污水处理厂出水水质变量区间预测建模

为了实现污水处理厂的有效运行,需要建立能够精确描述水厂行为的模型。根据水厂入水和出水数据,采用径向基函数神经网络建立污水处理过程模型。考虑到建模误差有界,使用参数线性集员辨识算法分别得到隐含层到输出层各神经元连接权值向量的不确定集合描述。与现有的单输出区间预测模型相比,该模型能够根据水厂入水数据同时给出多个出水水质变量的置信区间。这些区间能表征出水变量的存在范围,从而实现水质变量的可靠估计,进而评估出水水质或水厂性能。此外,还将此出水区间预测模型用于污水处理厂的故障检测与隔离,以提高水厂运行的可靠性。实验结果表明文中所提方法的有效性。     

基于RBF神经网络的轮胎滚动阻力建模研究

建立径向基函数(RBF)神经网络轮胎滚动阻力模型,充分利用RBF神经网络模型逼近精度高、训练速度快、无局部极小等优点,对轮胎滚动阻力进行全面、准确的预测。结果表明,轮胎滚动阻力RBF与反向传播算法(BP)神经网络模型预测值的平均相对误差分别为2%和6%左右,RBF神经网络模型在训练和预测结果上均有更大优势,能够有效预测轮胎滚动阻力。     

基于新型联合循环神经网络(RNN)模型的出水总氮预测

为了提高污水处理过程中出水TN的预测精度,提出一种基于联合RNN神经网络的出水TN预测算法.该联合RNN神经网络是基于传统RNN网络的改进型网络,其由一系列并排的单RNN网络构成,每个RNN网络的输出不仅与本网络的当前和历史输入相关,还与相邻RNN网络的历史输入相关.网络的训练过程分为两步,首先断开不同RNN网络的连接...     

基于模糊RBF神经网络的乙烯装置生产能力预测

针对传统的径向基函数(RBF)神经网络隐藏层节点的不确定和初始中心敏感性、收敛速度过慢等问题,提出一种基于模糊C均值的RBF神经网络(FCM-RBF)模型,通过模糊C均值聚类(FCM)得到各聚类中心,基于误差反传的梯度下降法训练隐藏层到输出层之间的权值,克服传统RBF模型对数据中心的敏感性,优化确定RBF神经网络隐藏层的节点数,提高网络训练速度和精度。最后将其用于乙烯装置生产能力预测中,分析预测不同技术、不同规模乙烯装置生产情况,指导乙烯生产,提高生产效率,结果验证了所提出算法的有效性和实用性。     

基于SOM-RBF神经网络的COD软测量方法

污水处理是一个复杂的非线性过程,化学需氧量（chemical oxygen demand,COD）是评价污水处理效果的关键指标之一。COD的传统测量方法耗时长、成本高,基于传统神经网络的软测量方法提高了COD参数的测量速度但精度较差。针对这些问题,设计一种结合自组织特征映射 (self-organizing map, SOM)和径向基函数(radial basis function, RBF)神经网络的COD参数软测量方法。该方法利用SOM网络聚类数据样本,根据所得聚类结果确定RBF网络的隐层节点数及节点的数据中心,综合提高RBF网络的收敛速度和拟合精度。利用污水处理厂部分水样数据建立COD软测量模型,模型仿真和硬件在线测试结果表明,相对于传统的BP、RBF等网络,基于SOM-RBF神经网络的COD软测量方法测量时间短、预测精度较高,具有较为广阔的应用前景。     

Process development of treatment plants for dyeing wastewater

A three‐step methodology that integrates experiments, modeling and synthesis has been developed for the systematic development of a plant for treating dyeing wastewater for discharge and/or reuse. First, wastewater characteristics, discharge water standards, and reuse water quality specifications, etc. are collected as input information. Heuristics developed in our industrial practice and gleaned from the literature are used to guide the designer to come up with preliminary flow sheet alternatives. Then, bench‐scale experiments and pilot plant tests for the relevant unit operations are performed. A computer code accepts the bench‐scale and pilot plant experimental data for regression of model parameters and determines the superior process configuration and equipment operating conditions through sensitivity analysis. The workflow among various stakeholders to reach the final design is presented. Possible extension of the methodology to other industrial wastewater treatment plants is discussed. © 2011 American Institute of Chemical Engineers AIChE J, 2012     

Neural virtual sensor for the inferential prediction of product quality from process variables

A predictive Fuzzy ARTMAP neural system and two hybrid networks, each combining a dynamic unsupervised classifier with a different kind of supervised mechanism, were applied to develop virtual sensor systems capable of inferring the properties of manufactured products from real process variables. A new method to construct dynamically the unsupervised layer was developed. A sensitivity analysis was carried out by means of self-organizing maps to select the most relevant process features and to reduce the number of input variables into the model. The prediction of the melt index (MI) or quality of six different LDPE grades produced in a tubular reactor was taken as a case study. The MI inferred from the most relevant process variables measured at the beginning of the process cycle deviated 5% from on-line MI values for single grade neural sensors and 7% for composite neural models valid for all grades simultaneously.     

一种基于自适应回归核函数的污水处理能耗模型

针对污水处理过程能耗模型难以建立的问题,提出了一种基于自适应回归核函数的建模方法。通过分析污水处理过程的运行特点,构建能耗与运行过程变量之间的关系,得到一种基于过程变量的能耗模型表达;同时利用梯度下降算法对能耗模型参数进行自适应调整,提高模型精度。最后,将设计的能耗模型应用于污水处理过程基准仿真平台BSM1和实际污水处理厂,实验结果表明该模型能够根据污水处理过程变量实时获得污水处理过程的能耗,具有较好的自适应特性和较高的精度。     

Handling sensor faults in economic model predictive control of batch processes

The problem of sensor fault detection and isolation (FDI) and fault-tolerant economic model predictive control (FT-EMPC) for batch processes is addressed. To this end, we first model batch processes using subspace-based system identification techniques. The analytical redundancy within the identified model is subsequently exploited to detect, isolate, and handle the faulty measurements. The reconciled fault-free measurements are then incorporated in an economic model predictive controller formulation. Simulation case studies involving the application of the proposed data-driven FDI and FT-EMPC algorithms to the energy intensive electric arc furnace process illustrate the improvement in economic performance under various fault scenarios. © 2018 American Institute of Chemical Engineers AIChE J, 65: 617–628, 2019     

Kinetic model for hydroisomerization reaction of C8-aromatics

Based on the reported reaction networks, a novel six-component hydroisomerization reaction network with a new lumped species including C₈-naphthenes and C₈-paraffins is proposed and a kinetic model for a commercial unit is also developed. An empirical catalyst deactivation function is incorporated into the model accounting for the loss in activity because of coke formation on the catalyst surface during the long-term operation. The Runge-Kutta method is used to solve the ordinary differential equations of the model. The reaction kinetic parameters are benchmarked with several sets of balanced plant data and estimated by the differential variable metric optimization method (BFGS). The kinetic model is validated by an industrial unit with sets of plant data under different operating conditions and simulation results show a good agreement between the model predictions and the plant observations. __________ Translated from Journal of Chemical Engineering of Chinese Universities, 2007, 21(3): 429–435 [译自: 高校化学工程学报]