Combined Back-propagation Neural Network Research Articles

A back propagation neural network-based approach for inverting layered seabed acoustic parameters in shallow waters

IntroductionExisting methods primarily focus on earth acoustic parameters inversion under specific layered structures. However, they face challenges with experimental data from unknown seabed stratification, hindering accurate parameter inversion.MethodsTo address this, a novel algorithm combines Back Propagation Neural Network (BPNN) for distinguishing seabed stratification and inverting acoustic parameters. Simulated sound pressure data disturb seabed parameters as input, enabling feature recognition for training the neural network inversion model. Acoustic parameters are then estimated under identified stratification using the sound field model.ResultsThe inversion model is validated using simulation and pool shrinkage data. Results show the neural network model effectively stratifies simulation and experimental data, providing accurate inversion results for acoustic parameters corresponding to distinct layers.DiscussionThe neural network model's accuracy and practicality are confirmed through hierarchical judgment of scale test data and acoustic parameter inversion. This approach introduces a new perspective for shallow sea acoustic parameter inversion, offering a promising application scenario.

Short term load prediction of regional heating and heat storage system based on neural network

Accurate heat load prediction is the key to achieve fine control, energy conservation, and carbon reduction of regional hydronics. Taking the regional hydronics of a city in the north of China as the research object, the author, respectively uses back propagation neural network (BPNN), genetic algorithm (GA) optimized BPNN (GA-BPNN), and autoregressive integrated moving average model (ARI?MA) combined BPNN (ARIMA BPNN) to predict its heat load, and compares the accuracy and applicability of each prediction method. The results indicate that GA-BPNN has the smallest prediction error, followed by ARIMA-BPNN, but the latter requires less data for prediction. In practical engineering, if there is a sufficient amount of data related to heat load, it is recommended to use GA-BPNN. If there is a small amount of data, ARIMA-BP prediction method can be used.

Research on prediction of slurry migration distance in aggregate stacking based on GA-PSO-BPNN algorithm

After the water inrush accident in coal mine tunnels, early-stage pouring of aggregate forms a high-resistance, low-permeability aggregate stacking, transforming the pipeline flow into percolation. In the later stage, grouting is carried out into the interior of the aggregate stacking, effectively accumulating and solidifying the cement slurry. Among these, whether the slurry can migrate over long distances and fill the voids inside the aggregate stacking is the critical determinant of the success or failure of sealing. To quantitatively analyze the migration distance of slurry inside the aggregate stacking after grouting, a single-hole grouting test platform was established, and an orthogonal experiment was designed with grouting pressure, water cement ratio, and aggregate stacking porosity as influencing factors. Based on 25 sets of experimental measurements, four neural network prediction models suitable for studying the slurry migration distance within the aggregate stacking were constructed separately as back propagation neural network (BPNN), genetic algorithm (GA) combined BPNN, particle swarm optimization (PSO) combined BPNN, and GA-PSO combined BPNN. Evaluation criteria such as Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), Mean Square Error (MSE), Root Mean Square Error (RMSE), and the coefficient of determination (R2) were used for comparative analysis of the calculation errors and prediction accuracy of each model. From the perspective of neural network prediction results, the weight value of each influencing factor was analyzed, and the ranking was as follows: grouting pressure > aggregate particle size > water cement ratio, with grouting pressure being the primary controlling factor. The study demonstrates that the GA-PSO-BP model exhibits the best prediction performance, with an average relative error of only 1.59% and an R² of 0.998. This neural network model overcomes issues such as slow learning and getting stuck in tricky spots in BP neural networks. The prediction model shows high accuracy and stability, enabling more effective and accurate prediction of slurry migration distances, making it worthy of dissemination and application. This study can improve safety measures by reducing waste, expediting disaster management efforts, and minimizing environmental hazards associated with mining incidents.

