Back Propagation Neural Network Prediction Research Articles

A Method for Optimizing Different Geometric Shields of D-T Neutron Generators by Combining BP Neural Network and Analytic Hierarchy Process

In this paper, an optimization of shielding structures with different geometries is established for the D-T neutron generator system by combining Back Propagation (BP) neural network and Analytic Hierarchy Process (AHP). The D-T neutron generator (Model NG-9) used in the system was developed independently by Northeast Normal University. After investigating the rule of shielding performance among spherical, cylindrical and cubic geometries, the spherical shield is selected for BP neural network prediction to determine the total dose rate through it. Information about spherical multilayer-shielding structures and properties calculated by MCNP code is used to train the neural network. The predicted result serves as a parameter of the evaluation function, which provides a comprehensive assessment of the dose rate penetrated the shield, the shielding mass, and the shielding volume. Together with AHP, the weight factors are determined for all the optimization objectives to construct the evaluation function. By comparing its values, the optimal shielding structures for spherical, cylindrical and cubic materials are found. Against MCNP simulated values, the total dose rates’ errors of the optimal shielding structures for the sphere, cylinder, and cube are 1.72%, -4.94%, and -5.17%, respectively. This result demonstrates that the combination of BP neural network and AHP is more effective in addressing multi-objective optimization problems related to the design of radiation shielding for various geometries.

Characterization of natural fracture development in coal reservoirs using logging machine learning inversion, well test data and simulated geostress analyses

Natural fractures directly affect the permeability and mechanical strength of reservoirs, and their development degree has an important impact on the design and implementation of engineering and development projects. Although there is some correlation between logging data and fracture development, studies using algorithms to optimize logging predictions are still scarce. Meanwhile, there is a scarcity of calculations and analyses concerning the distribution of geostress at the block scale, and the pivotal role that geostress plays as a tectonic factor in the development of fractures. In this study, machine learning methods are used to predict reservoir fracture development, and regional geostress distribution patterns derived from well test data and finite element methods are combined for verification. The support vector machine (SVM), random forest (RF), extreme gradient boosting (XGBoost) and back propagation neural network (BPNN) algorithms are used to predict the fracture development of No.15 coal reservoir in the block. The results showed that the accuracy of SVM is 83.3 %, RF is 91.6 %, XGBoost is 93.7 % and BPNN is 95.8 %. The BPNN can effectively predict the reservoir fracture development of the block. Combined with the regional finite element stress-strain analysis and geostress measurement, the prediction of No.15 coal geostress distribution and fracture development model is established. Under comprehensive verification, the established distribution of the degree of regional fracture development under the control of geostress is consistent with the results of the BPNN prediction of fracture development. These results show that the regional geostress calculated in association with finite element analysis (FEA) can reflect the development of fracture in coalbed methane (CBM) reservoirs, and the neural network has good performance in predicting regional fracture development. This work provides a new approach to the application of machine learning in the field of geological engineering, and the comprehensively validated model provides geologists and geological engineers with ideas in algorithmic practice.

Process Optimization for Antarctic Krill (Euphausia superba) Sauce Based on Back Propagation Neural Network Combined with Genetic Algorithm

Using Antarctic krill (Euphausia superba) as the research object, we optimized the process conditions for Antarctic krill sauce (AkS) by including three factors (salt addition, rock sugar addition, and the oil-to-material ratio) and sensory evaluation as response values. The data from the response surface were fed into the back propagation (BP) neural network training, generating a model mapping the process conditions and sensory scores, which were subsequently combined with the genetic algorithm (GA) for global optimization to determine the optimal process for AkS preparation. The results revealed that the response surface model was well suited to the BP neural network training and prediction sets, with correlation values of 0.98 and 0.95, respectively. The fitting prediction effect was obvious for the sensory scoring results of the product. The parameters obtained from the GA’s global optimization search accord with the analytical results of the response surface. The findings demonstrated that combining a BP neural network with a GA can enhance the AkS preparation technique. Under optimal processing conditions, AkS has a high sensory score and protein and carbohydrate contents, moderate fat content, minimal fat oxidation, and non-detectable pathogens, indicating that the AkS in this study was nutritious and safe to consume.

