A fully autonomous kernel-based online learning neural network model and its application to building cooling load prediction

Building cooling load prediction is critical to the success of energy-saving measures. While many of the computational models currently available in the industry have been developed for this purpose, most require extensive computer resources and involve lengthy computational processes. Artificial neural networks (ANNs) have recently been adopted for prediction, and pioneering works have confirmed the feasibility of this approach. However, users are required to predetermine an ANN model’s parameters. This hinders the applicability of the ANN approach in actual engineering problems, as most engineers may be unfamiliar with soft computing. This paper proposes a fully autonomous kernel-based neural network (AKNN) model for noisy data regression prediction. No part of the model’s mechanism requires human intervention; rather, it self-organises its structure according to the training samples presented. Unlike the other existing autonomous models, the AKNN model is an online learning model. It is particularly suitable for online steps-ahead prediction. In this paper, we benchmark the AKNN model’s performance according to other ANN models. It is also successfully applied to predicting the cooling load of a commercial building in Hong Kong. The occupancy areas and concentration of carbon dioxide inside the building are successfully adopted to mimic the building’s internal cooling load. Training data was adopted from actual measurements taken inside the building. Its results show reasonable agreement with actual cooling loads.     

Modeling users through an expert system and a neural network

With the number of Internet and Web users increasing rapidly, electronic service providers are competing to satisfy and better serve customers looking for information or channels of advertisement. A wide variety of browses, specialized sites, custom made software, etc. are being offered on a regular basis. However, the user has to filter through a large number of files before finding what he/she is really looking for. This paper presents a user modeling expert system, SIGMA, based on neural networks for encapsulating Internet and Web users' habits and preferences. SIGMA is an artificial intelligence application designed to answer an Internet client needs and preferences. It analyses the user supplied demographic data and the monitored transactions then generate a tailored profile that is ultimately used to filter what information is being passed on to him/her in an effort to reduce and hopefully eliminate the time and energy expended in sifting through raw and often unwanted data.     

Graph-based algorithmic design and decision-making framework for district heating and cooling plant positioning and network planning

Although buildings are the single largest users of energy in cities, many individual building HVAC (heating, ventilation, and air conditioning) systems are energy inefficient. Sustainable district heating and cooling systems have been developed to address this inefficiency, however, the implementation of district heating and cooling systems is a complex, capital intensive multivariate problem. One critical engineering problem, the location of energy production plant and pipe network topology, has a major influence on the system performance as well as the life cycle cost, and therefore should be comprehensively studied and chosen. This paper proposes a novel algorithmic design and decision-making framework that uses multiple simultaneous criteria to generate and evaluate design alternatives of site selection and pipe network layout. The framework consists of three components: site location and topology generation, network sizing and evaluation, and multi-objective decision-making assistance. The proposed framework has been validated in a real-world district cooling greenfield project in China's metropolitan area. Results were compared against the engineers’ best practices. It has shown that the proposed framework could find equally good or better designs in engineering and financial performance with CAPEX reduction of up to 15%.     

Vibration control of vehicle active suspension system using a new robust neural network control system

The main problem of vehicle vibration comes from road roughness. For that reason, it is necessary to control vibration of vehicle’s suspension by using a robust artificial neural network control system scheme. Neural network based robust control system is designed to control vibration of vehicle’s suspensions for full suspension system. Moreover, the full vehicle system has seven degrees of freedom on the vertical direction of vehicle’s chassis, on the angular variation around X-axis and on the angular variation around Y-axis. The proposed control system is consisted of a robust controller, a neural controller, a model neural network of vehicle’s suspension system. On the other hand, standard PID controller is also used to control whole vehicle’s suspension system for comparison.Consequently, random road roughnesses are used as disturbance of control system. The simulation results are indicated that the proposed control system has superior performance at adapting random road disturbance for vehicle’s suspension.     

