based on a distributed active set method and offline inversion of the Hessian matrix to efficiently solve a constrained distributed quadratic program. A dual-mode optimization strategy based on the value of unconstrained optimal solution is developed to accelerate the computation of control action. The proposed method allows for the optimization to be terminated before convergence to cope with the fast sampling periods. Furthermore, a warm-start strategy based on the solution of the previous sampling instant is integrated with the approach to further improve convergence speed. The approach is highly parallelized as constraints can be checked in parallel. The approach is demonstrated using an academic example as well as a chemical process network control.     

基于RISOMAP的非线性过程故障检测方法

化工过程监控数据存在非线性特点,且过程常常运行于多个模态,针对该类问题,提出基于相对等距离映射(relative isometric mapping, RISOMAP)的过程故障检测方法,该方法采用相对测地距离构造高维空间的距离关系阵,运用多维尺度变换(MDS)计算其低维嵌入输出,从高维数据中提取子流形信息和残差信息分别构造监控统计量进行故障检测,同时运用核ridge回归在线计算测试数据的低维输出,核矩阵通过综合相似度进行更新。数值算例和TE过程的仿真结果表明,RISOMAP方法可以更为有效地实施故障检测,故障检测的灵敏度较高,同时也为基于流形学习的多模态过程故障检测的实施提供了一条思路。     

基于Vague集的采矿方法优选研究

采矿方法是关系矿山企业生存与发展的重要因素,随着开采深度的增加,采矿方法的选择对于矿山生产具有极其重要的意义,因此有必要对采矿方法优选进行研究。将Vague集理论引入采矿方法选择中,目标集为矿山采矿方案,约束集为生产能力、采矿成本、千吨采切比、损失率、贫化率、安全状况、通风条件、劳动强度、工艺复杂程度、对矿体的适应性等10个指标,建立了采矿方案的Vague集模型,并以某铜铁矿为背景,提出了4种采矿方案,应用Vague集模型对其进行了优选,结果表明:盘区沿走向上向水平分层充填采矿法是最优的开采方案,这与专家讨论的结论一致,验证了Vague优选模型的实用性及可靠性。     

模型有界失配下基于MAP准则的扩展集员粒子滤波方法

在非线性非高斯系统中,当状态转移模型存在有界失配时,采用粒子滤波往往无法获得理想的状态估计值。考虑有界失配对粒子的约束条件,提出一种基于MAP准则的扩展集员粒子滤波算法（MAP-ESMPF）。该算法采用扩展集员求取真实状态的可信域,并基于MAP密度函数的准则,定义优化方程,从而将可信域外的粒子映射到可信域内,保证了状态估计的精度。在数值仿真和连续搅拌反应釜（CSTR）过程中的仿真应用,验证了该算法的有效性。     

一种基于改进KICA的非高斯过程故障检测方法

针对基于核独立元分析(kernel independent component analysis,KICA)的故障检测方法只考虑非高斯信息提取而忽略局部近邻结构保持的问题,提出基于改进KICA的过程故障检测方法。将KICA法中只考虑非高斯信息提取的负熵最大化准则转换为熵最小化准则,结合局部保持投影的相似局部近邻结构准则,提出了同时考虑非高斯信息提取和局部近邻结构保持的目标函数,通过粒子群优化算法进行全局寻优,然后建立监控统计量对过程进行监控。在Tennessee Eastman过程上的仿真结果说明,与基于KICA的故障检测方法相比,所提方法能够在保持数据集局部近邻结构的同时,提取非高斯信息,能够有效缩短故障检测的延迟时间,提高故障检测率。     

一种基于改进核Fisher的故障诊断方法

针对化工过程故障数据呈非线性分布,数据类别复杂,难以进行故障诊断的问题,提出一种改进核Fisher的故障诊断方法。对于原始样本数据投影后,样本出现因类间距离存在很大差异性而导致部分类别样本存在混叠的现象,以及不同类别的边界数据归类模糊问题,给出了统一的解决办法。该方法首先采用改进类间距的方法来改变样本投影空间的分布,使得样本具有较好的投影效果,然后通过定义阈值参数来筛选出边界数据,对于边界数据,采用改进的K近邻（K-NN）算法来分类,对于非边界数据,采用马氏距离来分类。最后在TE过程中进行了仿真验证,结果表明方法在兼顾了故障诊断时间的同时,有效提高了故障诊断精度。     

Multi-loop nonlinear internal model controller design based on a dynamic fuzzy partial least squares model

In this paper, a dynamic fuzzy partial least squares (DFPLS) modeling method is proposed. Under such framework, the multiple input multiple output (MIMO) nonlinear system can be automatically decomposed into several univariate subsystems in PLS latent space. Within each latent space, a dynamic fuzzy method is introduced to model the inherent dynamic and nonlinear feature of the physical system. The new modeling method combines the decoupling characteristic of PLS framework and the ability of dynamic nonlinear modeling in the fuzzy method. Based on the DFPLS model, a multi-loop nonlinear internal model control (IMC) strategy is proposed. A pH neutralization process and a methylcyclohexane (MCH) distillation column from Aspen Dynamic Module are presented to demonstrate the effectiveness of the proposed modeling method and control strategy.     