Research on the Development Height Prediction Model of Water-Conduction Fracture Zones under Conditions of Extremely Thin Coal Seam Mining

Addressing the difficult problem of predicting the height of water-conducting fracture zones in shallow and thin coal seams, a prediction model of water-conduction fracture zones based on a backpropagation (BP) neural network was developed by integrating theoretical analysis, field measurements, and algorithmic advancements. Firstly, through overburden migration analysis and correlation tests, the height index system of the water-conducting fracture zone was determined. This system includes mining height, buried depth, dip angle, working face width, and overburden rock lithology, with five groups of characteristic parameters. Then, 35 pairs of minefield-measured data were collected to establish the measured height data set of the water-conducting fracture zone. Secondly, a BP neural network prediction model and a traditional support vector regression (SVR) prediction model were constructed based on a Pytorch framework, and the models were trained and tested by selecting data sets. Thirdly, the optimal prediction model was determined by comparing the model with the empirical model and multiple regression model of mining regulations for coal pillar maintenance and pressure in buildings, water bodies, railways, and main shafts. Finally, a typical mine was selected for application to verify the suitability of the optimal model. The results show that: (1) the predicted value of the neural network model is consistent with the change trend of the measured value, which accords with the theoretical law; (2) compared with traditional forecasting methods, the error of the BP neural network prediction model is stable and the prediction effect is the best; (3) dropout can effectively mitigate mitigation training overfitting, achieve regularization, and improve prediction accuracy; (4) the field application further verified that the BP neural network model is the best for predicting the height of water-conducting fracture zones of extremely thin coal seams, and the research results can provide technical guidance for similar fragile coal seams.

Regularity analysis of resilient modulus for hot-mix asphalt with large temperature fluctuations

To evaluate the regularity of resilient modulus for hot-mix asphalt (HMA) under large temperature fluctuations, back propagation (BP) neural network technology was used to analyze the continuous change of HMA resilient modulus. Firstly, based on the abundant data, the training model of HMA resilient modulus was established by using BP neural network technology. Subsequently, BP neural network prediction and regression analysis were performed, and the prediction model of HMA resilient modulus at different temperatures (−50 °C to 60 °C) was obtained, which fully considered multi-factor and nonlinearity. Finally, the fitted theoretical model can be used to evaluate the HMA performance under the condition of large temperature fluctuations, and the rationality of theoretical model was verified by taking Harbin region as an example. It was found that the relationship between HMA resilient modulus and temperatures can be described by inverse tangent function. And the key parameters of theoretical model can be used to evaluate the continuous change characteristics of HMA resilient modulus with large temperature fluctuations. The results can further improve the HMA performance evaluation system and have certain theoretical value.

Performance prediction and optimization of lateral exhaust hood based on back propagation neural network and genetic algorithm

The lateral exhaust hood (LEH) is commonly used to capture contaminated airflow in industrial plants. Its performance is influenced by numerous factors and exhibits a strong nonlinear relationship, optimizing this performance demands significant computational time and cost. This study proposes an LEH optimization design method that combines a backpropagation neural network (BPNN) with a genetic algorithm (GA), verified through experiments and numerical simulations. A BPNN prediction model is established based on 520 CFD simulation results. This study discusses seven factors influencing LEH performance, including the buoyant jet's velocity and temperature, the LEH's geometry and position, and the exhaust velocity. The results indicate that within the parameter range of the prediction model, there is a critical aspect ratio value that significantly affects LEH capture efficiency. The horizontal distance between the LEH and the pollution source significantly affects capture efficiency, especially when the LEH's installation height is low (H/D < 0.5). In addition, when combined with GA for fast searching, the expected combination of LEH design parameters can be obtained. This study aims to enhance the design efficiency of industrial exhaust hoods and improve worker health protection.

Optimization Study of Steady-State Aerial-Towed Cable Circling Strategy Based on BP Neural Network Prediction

This paper presents models for UAV aerial-towed cables in free-end and fixed-end configurations, crucial for tasks like communication and aerial charging. By establishing a quasi steady-state model, computational results on cable shapes are obtained. To accelerate computations, a backpropagation (BP) neural network prediction model is trained, significantly reducing the computation time. An evaluation function has been developed that integrates both aircraft performance and cable shape considerations to evaluate circling parameters across various states. This function integrates techniques such as BP neural networks and particle swarm optimization (PSO) to refine parameters such as velocities and bank angles for both free-end and fixed-end cables. The results show that the BP neural network accurately predicts cable shapes, achieving a maximum error of 5% in towing force and verticality. Additionally, PSO efficiently optimizes circling parameters, thereby enhancing the effectiveness of the evaluation function in identifying optimal solutions. This approach significantly improves the efficiency of determining optimal circling parameters for UAV aerial-towed cables, thereby contributing to their operational efficacy.