Innovative modeling strategy of wind resistance for platoon vehicles based on real-time disturbance observation and parameter identification

Variable inter-vehicle distances influence significantly the wind resistance of platoon vehicles due to the sheltering of airflow. Accurate air drag estimation is extremely important for platoons in scenarios like energy-oriented driving and high-precision tracking. Most aerodynamic researchers have performed qualitative analysis of wind resistance for equally inter-spaced platoons, while the quantitative description of wind resistance variation with coupled inter-vehicle distances is rare. In addition, data measured through offline wind tunnel experiments, Computational Fluid Dynamics (CFD) simulations, or road tests via fuel consumption calibration for a period of time is unsynchronized, which may be unconvincing for real air drag estimation on road. Aiming at the quick and accurate approximation of platoon wind resistance, this paper proposes a novel and universal modeling strategy combining offline CFD simulation, online air drag observation, and real-time parameter identification. The variation characteristics of air drag with distance of a longitudinal platoon consisting of three homogeneous C-class Notchback cars are analyzed by CFD simulation. With appropriate data processing, a well-designed basis function is summarized. Then a novel wind resistance separation method combining Back-Propagation Neural Network (BPNN) and Extended State Observer (ESO) is proposed. Using the observed data stored in experience memory, the hybrid optimization method via particle swarm optimization (PSO) and gradient descent with momentum (GDM) is employed to identify the model parameters toward high accuracy and global optimality. Results of Hardware-In-the-Loop (HIL) experiment show that the proposed modeling strategy realizes effective real-time observation and accuracy description; the developed approximation model can describe the platoon wind resistance with continuous and coupled inter-vehicle distances, with the RMSE less than 11.4%.

Amalgamation of Geometry Structure, Meteorological and Thermophysical Parameters for Intelligent Prediction of Temperature Fields in 3D Scenes.

Temperature field calculation is an important step in infrared image simulation. However, the existing solutions, such as heat conduction modelling and pre-generated lookup tables based on temperature calculation tools, are difficult to meet the requirements of high-performance simulation of infrared images based on three-dimensional scenes under multi-environmental conditions in terms of accuracy, timeliness, and flexibility. In recent years, machine learning-based temperature field prediction methods have been proposed, but these methods only consider the influence of meteorological parameters on the temperature value, while not considering the geometric structure and the thermophysical parameters of the object, which results in the low accuracy. In this paper, a multivariate temperature field prediction network based on heterogeneous data (MTPHNet) is proposed. The network fuses geometry structure, meteorological, and thermophysical parameters to predict temperature. First, a Point Cloud Feature Extraction Module and Environmental Data Mapping Module are used to extract geometric information, thermophysical, and meteorological features. The extracted features are fused by the Data Fusion Module for temperature field prediction. Experiment results show that MTPHNet significantly improves the prediction accuracy of the temperature field. Compared with the v-Support Vector Regression and the combined back-propagation neural network, the mean absolute error and root mean square error of MTPHNet are reduced by at least 23.4% and 27.7%, respectively, while the R-square is increased by at least 5.85%. MTPHNet also achieves good results in multi-target and complex target temperature field prediction tasks. These results validate the effectiveness of the proposed method.

Deep learning for assessment of environmental satisfaction using BIM big data in energy efficient building digital twins

Energy efficient Building Digital Twins (BDTs) are researched using Building Information Model (BIM) to explore the key techniques of Digital Twins (DTs). DTs in buildings can be regarded as an expression of “BIM+,” born to digital descriptions. Comprehensive perception of physical systems is the preconditions for DTs implementation. BIM’s energy-saving design includes the selection of building orientation and building shape. BIM energy consumption analysis can compare different materials, examine the performance of various materials, and select the most suitable and most energy-efficient materials for building structure maintenance. Data Fusion Algorithm (DFA) in Wireless Sensor Networks (WSNs) is improved. A novel DFA is constructed by combining Backpropagation Neural Network (BPNN) with Dynamic Host Configuration Protocol (DCHP), recorded as BP-DCHP. Simulation experiment proves that BP-DCHP can prolong sensor nodes’ survival time and provide the highest data fusion quality. BP-DCHP runs for about 310 s, 500 s, and 705 s in WSNs consisting of 20, 50, and 100 WSNs, respectively. Moreover, BP-DCHP can provide higher quality given insufficient data fusion degree. Once the WSNs consume 50% of the total initial energy, BP-DCHP presents a shorter network delay, only 0.6 s on average in the 100-sensor-node-WSN. To validate BDTs’ effectiveness, the environmental satisfaction of residents from two Beijing intelligent communities is assessed using Deep Learning (DL) approach. Taking the data as the clue, the study establishes DTs serving the application of urban scene, which plays a certain role in promoting the technological innovation of BDTs, better optimizing the city and managing the city.

