基于神经网络的化工过程测量数据在线校正技术的研究

研究了人工神经网络在化工过程测量数据校正中的应用，提出了新的样本构造方法和神经网络的在线训练策略。对某乙烯装置裂解气分离系统测量数据，应用自行设计开发的改进算法的神经网络与数据校正系统集成运行，结果表明基于神经网络的数据校正技术能对测量数据中所含的随机误差和过失误差进行同时校正，提高过程数据的精度和校正过程的稳定性，同时满足数据校正的实时性要求。     

一种改进的鲁棒数据校正方法

鲁棒数据校正方法能有效得进行数据协调和过失误差侦破。然而,当泄露、积累等问题导致模型不准确时,传统的鲁棒数据校正技术无法得到满意的结果。针对这种情况,本文对鲁棒数据校正技术做了改进,将相对残差大于临界值的等式约束变为不等式约束,同时将原来等式约束的相对残差加入到鲁棒目标函数中,使得该方法适用于过程模型不准确的情况。实例运算表明,本文提出的方法对模型准确或者模型不准确的情况都有较好的数据校正结果。     

化工过程数据校正方法的探讨

简要地介绍了数据校正中的基本概念;然后从数据协调、显著误差检测、数据校正三个方面较系统地、全面地总结了数据校正方法的研究进展和成就;最后探讨了这一领域中值得进一步研究的问题和可能的发展方向。     

关于化工数据校正问题的研究

介绍化工过程数据校正和参数估计的基本原理。提出对化工过程数据进行分类的新方法─—有序规则法，为正确设计测量点的位置提供了理论依据；提出非线性数据校正问题的线性化方法，有效地避免了易使校正问题繁解的非线性方程的出现。     

基于双权M-估计子的鲁棒数据校正新方法

准确可靠的测量数据是实现装置过程控制、模拟、优化和生产管理的前提条件,而通过仪表测量获取的过程数据中可能存在过失误差,直接影响数据校正的准确性,现有的数据校正方法不能完全有效地避免过失误差的影响。根据双权M-估计的原理,今以相对残差为变量构造了一种新型的具有强鲁棒性的目标函数,使含有过失误差的变量对函数的贡献为一常数,从而避免了过失误差对数据校正过程的影响。选取了具有代表性的一个线性问题和一个非线性问题进行实例研究,并与现有的Huber法和Cauchy法进行了对比分析。计算结果表明,对线性系统和非线性系统,新方法的过失误差侦破性能均优于Huber法和Cauchy法,且其稳定性更高。因此,在进行数据校正时应首选新方法。     

复杂大系统的数据校正

由于复杂大系统包含设备较多、物料品种多、工艺流程复杂,所需测量的数据繁多、解题规模庞大.传统的数据校正算法耗费了计算机大量的存储单元,而且计算时间较长.而新的分解协调算法无需协调参数和迭代计算,故计算速度加快,很好地解决了这个问题.     

基于模糊自适应遗传算法的数据校正方法

针对数据校正中的非线性数据分类还比较困难的问题,提出表上作业法,结合遗传算法同时进行数据协调与过失误差侦破.并将模糊数据协调模型中的三角形约束引入遗传算法界定其变异上下限,还对交叉算子进行自适应改进,从而形成了基于模糊自适应的数据校正方法并用该方法对一个稳态多组分精馏过程进行大量仿真试验,结果表明了算法的有效性.     

数据样正在精馏塔系中的应用

针对某芳烃联合装置中精馏塔系的双线性数据校正问题，通过此入测度因子函数，将双一问题转化为线性数据校正问题，进行了稳态数据校正的研究，得出了该类问题的一种解法，仿真实验表明该方法迅速、准确。     

一种用于显著误差检测与数据校正的NT-MT组合算法

An NT-MT combined method based on nodal test （NT） and measurement test （MT） is developed for gross error detection and data reconciliation for industrial application. The NT-MT combined method makes use of both NT and MT tests and this combination helps to overcome the defects in the respective methods. It also avoids any artificial manipulation and eliminates the huge combinatorial problem that is created in the combined method based on the nodal test in the case of more than one gross error for a large process system. Serial compensation strategy is also used to avoid the decrease of the coefficient matrix rank during the computation of the proposed method. Simulation results show that the proposed method is very effective and possesses good performance.     