Height Prediction of Water-Conducting Fracture Zone in Jurassic Coalfield of Ordos Basin Based on Improved Radial Movement Optimization Algorithm Back-Propagation Neural Network

The development height of the water-conducting fracture zone (WCFZ) is crucial for the safe production of coal mines. The back-propagation neural network (BP-NN) can be utilized to forecast the WCFZ height, aiding coal mines in water hazard prevention and control efforts. However, the stochastic generation of initial weights and thresholds in BP-NN usually leads to local optima, which might reduce the prediction accuracy. This study thus invokes the excellent global optimization capability of the Improved Radial Movement Optimization (IRMO) algorithm to optimize BP-NN. The influences of mining thickness, coal seam depth, working width, and hard rock lithology proportion coefficient on the height of WCFZ are investigated through 75 groups of in situ data of WCFZ heights measured in the Jurassic coalfield of the Ordos Basin. Consequently, an IRMO-BP-NN model for predicting WCFZ height in the Jurassic coalfield of the Ordos Basin was constructed. The proposed IRMO-BP-NN model was validated through monitoring data from the 4−2216 working faces of Jianbei Coal Mine, followed by a comparative analysis with empirical formulas and conventional BP-NN models. The relative error of the IRMO-BP-NN prediction model is 4.93%, outperforming both the BP-NN prediction model, the SVR prediction model, and empirical formulas. The results demonstrate that the IRMO-BP-NN model enhances the accuracy of predicting WCFZ height, providing an application foundation for predicting such heights in the Jurassic coalfield of the Ordos Basin and protecting the ecological environment of Ordos Basin mining areas.

Global Navigation Satellite System Receiver Positioning in Harsh Environments via Clock Bias Prediction by Empirical Mode Decomposition and Back Propagation Neural Network Method.

This paper proposes a novel method to improve the clock bias short-term prediction accuracy of navigation receivers then solve the problem of low positioning accuracy when the satellite signal quality deteriorates. Considering that the clock bias of a navigation receiver is equivalent to a virtual satellite, the predicted value of clock bias is used to assist navigation receivers in positioning. Consequently, a combined prediction method for navigation receiver clock bias based on Empirical Mode Decomposition (EMD) and Back Propagation Neural Network (BPNN) analysis theory is demonstrated. In view of systematic errors and random errors in the clock bias data from navigation receivers, the EMD method is used to decompose the clock bias data; then, the BPNN prediction method is used to establish a high-precision clock bias prediction model; finally, based on the clock bias prediction value, the three-dimensional positioning of the navigation receiver is realized by expanding the observation equation. The experimental results show that the proposed model is suitable for clock bias time series prediction and providing three-dimensional positioning information meets the requirements of navigation application in the harsh environment of only three satellites.

Prediction of the whole society electricity consumption in northeast China based on the BP neural network and Markov

Northeast China has been facing a severe power shortage situation. Since September 2021, “power rationing” events occurring in many places in the three provinces of northeast China have been causing inconvenience to people’s production and life. Therefore, it is particularly important to accurately predict the power load combined with the influencing factors of local power consumption. At the same time, the northeast region is about to enter the heating season, and the pressure on coal and electricity will further increase. In Heilongjiang Province, due to coal capacity control, limited production led to the high price of thermal coal; wind power photovoltaic output fluctuations, the epidemic, and other reasons also resulted in a large gap in the power supply side. Improving the power demand forecasting ability is of great significance to strengthen the reliability of people’s daily electricity consumption, rational distribution of power generation plans, and deployment of power grid resources. In order to improve the accuracy of electricity consumption prediction in Heilongjiang Province, Markov error correction is carried out on the basis of the backpropagation (BP) neural network prediction model so that the final prediction results have the advantages of the BP neural network prediction model and Markov model. In addition, it is more suitable for the prediction of random series data with high volatility, the prediction accuracy can be improved significantly, and the overall trend of electricity consumption can be predicted more accurately.