Cutting Force and Surface Roughness Optimizations in End Milling of GFRP Composites Utilizing BPNN-Firefly Method

The excessive cutting force that is generated in the end milling process of glass fiber-reinforced polymer (GFRP) composites can lower the surface quality. Hence, it is necessary to select the correct levels of end milling parameters to minimize the cutting force (CF) and surface roughness (SR). The parameters of the end milling process comprised the depth of cut (doc), spindle speed (n), and feeding speed (Vf). This study emphasized on the modeling and minimization of both CF and SR in the end milling of GFRP combo fabric by combining backpropagation neural network (BPNN) method and firefly algorithm (FA). The FA based BPNN was first performed to model the end-milling process and predict CF and SR. It was later also executed to obtain the best combination of end-milling parameter levels that would provide minimum CF and SR. The outcome of the confirmation experiments disclosed that the integration of BPNN and FA managed to accurately predict and substantially enhance the multi-objective characteristics.

Multiparameter prediction control of in vitro drug delivery into mycobacterium smegmatis induced by microbubble-enhanced sonoporation

The antibacterial method induced by microbubble-enhanced sonoporation has shown its great potential in facilitating drug delivery into thallus. The enhanced drug delivery induced by microbubble-enhanced sonoporation is a complex event which can be affected by various physical parameters. How to determine the correlation between experimental parameters and the drug delivery efficiency to give the instruction on reasonably choosing the parameters and achieve the control of drug delivery efficiency is impeding further investigations for this complex biophysical process. In the present work, we have explored a number of key parameters affecting the drug delivery efficiency induced by microbubble-enhanced sonoporation using multivariate biological experiments. To achieve the control of the drug delivery efficiency, a multiparameter prediction control method based on modified artificial neural network is presented in this paper. This method is a new modeling method based on combined back-propagation neural network and the multiple model idea to establish quantitative relationship between experimental parameters and drug delivery efficiency. By analyzing the experimental samples, a mapping relationship expression can be deduced to determine the input and output variables of artificial neural network models. Experimental samples were divided into training and test samples. We trained models based on back-propagation neural network to establish their quantitative relationship. In this model, the multiple model idea was introduced into the selection of training samples to modify the traditional back-propagation neural network model to avoid model mismatch caused by poor training sample selection. Numerical experiments results have shown that compared with the traditional model, the identification results obtained by modified model are more closed to experimental results. It is elucidated that an appropriately trained network can act as a good alternative for costly and time-consuming experiments. The findings of this study indicate that this approach can realize the prediction of drug delivery efficiency induced by microbubble-enhanced sonoporation under different experimental parameters, and then achieve the control of drug delivery efficiency through reasonable parameter selection, finally achieve the purpose of efficiently killing bacteria.

Lung Cancer Detection from X-ray images by combined Backpropagation Neural Network and PCA

The lungs are portion of a complex unit, enlarging and relaxing numerus times every day to supply oxygen and exude CO2. Lung disease might occur from troubles in any part of it. Carcinoma often called Cancer is the generally rising and it is the most harmful disease happened in humankind. Carcinoma occurs because of uncontrolled growth of malignant cells inside the tissues of the lungs. Earlier diagnosis of cancer can help save large numbers of lives, while any delay or fail in detection may cause additional serious problems leading to sudden fatal death. The objective of this study is to design an automated system with an ability to improve the detection process in order to perform advanced recognition of the disease. The diagnosis techniques include: X-rays, MRI, CT images etc. X-ray is the common and low-cost technique that is widely used and it is relatively available for everyone. Rather than new techniques like CT and MRI, X-ray is human dependable, meaning it needs a Doctor and X-ray specialist in order to determine lung cases, so developing a system which can enhance and aid in diagnosis, can help specialist to determine cases in easily.

Regional Short-term Micro-climate Air Temperature Prediction with CBPNN

This paper proposes a novel short-term air temperature prediction with three-layer Back Propagation Neural Network (BPNN) for the regional application of next 1-12 hours. With the continuous collection of eight real-time micro-climate parameters in the experimentation and demonstration stations in our university, the Multiple Stepwise Regression (MSR) is employed to screen the original historical data to find the parameter factors with greater contribution rate. On the basis of the Root Mean Square Error (RMSE) value evaluating the optimal fitting degree of the stepwise regression, the Levenberg-Marquardt (LM) and the Resilient Propagation (R-Prop) training algorithm are employed to construct a Combined BPNN (CBPNN) with two MSR inputs. Compared with the known micro-climate data sets, the Mean Absolute Error (MAE) is to evaluate the applicability of CBPNN prediction model. The experimentation shows that the MAE is within 4°C in the next 12 hours. This proposal will be deployed in stations in our university for extreme weather warnings, and could be applied to some regional short-term parameter prediction for the future agricultural production service.

