An artificial neural-network model for impact properties in X70 pipeline steels

An artificial neural-network (ANN) model has been developed for the analysis and simulation of the correlation between the mechanical properties and composition and thermomechanical treatment parameters of high strength, low alloy steels. The input parameters of the model consist of alloy compositions (C, Si, Mn, P, S, Cu, Ni, Cr, Mo, Ti, V, Nb, Ca, Al, B) and tensile test results (yield strength, ultimate tensile strength, percentage elongation). The outputs of the ANN model include impact energy (?10 °C). The model can be used to calculate the properties of low alloy steels as a function of alloy composition and thermomechanical treatment variables. The current study achieved a good performance of the ANN model, and the results are in agreement with experimental knowledge.     

Prediction of the natural gas consumption in chemical processing facilities with genetic programming

Cinkarna Ltd. is a chemical processing company in Slovenia and the country’s largest manufacturer of titanium oxides (TiO₂). Chemical processing and titanium oxide manufacturing in particular requires high natural gas consumption, and it is difficult to accurately pre-order gas from suppliers. In accordance with the Energy Agency of the Republic of Slovenia regulations, each natural gas supplier regulates and determines the charges for the differences between the ordered (predicted) and the actually supplied quantities of natural gas. Yearly charges for these differences total 1.11 % of supplied natural gas costs (average 50,960 EUR per year). This paper presents natural gas consumption prediction and the minimization of associated costs. The data on daily temperature, steam boilers, sulfur acid and TiO₂ production was collected from January 2012 until November 2014. Based on the collected data, a linear regression and a genetic programming model were developed. Compared to the specialist’s prediction of natural gas consumption, the linear regression and genetic programming models reduce the charges for the differences between the ordered and the actually supplied quantities by 3.00 and 5.30 times, respectively. Also, from January until November 2014 the same genetic programming model was used in practice. The results show that in a similar gas consumption regime the differences between the ordered and the actually supplied quantities are statistically significant, namely, they are 3.19 times lower (t test, p < 0.05) than in the period in which the specialist responsible for natural gas consumption made the predictions.     

Elementary function generators for neural-network emulators

Piecewise first- and second-order approximations are employed to design commonly used elementary function generators for neural-network emulators. Three novel schemes are proposed for the first-order approximations. The first scheme requires one multiplication, one addition, and a 28-byte lookup table. The second scheme requires one addition, a 14-byte lookup table, and no multiplication. The third scheme needs a 16-byte lookup table, no multiplication, and no addition. A second-order approximation approach provides better function precision; it requires more hardware and involves the computation of one multiplication and two additions and access to a 28-byte lookup table. We consider bit serial implementations of the schemes to reduce the hardware cost. The maximum delay for the four schemes ranges from 24- to 32-bit serial machine cycles; the second-order approximation approach has the largest delay. The proposed approach can be applied to compute other elementary function with proper considerations.     

Estimations of error bounds for neural-network functionapproximators

Neural networks are being increasingly used for problems involving function approximation. However, a key limitation of neural methods is the lack of a measure of how much confidence can be placed in output estimates. In the last few years many authors have addressed this shortcoming from various angles, focusing primarily on predicting output bounds as a function of the trained network's characteristics, typically as defined by the Hessian matrix. In this paper the problem of the effect of errors or noise in the presented input vector is examined, and a method based on perturbation analysis of determining output bounds from the error in the input vector and the imperfections in the weight values after training is also presented and demonstrated.     

一种自然人工对象识别与诊断的方法

为了能有效提高识别自然人造对象并诊断其缺陷的效率,提出了一种基于改进的形状上下文实现采集图像和样本图像匹配的方法。该方法对采样图像进行轮廓均匀采样并改进其形状上下文,将采样图像与样本图像进行匈牙利匹配,计算出所有匹配点间欧式距离。对采样图像进行TPS变换,计算匹配点间的欧式距离;根据TPS变换前后匹配点间距离变化比例判定样本图像与采样图像是否是同类图像;在同类图像的基础上计算形状上下文距离,根据匹配代价衡量匹配点集中匹配上的点数实现图像的缺陷诊断。该方法同时实现了图像识别与缺陷诊断,实验结果较好,能应用到架空输电线路直升机巡检系统中。     

乙烯裂解炉燃料气消耗的实时优化

乙烯装置作为石化行业能耗大户,在乙烯装置能量优化过程中,炉群系统能耗优化起到至关重要的作用。在保证工业装置产品收率不变的情况下,本文采用调整操作变量裂解炉出口温度,达到炉群整体燃料气消耗降低的目的。本文采用K均值聚类算法结合即时学习局部建模方法,建立了精确的燃料气消耗预测模型,模型平均绝对百分比误差为0.0626%,相对误差在5%以内,满足实际工业过程对预测模型精度的要求。以某组工业数据为例,通过差分进化算法,炉群整体燃料气消耗量降低2.5%,有效的通过操作变量优化达到整体乙烯装置经济效益提高。     

ScaleNet-multiscale neural-network architecture for time series prediction.