Operations and learning in neural networks for robust prediction

We consider stochastic neural networks, the objective of which is robust prediction for spatial control. We develop neural structures and operations, in which the representations of the environment are preprocessed and provided in quantized format to the prediction layer, and in which the response of each neuron is binary. We also identify the pertinent stochastic network parameters, and subsequently develop a supervised learning algorithm for them. The on-line learning algorithm is based an the Kullback-Leibler performance criterion, it induces backpropagation, and guarantees fast convergence to the prediction probabilities induced by the environment, with probability one.     

Level set topology optimization of cooling and heating devices using a simplified convection model

This paper studies topology optimization of convective heat transfer problems in two and three dimensions. The convective fluxes are approximated by Newton’s Law of Cooling (NLC). The geometry is described by a Level Set Method (LSM) and the temperature field is predicted by the eXtended Finite Element Method (XFEM). A constraint on the spatial gradient of the level set field is introduced to penalize small, sub-element-size geometric features. Numerical studies show that the LSM-XFEM provides improved accuracy over previously studied density methods and LSMs using Ersatz material models. It is shown that the NLC model with an iso-thermal fluid phase may over predict the convective heat flux and thus promote the formation of very thin fluid channels, depending on the Biot number characterizing the heat transfer problem. Approximating the temperature field in the fluid phase by a diffusive model mitigates this issue but an explicit feature size control is still necessary to prevent the formation of small solid members, in particular at low Biot numbers. The proposed constraint on the gradient of the level set field is shown to suppress sub-element-size features but necessitates a continuation strategy to prevent the optimization process from stagnating as geometric features merge.     

Combining a neural network with a rule-based expert system approach for short-term power load forecasting in Taiwan

A back-propagation neural network with the output provided by a rule-based expert system is designed for short-term power load forecasting. To demonstrate that the inclusion of the prediction from a rule-based expert system of a power system would improve the predictive capability of the network, load forecasting is performed on the Taiwan power system. The hourly load for one typical day was evaluated and, in that case, the inclusion of the rule-based expert system prediction significantly improved the neural network's prediction of power load. Moreover, the proposed combined approach converges much faster than the conventional neural network and the rule-based expert system method. Extensive studies were performed on the robustness of the built network model by using different specified censoring time. The prediction intervals of future power load series are also provided, to evaluate the prediction efficiency of the neural network model.     

Particle filter robot localisation through robust fusion of laser,WiFi, compass,and a network of external cameras

In this paper, we propose a multi-sensor fusion algorithm based on particle filters for mobile robot localisation in crowded environments. Our system is able to fuse the information provided by sensors placed on-board, and sensors external to the robot (off-board). We also propose a methodology for fast system deployment, map construction, and sensor calibration with a limited number of training samples. We validated our proposal experimentally with a laser range-finder, a WiFi card, a magnetic compass, and an external multi-camera network. We have carried out experiments that validate our deployment and calibration methodology. Moreover, we performed localisation experiments in controlled situations and real robot operation in social events. We obtained the best results from the fusion of all the sensors available: the precision and stability was sufficient for mobile robot localisation. No single sensor is reliable in every situation, but nevertheless our algorithm works with any subset of sensors: if a sensor is not available, the performance just degrades gracefully.     

Time series prediction using PSO-optimized neural network and hybrid feature selection algorithm for IEEE load data

Artificial neural networks have been widely used in time series prediction. In this paper, multi-layer feedforward neural networks with optimized structures, using particle swarm optimization (PSO) algorithm, are used for hourly load prediction based on load time series of IEEE Reliability Test System. To have a small and appropriate feature subset, a hybrid method is used for feature selection in this paper. This hybrid method uses the combination of genetic algorithm (GA) and ant colony optimization (ACO) algorithm. The season, day of the week, time of the day and history load are considered as load influencing factors in this study based on the mentioned standard load dataset. The optimized number of neurons in the hidden layers of multi-layer perceptron (MLP) is determined using PSO algorithm. Experimental results show that the proposed hybrid model offers superior performance, in terms of mean absolute percentage error (MAPE), in time series prediction as compared to some recent researches in this field.     