Warpage optimization of fiber-reinforced composite injection molding by combining back propagation neural network and genetic algorithm

Warpage is an important index in measuring part quality in the short fiber-reinforced composite injection molding process. In addition to process parameters, fiber parameters have an important influence on part warpage. In this study, the objective function is a minimum problem of the warpage. The design parameters include fiber aspect ratio, fiber content, injection time, melt temperature, mold temperature, and holding pressure. The combined back propagation neural network (BPNN) and genetic algorithm (GA) approach is proposed to optimize the fiber-reinforced composite injection molding process. On the basis of orthogonal experiment design, Moldflow software is applied in the short fiber-reinforced composite injection molding process. The importance of various parameters on the influence of warpage is researched by the range analysis method. Thereafter, a back propagation neural network model is developed on the basis of the simulation results to map the complex nonlinear relationship between design parameters and warpage. The GA is interfaced with this predictive model to improve the warpage significantly by optimizing the design parameters. A case study of a part is presented. The optimum values of design parameters are determined to minimize the warpage. Results show that the combined BPNN/GA approach is an effective method for the warpage optimization of fiber-reinforced composite injection molding.

Application of Improved NSGA-II to Multi-objective Optimization of a Coal-fired Boiler Combustion Electronical Systems in Green Food Bases

In this study, we have a research of a hybrid algorithm by combining BP neural network and improved Non-dominated Sorting Genetic Algorithm-II (NSGA-II) to solve the multi-objective optimization problem of a nanoscale coal-fired boiler combustion electronical systems, the two objectives considered are minimization of overall heat loss and NOx emissions from coal-fired boiler. First, Back Propagation (BP) neural network was dopted to establish a mathematical model predicting the NOx emissions and overall heat loss of the coal-fired boiler with the inputs such as operational parameters of the nanoscale coal-fired boiler. Then, BP model and the Non-dominated Sorting Genetic Algorithm II (NSGA-II) were combined to gain the optimal operating parameters which lead to lower NOx emissions and overall heat loss boiler. According to the problems such as premature convergence and uneven distribution of Pareto solutions exist in the application of NSGA-II, corresponding improvements in the crowded-comparison operator and crossover operator were performed. The optimization results show that hybrid algorithm by combining BP neural network and improved NSGA-II can be a good tool to solve the problem of multi-objective optimization of a nanoscale coal-fired boiler combustion electronical systems in green food bases, which can reduce NOx emissions and overall heat loss effectively for the nanoscale coal-fired boiler combustion electronical systems in green food bases. Compared with original NSGA-II, the Pareto set obtained by the improved NSGA-II shows a better distribution and better quality.

Multibiometric Personal Identification based on Hybrid Artificial Intelligence Technique using Serial Mode Architecture

this research a multibiometric identification system with serial mode using palmprint, Dental and DNA biometric traits is proposed to identify person. A proposed system consists of three steps, first step is to create a database where all the information about the authenticated users is stored. Second step is to extracts the features from a database images for palmprint and dental estimated the eigenvalues using KL transforming while the DNA data converted into ASCII code. The last step used hybrid Artificial technique combined backpropagation Neural Network and particle swarm optimization (PSO) as classifier and compare the test image with the stored information in the database. The experimental results applied cascaded (serial) mode show that the proposed system has an encouraging performance. The Genuine acceptance rate 98%, False Acceptance Rate (2%), False Rejected Rate (5%) and Recognition Rate (93%) KeywordsTerms -Multibiometric - serial mode-

Design of Control System of Eddy Current Retarder Based on BP Neural Network PID Controller

A new kind of algorithm of controller for eddy current retarder is designed in this paper. The eddy current retarder control system with traditional PID controller can't achieve a perfect performance in the rapid response. Back propagation (BP) neural network is one of artificial neural networks which has a good learning ability with a simple and recurrent structure, so it is suitable for controlling complicated eddy current retarder system. This paper introduces the principle, characteristics and learning algorithm of the BP neural network and designs the control system of eddy current retarder based on BP neural network PID controller by combining BP neural network and traditional PID. Making use of MATLAB, simulate this new kind of controller for eddy current retarder in the rapid response. Simulation results show it can improve the dynamic response performance and enhance the static precision compared to the traditional PID controller.