A tree search methodology based on sequential dynamic programming and set analysis

This methodology permits a systematic tree search using a rank ordered list of alternative structures. After initially using all of the subproblems as the data base, dynamic programming (DP) is repeatedly completed on subsets of this data base. As sequences are identified, their subproblem identification numbers are stored and used to reduce the number of subproblem combinations (i.e. DP trials) that must be completed to generate either the complete rank order list or a part of it. Then starting with the apparently best, next best and subsequent sequences, the engineer can systematically evaluate process interactions and integration options to find the best design overall. The structural accuracy of the initial rank order list depends upon the interactions between subproblems and energy integration options for the network.     

基于加权概率CVDA的动态化工系统微小故障检测

典型变量差异度分析（CVDA）是近年来提出的一种新型动态过程监控方法,已在微小故障检测领域获得成功应用。针对传统CVDA方法忽视了特征量的概率信息挖掘问题,提出一种基于加权概率CVDA（WPCVDA）的动态化工系统微小故障检测方法。一方面,该方法在基本CVDA模型特征基础上引入Wasserstein距离（WD）度量特征量概率分布的变化,构造概率化的WD特征提高CVDA模型对微小故障的灵敏度;另一方面,进一步考虑不同的WD特征成分携带故障信息的差异性,设计一种自适应权值计算策略,为关键的故障敏感特征成分设置大的权值,突出其在监控统计量中的作用。在一个标准化工过程的验证结果说明,所提出的WPCVDA方法比传统CVDA方法具有更好的微小故障检测性能。     

基于粗糙集理论的决策树构造算法

应用粗糙集理论,提出了一种利用新的启发式函数构造决策树的方法。该方法以属性重要性评价指标作为信息熵函数,对条件属性进行选择,充分考虑了属性间的依赖性和冗余性,弥补了ID3算法对属性间依赖性强调不够的缺点,解决了决策树中子树的重复和有些属性在同一决策树上被多次选择的问题,该方法还能对不相容决策表进行正确分类。实例表明该方法是正确有效的,而且明显优于传统的决策树构造方法。     

基于多层优化PCC-SDG方法的化工过程故障诊断

化工过程的故障发生往往都是一个变量带动多个变量的连锁效应。本文基于变量的相关性变化特点,用符号有向图SDG（signed directed graph）描述系统因果影响关系,以皮尔逊相关系数PCC（Pearson correlation coefficient）计算网络统计指标,提出了一种基于多层优化PCC-SDG的故障诊断方法。该方法基于全工艺的网络拓扑结构,首先对选取的变量进行初步优化。然后,为有效提取工艺特征信息,运用PCA（principal component analysis）权重思想从多层相关系数集中选取了权重较大的关键变量,结合SDG建立最优PCC-SDG网络。最后,针对最优PCC-SDG网络变量的相关性规律重构聚集权重系数Q,进行过程故障检测与诊断。TE（Tennessee Eastman）仿真过程的应用结果表明,PCC-SDG建模及故障诊断步骤较为简洁,可以充分挖掘SDG深层次关联特性,定量简化SDG的故障诊断效果明显,具有较好的过程监控优势。     

基于模糊综合评价的烟道气辅助SAGD过程评价

目前烟道气辅助蒸汽辅助重力泄油（SAGD）技术已在机理研究和数值模拟方面取得了进展,但由于实际投入油田试验的操作成本较高以及存在额外的能量消耗,所以无法直接判断和验证实际应用效果。为了更全面地评估注采方案,本文运用模糊综合评价法建立了多元评价体系,以环境、能量、工艺、经济为4个评价指标,对不同注采方案进行综合评价以选出综合性能最优的生产设计方案,为实际生产方案的选取提供参考,首次对烟道气辅助SAGD实际工程进行评价研究。基于“有无对比法”,将烟道气辅助SAGD与常规SAGD进行对比,结果表明烟道气辅助SAGD的综合效益更好,佐证了烟道气辅助SAGD驱油的优越性。为方便工程应用,开发了“烟气辅助SAGD驱油评价软件”,通过软件进行实例分析并进行方法对比,验证了本文方法的有效性、适用性和准确性。