A hybrid evolutionary dynamic neural network for stock market trend analysis and prediction using unscented Kalman filter

Stock market prediction is of great interest to stock traders and investors due to high profit in trading the stocks. A successful stock buying/selling generally occurs near price trend turning point. Thus the prediction of stock market indices and its analysis are important to ascertain whether the next day's closing price would increase or decrease. This paper, therefore, presents a simple IIR filter based dynamic neural network (DNN) and an innovative optimized adaptive unscented Kalman filter for forecasting stock price indices of four different Indian stocks, namely the Bombay stock exchange (BSE), the IBM stock market, RIL stock market, and Oracle stock market. The weights of the dynamic neural information system are adjusted by four different learning strategies that include gradient calculation, unscented Kalman filter (UKF), differential evolution (DE), and a hybrid technique (DEUKF) by alternately executing the DE and UKF for a few generations. To improve the performance of both the UKF and DE algorithms, adaptation of certain parameters in both these algorithms has been presented in this paper. After predicting the stock price indices one day to one week ahead time horizon, the stock market trend has been analyzed using several important technical indicators like the moving average (MA), stochastic oscillators like K and D parameters, WMS%R (William indicator), etc. Extensive computer simulations are carried out with the four learning strategies for prediction of stock indices and the up or down trends of the indices. From the results it is observed that significant accuracy is achieved using the hybrid DEUKF algorithm in comparison to others that include only DE, UKF, and gradient descent technique in chronological order. Comparisons with some of the widely used neural networks (NNs) are also presented in the paper.     

Gaussian filters and filter synthesis using a Hermite/Laguerre neural network

A neural network for calculating the correlation of a signal with a Gaussian function is described. The network behaves as a Gaussian filter and has two outputs: the first approximates the noisy signal and the second represents the filtered signal. The filtered output provides improvement by a factor often in the signal-to-noise ratio. A higher order Gaussian filter was synthesized by combining several Hermite functions together.     

Design and implementation of an autonomous EGR cooling system using deep neural network prediction to reduce NOx emission and fuel consumption of diesel engine

Neural Computing and Applications - This study includes the design of an autonomous exhaust gas recirculation (EGR) cooling system and implementation of the system on diesel engine by using deep...     

Short-term Maharashtra state electrical power load prediction with special emphasis on seasonal changes using a novel focused time lagged recurrent neural network based on time delay neural network model

In this paper, the parameter-wise optimization training process is implemented to achieve an optimal configuration of focused time lagged recurrent neural network (FTLRNN) models by embedding the gamma, laguarre, and multi-channel tapped delay line memory structure. The aim is to examine the prediction ability of the proposed models in order to predict one-day-ahead electric power load simultaneously as usual to oppose 1–24 h forecast in sequel with a special emphasis on seasonal changes over a year. An improved delta-bar-delta algorithm is used to accelerate the training of neural networks and to improve the stability of the convergence.Experimental results indicate that the FTLRNN with time delay neural network (TDNN) clearly outperformed the gamma and laguarre based short-term memory structure in various performance metrics such as mean square error (MSE), normalized MSE, correlation coefficient (r) and mean absolute percentage error (MAPE) during evaluation process. Empirical results show that the proposed dynamic NN model consistently performs well on daily, weekly, and monthly average basis in terms of prediction accuracy. It is noticed from the literature review that an optimally configured FTLRNN with multi-channel tapped delay line memory structure is not currently available to solve short-term electrical power load prediction. The proposed method gives acceptable errors in all seasons, months and on daily basis. The average prediction error on three weeks is obtained as low as 1.67%.     