Multi-objective Optimization of Coal-fired Boiler Combustion Based on NSGA-II

NOx emission characteristics and overall heat loss model for a 300MW coal-fired boiler were established by Back Propagation (BP) neural network, by which the the functional relationship between outputs (NOx emissions & overall heat loss of the boiler) and inputs (operational parameters of the boiler) of a coal-fired boiler can be predicted. A number of field test data from a full-scale operating 300MWe boiler were used to train and verify the BP model. The NOx emissions & heat loss predicted by the BP neural network model showed good agreement with the measured. Then, BP model and the non-dominated sorting genetic algorithm II (NSGA-II) were combined to gain the optimal operating parameters which lead to lower NOx emissions and overall heat loss boiler. The optimization results showed that hybrid algorithm by combining BP neural network with NSGA-II can be a good tool to solve the problem of multi-objective optimization of a coal-fired combustion, which can reduce NOx emissions and overall heat loss effectively for the coal-fired boiler.

Optimal Design of Pre-Forging for Gear Blank Using BP Neural Network and Genetic Algorithm

The design of pre-forging is very important during multistage forging of producing gear blank. It directly affects the behavior of metal flowing, filled situation of die cavity of finish forging, quality of products and die life. Most designs of pre-forging for gear blank are based on trial and error method. This paper presents a suitable method for practical designation of pre-forging for gear blank by proposing an improved algorithm, which combines Back Propagation Neural Network and Genetic Algorithm. Firstly, the mathematical model between the size parameters of pre-forging and forming force and maximum die stress of finish forging was established by using Back Propagation Neural Network which has the feature of processing highly nonlinear problems. Secondly, the established model was set as the fitness function of Genetic Algorithm. At last, the most superior pre-forging shape and the size parameters were solved by using the Genetic Algorithm with the function of overall situation optimization. These can lead to lower cost and time in the stages of designing pre-forging for gear blank.

Wavelet-coupled backpropagation neural network as a chamber leak detector of plasma processing equipment

In order to improve equipment throughput and device yield, chamber leaks needs to be strictly monitored. A new technique for leak detection is presented and this was accomplished by combining backpropagation neural network, discrete wavelet transformation (DWT), and continuous transformation (CWT). Different types of BPNN models were constructed with raw, DWT, and CWT data and these are referred to as raw, DWT, and CWT models, respectively. Constructed models were validated with a total of 47 data sets for normal and leaky chamber conditions. The experimental data were in-situ collected by using an optical emission spectroscopy. Both raw and DWT models could detect all abnormal data sets. Worst detection by CWT model was noted. Wider detection margin provided by DWT model was attributed to enhanced sensitivity of model to leaky condition. A modified cumulative control chart was applied to the statistical mean of raw OES spectra as well as to DWT and CWT data. The statistical mean-based CUSUM control chart was unable to detect chamber leaks. In contrast, chamber leaks could be identified by all model-based CUSUM control charts. Of the proposed models, DWT model is identified to be the most appropriate to chamber leak detection.

Geometric and force errors compensation in a 3-axis CNC milling machine

This paper proposes a new off line error compensation model by taking into accounting of geometric and cutting force induced errors in a 3-axis CNC milling machine. Geometric error of a 3-axis milling machine composes of 21 components, which can be measured by laser interferometer within the working volume. Geometric error estimation determined by back-propagation neural network is proposed and used separately in the geometric error compensation model. Likewise, cutting force induced error estimation by back-propagation neural network determined based on a flat end mill behavior observation is proposed and used separately in the cutting force induced error compensation model. Various experiments over a wide range of cutting conditions are carried out to investigate cutting force and machine error relation. Finally, the combination of geometric and cutting force induced errors is modeled by the combined back-propagation neural network. This unique model is used to compensate both geometric and cutting force induced errors simultaneously by a single model. Experimental tests have been carried out in order to validate the performance of geometric and cutting force induced errors compensation model.