数据校正在精馏塔系中的应用

针对某芳烃联合装置中精馏塔系的双线性数据校正问题，通过引入测度因子函数，将双线性问题转化为线性数据校正问题，进行了稳态数据校正的研究，得出了该类问题的一种解法。仿真实验表明该方法迅速、准确     

面向对象的炼油厂全厂调度优化模型及程序框架

基于考虑炼油装置优化操作模式切换过程的总体思想,构建了一套炼油厂全厂调度优化离散时间模型结构,并形成配套的程序框架。采用面向对象的建模方式,引入模态指示矩阵等表达,为炼油厂生产调度建模提供了较为清晰的参考思路。通过GAMS和MATLAB的数据交互,实现二者优势互补,为进一步研究炼油生产调度模型提供便利、奠定基础。案例研究验证了所提模型结构及程序框架的有效性。     

Sparse probabilistic principal component analysis model for plant-wide process monitoring

In the industrial monitoring process, probabilistic principal component analysis (PPCA) is a popular algorithm for reducing the dimension. However, the principal components (PCs) are not easy to interpret and its preserved number cannot be determined automatically. In this paper, we propose a sparse PPCA (SPPCA) to improve the interpretability by adding a prior and introducing sparsification to the loading matrix of PPCA. An expectation-maximization (EM) algorithm is used to obtain the parameters of the probabilistic formulation, and the dimensionality of the latent variable space can be automatically determined during the iterative process. With the sparse representation, a process monitoring strategy is then developed with the construction of several partial PPCA models. Case studies of SPPCA to a numerical case and Tennessee Eastman (TE) benchmark process demonstrate its feasibility and efficiency.     

A multiscale mass transfer model for gas-solid riser flows: Part 1 — Sub-grid model and simple tests

The gas-solid mass transfer in circulating fluidized bed (CFB) riser flow is both structure-dependent and dynamic in nature. Recent progress in multiscale computational fluid dynamics (CFD) allows fresh insight into the dynamic flow structure, yet its influence on the mass transfer remains to be settled. To this end, a multiscale mass transfer model is established in this paper based on the extended framework of the energy-minimization multiscale (EMMS) model. The relevant algorithm named EMMS/mass is proposed for CFD-coupled mass transfer computation. Two testing cases accounting for sublimation of naphthalene and decomposition of ozone, respectively, are presented to demonstrate the characters of the model. It is shown that structural consideration can have significant effects on the model prediction. The normally used Reynolds number is not adequate to characterize these effects, while the combination of gas velocity and solids flux seems to capture the structural effects and allows to explain the variation of Sherwood number reported for CFB risers in the literature. Sub-grid coupling of this multiscale mass transfer model and CFD approach can be expected to provide a promising tool to probe the dynamic and structure-dependent nature of mass transfer in CFB risers.     

Segregated flows in model cells

The flow of an electrolyte in a parallel channel electrochemical reactor has been simulated in a simple Perspex flow channel, operating at Reynolds numbers between 20–1000. Age distribution information was obtained by the marker-pulse technique and analysed with the aid of various simple mathematical models of the flow. It was found that the flow appeared to be segregated and that, in particular, electrogenerated species were retained for long periods in the vicinity of the electrode. Regression was performed in both the time domain and Laplace plane and the effect of experimental noise on the accuracy of the results was estimated. Fitting in real-time was found to be inherently less sensitive to noise than fitting in the Laplace plane. Data are given for the highly structured flow within a model of an Mg-AgCl sea water battery as well as for channel flow.     

Dantzig–Wolfe decomposition and plant-wide MPC coordination

Due to the enormous success of model predictive control (MPC) in industrial practice, the efforts to extend its application from unit-wide to plant-wide control are becoming more widespread. In general, industrial practice has tended toward a decentralized MPC architecture. Most existing MPC systems work independently of other MPC systems installed within the plant and pursue a unit/local optimal operation. Thus, a margin for plant-wide performance improvement may be available beyond what decentralized MPC can offer. Coordinating decentralized, autonomous MPC has been identified as a practical approach to improving plant-wide performance. In this work, we propose a framework for designing a coordination system for decentralized MPC which requires only minor modification to the current MPC layer. This work studies the feasibility of applying Dantzig–Wolfe decomposition to provide an on-line solution for coordinating decentralized MPC. The proposed coordinated, decentralized MPC system retains the reliability and maintainability of current distributed MPC schemes. An empirical study of the computational complexity is used to illustrate the efficiency of coordination and provide some guidelines for the application of the proposed coordination strategy. Finally, two case studies are performed to show the ease of implementation of the coordinated, decentralized MPC scheme and the resultant improvement in the plant-wide performance of the decentralized control system.     

Effects of turbulence on heat and mass transfer in stagnation flows

The influence of turbulence on local heat and mass transfer is systematically analyzed. Essential prerequisite is the availability of measurement methods, based on convective mass transfer, which allow the visualization of heat and mass transfer distributions. The results show that, even at low turbulence intensities, the flow and transport phenomena are decisively influenced by the specific design of turbulence grids and the orientation of the grid wires with respect to the stagnation line. Thereby, two different flow mechanisms occur which may interact. The anisotropy of grid turbulence leads to the “wire-gap effect” depending on the actual position of the grid wires in relation to the stagnation line. The other mechanism is based on the wavy wake velocity distribution behind screens and grid which, in conjunction with the centrifugal instability of stagnation flows, may lead to longitudinal vortices. The results indicate why turbulence has not been sufficiently considered in the available (semi-) empirical correlations. Information needed for an accurate prediction of heat and mass transfer in turbulence flows is discussed.     