灰色神经网络在股价预测中的应用研究

股票价格受多种因素的综合影响,具有趋势性、较大波动性和随机性等变化特点,单一模型难准确对其变化规律进行准确描述,将灰色理论和BP神经网络相结合构建一种股票价格组合预测模型。采用灰色GM（1,1）预测模型动态预测股票价格变化趋势,运用BP神经网络对灰色GM（1,1）模型预测结果进行修正,以提高股票价格预测精度。采用ST东北高（600003）股票价格对预测模型性能进行测试,结果表明,组合预测模型提高了股票价格的预测精度,更能挖掘股票价格变化规律。     

Intelligent task planning and action selection of a mobile robot in a multi-agent system through a fuzzy neural network approach

This paper proposes an intelligent task planning and action selection mechanism for a mobile robot in a robot soccer system through a fuzzy neural network approach. The proposed fuzzy neural network system is developed through the two dimensional fuzzification of the soccer field. A five layer fuzzy neural network system is trained through error back propagation learning algorithm to impart a strategy based action selection. The action selection depends on the field configuration, and the emergence of a particular field configuration results from the game dynamics. Strategy of the robot changes when the configuration of the objects in the field changes. The proposed fuzzy neural network structure is flexible to accommodate all possible filed configurations. Simulation results indicate that the proposed approach is simple and has the capability in coordinating the multi-agent system through selection of sensible actions.     

Artificial neural network and fuzzy expert system comparison for prediction of performance and emission parameters on a gasoline engine

This study is deals with artificial neural network (ANN) and fuzzy expert system (FES) modelling of a gasoline engine to predict engine power, torque, specific fuel consumption and hydrocarbon emission. In this study, experimental data, which were obtained from experimental studies in a laboratory environment, have been used. Using some of the experimental data for training and testing an ANN for the engine was developed. Also the FES has been developed and realized. In this systems output parameters power, torque, specific fuel consumption and hydrocarbon emission have been determined using input parameters intake valve opening advance and engine speed. When experimental data and results obtained from ANN and FES were compared by t-test in SPSS and regression analysis in Matlab, it was determined that both groups of data are consistent with each other for p > 0.05 confidence interval and differences were statistically not significant. As a result, it has been shown that developed ANN and FES can be used reliably in automotive industry and engineering instead of experimental work.     

神经网络在混凝土碳化深度预测中的研究应用

针对BP神经网络搜索速度慢、易陷入局部极值的缺陷,采用PSO算法优化BP神经网络后建立各影响因素与部分碳化区长度的关系模型。将改进后的模型进行实验仿真训练并应用于某混凝土大桥部分位置的碳化深度预测中,仿真应用结果表明,网络输出值和期望值很好吻合,收敛速度更快。所以该模型能够对混凝土部分碳化区长度进行预测,为混凝土结构耐久性设计、评估和寿命预测提供科学指导。     

Artificial neural network for prediction of air flow in a single rock joint

In this paper, an attempt has been made to evaluate and predict the air flow rate in triaxial conditions at various confining pressures incorporating cell pressure, air inlet pressure, and air outlet pressure using artificial neural network (ANN) technique. A three-layer feed forward back propagation neural network having 3-7-1 architecture network was trained using 37 data sets measured from laboratory investigation. Ten new data sets were used for the, validation and comparison of the air flow rate by ANN and multi-variate regression analysis (MVRA) to develop more confidence on the proposed method. Results were compared based on coefficient of determination (CoD) and mean absolute error (MAE) between measured and predicted values of air flow rate. It was found that CoD between measured and predicted air flow rate was 0.995 and 0.758 by ANN and MVRA, respectively, whereas MAE was 0.0413 and 0.1876.     

Flank wear prediction in drilling using back propagation neural network and radial basis function network

In the present work, two different types of artificial neural network (ANN) architectures viz. back propagation neural network (BPNN) and radial basis function network (RBFN) have been used in an attempt to predict flank wear in drills. Flank wear in drill depends upon speed, feed rate, drill diameter and hence these parameters along with other derived parameters such as thrust force, torque and vibration have been used to predict flank wear using ANN. Effect of using increasing number of sensors in the efficacy of predicting drill wear by using ANN has been studied. It has been observed that inclusion of vibration signal along with thrust force and torque leads to better prediction of drill wear. The results obtained from the two different ANN architectures have been compared and some useful conclusions have been made.