Prediction of the mechanical properties of TPMS structures based on Back propagation neural network

Triply Periodic Minimal Surface (TPMS) has the characteristics of high porosity, a highly interconnected network, and a smooth surface, making it an ideal candidate for bone tissue engineering applications. However, due to the complex relationship between multiple parameters of the TPMS structure and mechanical properties, it is a challenging task to optimize the properties of TPMS structures with different parameters. In this study, a Back-Propagation Neural Network (BPNN) was utilized to construct the relationship between TPMS parameters. Its mechanical performance and the TPMS structure were optimized using the BPNN. Results indicated that after training the correlation coefficient (R) between the BPNN prediction and the experimental results is 0.955475, it shows that our BPNN model has an adequate accuracy in describing the TPMS structures properties. Result of TPMS structure optimization shows that after optimization the yield strength of Hybridized Gyroid-Diamond Structure (HGDS) is 6.20 MPa, which is increased by 102.61% when compared with the original Hybridized Gyroid-Diamond Structure (3.06 MPa). Result of topological morphology indicates the effective bearing area of the optimized model was increased by 12.92% compared with the original model, which ascribe the increase in yield strength of the optimization model.

Optimization of parameters for the best shot peening effect based on surface response and neural network model

To optimize the peening effect of different shot peening process parameters on metal surfaces, the mapping relationship between different shot peening process parameters and metal surface integrity was obtained. In this paper, ABAQUS software was used to establish a DE-FE (Discrete element-Finite element) random multi-shot analysis model to simulate shot peening, then optimize the shot peening process parameters based on the surface response method(RSM), and finally validate it through experiments and BP(back propagation) neural network model. The result shows that when the shot velocity is 70 m s−1, the impact angle of shot is 61.45°, and the shot diameter is 0.78 mm, the shot peening effect is the best, the surface roughness value is reduced by 101.84%, and the arc height value is increased by 54.66%; the error between the predicted results of BP neural network and the results of numerical analysis is less than 8%. Therefore, the optimized process parameters significantly improve the shot peening effect, but also shows that the BP neural network prediction model can more accurately predict the mapping relationship between the input parameters of shot velocity, shot diameter, and impact angle of shot and the output parameters of roughness value and arc height value.

Optimization of extraction process of Astragali radix-Paeoniae radix alba by comparing BBD-RSM and SA-BPNN and antioxidant activity

The Box-Behnken design-response surface methodology (BBD-RSM) and Back-propagation neural network prediction model optimized by Simulated annealing (SA-BPNN) were used to optimize the extraction process of Astragali radix-Paeoniae radix alba by the green natural deep eutectic solvent (NADES), and the antioxidant activity of them was discussed. The best NADES for extracting Astragali radix-Paeoniae radix alba need to be screened out. The comprehensive evaluation value of the extraction yield was used as the index. The BBD-RSM and SA-BPNN were established for prediction and optimization. The antioxidant activity of Astragali radix-Paeoniae radix alba in vitro and in vivo was studied by detecting 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical scavenging rate and ferric reducing antioxidant power (FRAP) and the expression levels of SOD, MDA and GSH-Px in rat plasma. The results showed that the best NADES formula screened was Betaine-Urea (Bet-Ur). Compared with the BBD-RSM, the SA-BPNN had a higher degree of fitting, the comprehensive evaluation value was higher, and the error between tested and predicted value was smaller, which was more suitable as a prediction model for the extraction yield of Astragali radix-Paeoniae radix alba. Through the determination of antioxidant kits and oxidative factors, the results showed that Astragali radix-Paeoniae radix alba had antioxidant activity in vitro and in vivo. The study on the extraction process optimization and antioxidant activity of Astragali radix-Paeoniae radix alba were expected to provide a theoretical basis for its effective utilization.

Novel Life Prediction Method of PMMA for Cultural Relics Protection Based on the BP Neural Network.

Poly(methyl methacrylate) (PMMA) is widely used in the preservation and exhibition of cultural relics in museums. Accurately predicting its service life can help avoid many negative effects caused by PMMA aging. To study the change in the yellowing index of PMMA after aging in a UV light environment, an aging experiment was conducted. A prediction model for the service life of PMMA was established using nonlinear curve fitting and a back propagation (BP) neural network. By comparing the goodness of fit, simulation and modeling capabilities of the initial data, and the predictive ability for new data, it was found that the BP neural network prediction model outperformed the nonlinear curve fitting prediction model. In this study, the service life of newly produced PMMA samples was calculated as 7.83, 8.47, and 8.42 years, based on the yellowing index of retired PMMA as a benchmark and using the output data from the BP neural network prediction model. At this time, the performance and exhibition effect of the PMMA are poor, and the batch of PMMA needs to be updated.