A multiscale mass transfer model for gas-solid riser flows: Part II—Sub-grid simulation of ozone decomposition

This article is to test the EMMS-based multiscale mass transfer model through computational fluid dynamics (CFD) simulation of ozone decomposition in a circulating fluidized bed (CFB) reactor. Three modeling approaches, namely types A, B and C, are classified according to their drag coefficient closure and mass transfer equations. Simulation results show that the routine approach (type C) with assumption of homogeneous flow and concentration overestimates the ozone conversion rate, introduction of structure-dependent drag force will improve the model prediction (type B), while the best fit to experimental data is obtained by the multiscale mass transfer approach (type A), which takes into account the sub-grid heterogeneity of both flow and concentration. In general, multiscale behavior of mass transfer is more distinct especially for the dense riser flow. The fair agreement between our new model with literature data suggests a fresh paradigm for the CFB related reaction simulation.     

IDEAL算法在非稳态两相流问题中的性能分析

对于非稳态两相流问题,联立求解连续性方程和动量方程最为广泛使用的方法为：分步方法和SIMPLE系列算法。分步方法的优点是：收敛速度快,缺点是：初值问题条件稳定。SIMPLE系列算法优点是：初值问题绝对稳定,缺点是：收敛速度慢。为了克服传统的SIMPLE系列算法收敛速度慢这一缺点,本文引入了IDEAL算法。本文通过2个非稳态的两相流问题对SIMPLER和IDEAL两种不同算法在收敛速度方面进行了比较。通过比较可以看出IDEAL算法的收敛速度较SIMPLER算法得到了大幅提高,提高了约6~87倍,克服了传统SIMPLE系列算法收敛速度较慢的缺点,同时具有了初值问题绝对稳定和收敛速度快这两个优点。     

Study of thermal rectification in a column with low mass transfer on the steps

The thermal rectification method, which comprises the condensing of vapors on steps and subsequent exposure of the heat flow on condensate before it is introduced into the falling reflux, has been studied. The results of the study of strengthening ethanol–water mixture upon the condensation of vapors on the heat-transfer surface have been presented. The basic parameters of the process have been revealed, which ensure the strengthening of the mixture on the stage upon condensate evaporation, and the empirical relationship for determining the effectiveness has been obtained. The studies of the thermal rectification of the ethanol–water mixture have been performed in a column with 24 contact stages made of horizontal plates, and conditions have been established that provide the greatest efficiency. The evaluation of thermal rectification contribution to the overall efficiency of the stage is performed that allows one to intensify the process by up to four times.     

Mercury species, mass flows and processes in a cement plant

The aim of the study was to evaluate the behaviour of mercury in the cement clinker production process. Simultaneous measurements of mercury in all important materials and gas streams were performed in three sampling periods on about 300 solid samples and about 80 samples taken from gas streams. Mercury species in flue gases at characteristic parts of the process were measured as total Hg(t), particulate Hg(p), elemental Hg⁰(g) and reactive Hg²⁺(g) mercury. Based on the results of measurements, mercury mass flows and mass balances of the whole and in certain parts of the process were evaluated. It was shown that the process comprises many mercury cycles which are strongly dependent on the operating conditions and technological specifics. Cycling of mercury causes a significant enrichment of mercury inside the process. In the annual mercury input of about 27 kg, raw materials accounted near by 60% and fuels about 40% (i.e. petrol coke 31%, waste tyres 10% and waste oil 0.4%). The annual emission of mercury represented 40-70% of the inputs, while cement clinker only contained about 10%. The difference between inputs and outputs (11-45%) obtained in the annual mass balance could be assigned to mercury recycling and significant enrichment inside the process, as well as variability between spot measurements. The parts of the process with the highest mercury mass flows and the lowest material/gas flows were identified. Such points represent an opportunity to remove a significant amount of mercury from the process at low material flows and to improve mercury control. Mercury was mainly emitted in gaseous form with 92% (direct mode) or 89% (combined mode) as Hg(g) on average, of which about 2/3 was as Hg²⁺(g), and about 1/3 as Hg⁰(g). Only a small part (the rest) was emitted as particulate Hg(p). Shares of individual mercury species in the last sampling period were 65.7% Hg²⁺(g), 34.0% Hg⁰(g) and 0.3% Hg(p) on average. Ratios between individual mercury forms were found to be related to operating modes. The quantities of Hg(t), Hg(g) and Hg²⁺(g) emitted were higher when operating with the raw mills off (direct mode). It was seen that the efficiency of Hg removal was strongly related to the dust removal efficiency. Bag filters very efficiently removed all mercury species.