The effectiveness of a multiscale neural net architecture for time series prediction of nonlinear dynamic systems is investigated. The prediction task is simplified by decomposing different scales of past windows into different scales of wavelets, and predicting the coefficients of each scale of wavelets by means of a separate multilayer perceptron. The short-term history is decomposed into the lower scales of wavelet coefficients, which are utilized for detailed analysis and prediction, while the long-term history is decomposed into higher scales of wavelet coefficients that are used for the analysis and prediction of slow trends in the time series. These coordinated scales of time and frequency provide an interpretation of the series structures, and more information about the history of the series, using fewer coefficients than other methods. Results concerning scales of time and frequencies are combined by another expert perceptron, which learns the weight of each scale in the goal-prediction of the original time series. Each network is trained by backpropagation. The weights and biases are initialized by a clustering algorithm of the temporal patterns of the time series, which improves the prediction results as compared to random initialization. The suggested multiscale architecture outperforms the corresponding single-scale architectures. The employment of improved learning methods for each of the ScaleNet networks can further improve the prediction results.     

Determination of neural-network topology for partial dischargepulse pattern recognition

A time-series approach has been employed to devise neural-network topologies for time dependent partial discharge pulse pattern recognition applications. The cascaded output neural-network structure was found to provide the highest success rate in differentiating between two different partial discharge patterns. This was accomplished by utilizing the indexed feature of the first stage output as one of the inputs into the second stage of the cascaded neural network.     

A parameter identification method for natural gas supply systems under unsteady gas flow

We consider the identification problem for the technical parameters, i.e., hydraulic efficiency coefficients, of gas transmission systems (GTS) under standard unsteady gas flow. Coefficient estimation is reduced to a conditional optimization problem with equality type constraints. We propose an algorithm to solve it. The algorithm’s efficiency has been tested with computational experiments, including a looped gas supply system. The resulting model operates under the current level of GTS information support.     

Inverse lighting design for interior buildings integrating natural and artificial sources

In this paper we propose a new method for solving inverse lighting design problems that can include diverse sources such as diffuse roof skylights or artificial light sources. Given a user specification of illumination requirements, our approach provides optimal light source positions as well as optimal shapes for skylight installations in interior architectural models. The well known huge computational effort that involves searching for an optimal solution is tackled by combining two concepts: exploiting the scene coherence to compute global illumination and using a metaheuristic technique for optimization.Results and analysis show that our method provides both fast and accurate results, making it suitable for lighting design in indoor environments while supporting interactive visualization of global illumination.     

Simulation model for natural gas transmission pipeline network system

This paper focuses on developing a simulation model for the analysis of transmission pipeline network system (TPNS) with detailed characteristics of compressor stations. Compressor station is the key element in the TPNS since it provides energy to keep the gas moving. The simulation model is used to create a system that simulates TPNS with different configurations to get pressure and flow parameters. The mathematical formulations for the TPNS simulation were derived from the principles of flow of fluid through pipe, mass balance and compressor characteristics. In order to determine the unknown pressure and flow parameters, a visual C++ code was developed based on Newton–Raphson solution technique. Using the parameters obtained, the model evaluates the energy consumption for various configurations in order to guide for the selection of optimal TPNS. Results from the evaluations of the model with the existing TPNS and comparison with the existing approaches showed that the developed simulation model enabled to determine the operational parameters with less than 10 iterations. Hence, the simulation model could assist in decisions regarding the design and operations of the TPNS.     

天然气脱酸气装置动态建模及分析

天然气脱酸气装置是天然气开发过程中重要的组成部分,由于脱酸气装置的来气量经常发生变动,致使装置运行偏离最优工况。本文基于川渝地区某高含硫天然气脱酸气装置,首次应用HYSYS Dynamic流程模拟软件,建立天然气脱酸气装置动态模型,研究脱酸气装置在原料气的进气量在80%～130%负荷变化时,整套装置的动态响应过程,确定了常规PID控制方案对吸收剂循环量控制滞后,造成装置在高负荷下,产品气质量达不到标准;针对动态分析发现的问题,本文提出了一种比例控制方案,依据原料气的处理量的变化比例调节吸收剂循环量,并根据进入再生塔的富吸收液流量比例调节再生塔重沸器的负荷。动态分析表明,在原料气负荷变动条件下,装置可以始终保持产品气质量在要求之内,再生塔负荷也可以根据处理量的变化而自动调节,在低负荷运行时,降低装置的能耗。此比例控制方案大大提高了装置的操作弹性。     

基于改进神经网络的非线性系统观测器设计

根据非线性系统利用前馈网络的函数逼近能力,设计了一种神经网络观测器,并利用网络权值校正法,建立Lyapunov函数对观测器的稳定性进行了分析。为了加快训练速度,在训练网络时采用LM优化算法来实现,仿真结果不仅证明了所设计的神经网络观测器的有效性,还证实了神经网络改进算法后的优越性。     

A test for the homoscedasticity of the residuals in fuzzy rule-based forecasters

Heteroscedasticy is the property of having a changing variance throughout the time. Homoscedasticity is the converse, that is, having a constant variance. This is a key property for time series models which may have serious consequences when making inferences out of the errors of a given forecaster. Thus it has to be conveniently assessed in order to establish the quality of the model and its forecasts. This is important for every model including fuzzy rule-based systems, which have been applied to time series analysis for many years. Lagrange multiplier testing framework is used to evaluate wether the residuals of an FRBS are homoscedastic. The test robustness is thoroughly evaluated through an extensive experimentation. This is another important step towards a statistically sound modeling strategy for fuzzy rule-based systems.     

Development of a predictive system for car fuel consumption using an artificial neural network

A predictive system for car fuel consumption using a back-propagation neural network is proposed in this paper. The proposed system is constituted of three parts: information acquisition system, fuel consumption forecasting algorithm and performance evaluation. Although there are many factors which will effect the fuel consumption of a car in a practical drive procedure, however, in the present system the impact factors for fuel consumption are simply decided as make of car, engine style, weight of car, vehicle type and transmission system type which are used as input information for the neural network training and fuel consumption forecasting procedure. In the fuel consumption forecasting, to verify the effect of the proposed predictive system, an artificial neural network with back-propagation neural network has a learning capability for car fuel consumption prediction. The prediction results demonstrated that the proposed system using neural network is effective and the performance is satisfactory in fuel consumption prediction.     

Prediction of building energy consumption by using artificial neural networks

In this study, the main objective is to predict buildings energy needs benefitting from orientation, insulation thickness and transparency ratio by using artificial neural networks. A backpropagation neural network has been preferred and the data have been presented to network by being normalized. The numerical applications were carried out with finite difference approach for brick walls with and without insulation of transient state one-dimensional heat conduction. Three different building samples with different form factors (FF) were selected. For each building samples 0–2.5–5–10–15 cm insulations are assumed to be applied. Orientation angles of the samples varied from 0° to 80° and the transparency ratios were chosen as 15–20–25%. A computer program written in FORTRAN was used for the calculations of energy demand and ANN toolbox of MATLAB is used for predictions. As a conclusion; when the calculated values compared with the outputs of the network, it is proven that ANN gives satisfactory results with deviation of 3.43% and successful prediction rate of 94.8–98.5%.     

Finite-difference method for solving gas dynamics equations using adaptive artificial viscosity

A finite-difference method is proposed for solving gas dynamics equations, i.e., a homogeneous, monotonous finite-difference scheme of the second- order time approximation and space variables outside the areas of discontinuities and compression waves. A new way to introduce adaptive artificial viscosity (AAV) in a difference scheme is considered. The stability of the proposed difference scheme is numerically studied. Test calculations are presented for the motion of contact discontinuities, blast waves, and disintegration of discontinuities.     

STRIP - a strip-based neural-network growth algorithm for learningmultiple-valued functions

We consider the problem of synthesizing multiple-valued logic functions by neural networks. A genetic algorithm (GA) which finds the longest strip in V subset, dbl equalsK(n) is described. A strip contains points located between two parallel hyperplanes. Repeated application of GA partitions the space V into certain number of strips, each of them corresponding to a hidden unit. We construct two neural networks based on these hidden units and show that they correctly compute the given but arbitrary multiple-valued function. Preliminary experimental results are presented and discussed.     

A gradual neural-network approach for frequency assignment insatellite communication systems

A novel neural-network approach called gradual neural network (GNN) is presented for a class of combinatorial optimization problems of requiring the constraint satisfaction and the goal function optimization simultaneously. The frequency assignment problem in the satellite communication system is efficiently solved by GNN as the typical problem of this class. The goal of this NP-complete problem is to minimize the cochannel interference between satellite communication systems by rearranging the frequency assignment so that they can accommodate the increasing demands. The GNN consists of NxM binary neurons for the N-carrier-M-segment system with the gradual expansion scheme of activated neurons. The binary neural network achieves the constrain satisfaction with the help of heuristic methods, whereas the gradual expansion scheme seeks the cost optimization. The capability of GNN is demonstrated through solving 15 instances in practical size systems, where GNN can find far better solutions than the existing algorithm.