An indentation method for measuring welding residual stress: Estimation of stress-free indentation curve using BP neural network prediction model

The measurement of welding residual stress (WRS) is necessary to formulate the stress control strategies, in order to ensure the reliability of engineering structures during their service life. The instrumented indentation technique can detect surface residual stress nondestructively with an acceptable accuracy, making it a desirable option for in-situ applications. However, it is difficult to meet the requirement for stress-free samples, particularly in the case of welded joints with non-uniform mechanical properties. In this work, the impact of stress-free indentation curves on determining WRS is furtherly studied based on the previously proposed indentation energy difference method. The WRS in a mismatched welded plate is measured using single reference method and distributed reference method, respectively. The study suggests that using the single reference method can lead to significant errors in determining the sign and magnitude of WRS in the weld zone. A prediction model combining the continuous spherical indentation (CSI) method and Back Propagation (BP) neural network is proposed to obtain the stress-free indentation curves of local positions nondestructively. And the CSI-BP method is verified through reverse finite element analysis and experiments.

Multi-Algorithm Hybrid Optimization of Back Propagation (BP) Neural Networks for Reference Crop Evapotranspiration Prediction Models

The reference crop evapotranspiration (ET0) statistic is useful for estimating agricultural system water requirements and managing irrigation. In dry areas, the accurate calculation of ET0 is crucial for optimal agricultural water resource utilization. By investigating the relationship between meteorological information and ET0 in Shihezi City, four prediction models were developed: a BP neural network prediction model, a BP neural network prediction model improved by genetic algorithm (GA-BP), a BP neural network prediction model improved by particle swarm algorithm (PSO-BP), as well as an improved hybrid BP neural network prediction model (GA-PSO-BP). The Pearson correlation analysis found that the key parameters influencing ET0 were temperature (Tmax, Tave, Tmin), hours of sunshine (N), relative humidity (RH), wind speed (U), as well as average pressure (AP). Based on the analysis results, different combinations of meteorological input factors were established for modeling, and the results showed that when the input factors were temperature (Tmax, Tave, Tmin), hours of sunshine (N), as well as relative humidity (RH), the overall effect of the ET0 prediction model was better than the other input combinations, and the GA-PSO-BP prediction model was the best, which could provide some guidance for the deployment and use of water resources. This may assist in the allocation and utilization of agricultural water resources in Shihezi.

Prediction of Aircraft Arrival Runway Occupancy Time Based on Machine Learning

Wake re-categorization (RECAT) has been implemented to improve runway capacity, and consequently, aircraft arrival runway occupancy time has become a crucial factor influencing runway capacity. Accurate prediction of the runway occupancy time can assist controllers in determining aircraft separation, thereby enhancing the operational efficiency of the runway. In this study, the GA–PSO algorithm is utilized to optimize the Back Propagation neural network prediction model using Quick access recorder data from various domestic airports, achieving high-precision prediction. Additionally, the SHapley Additive explanation model is applied to quantify the effect of each characteristic parameter on the arrival runway occupancy time, resulting in the prediction of aircraft arrival runway occupancy time. This model can provide a foundation for improving runway operation efficiency and technical support for the design of airport runway and taxiway structure.

Non-destructive prediction techniques for asphalt mixture based on back propagation neural networks

ABSTRACT The density of the asphalt mixture surface is an important indicator to evaluate the quality of construction and is of great importance to ensure the quality of construction and maintain the long-term serviceability of the pavement. The 3D ground-penetrating radar (3D GPR) and density prediction model can achieve rapid and continuous prediction of the density of asphalt pavement. In this study, a back propagation neural network (BPNN) density prediction model was established and combined with 3D GPR to conduct field experiments on a new asphalt concrete mid-surface layer to compare the density values obtained by the coring method in order to verify the model performance. The results show that the BPNN prediction model established in this study has good application to the overall density diagnosis of the road in the field. The average prediction error of the BPNN density prediction model is 0.184%, which is highly reproducible in practical applications.

Triaxial Compression Strength Prediction of Fissured Rocks in Deep-Buried Coal Mines Based on an Improved Back Propagation Neural Network Model

In deep coal mine strata, characterized by high ground stress and extensive fracturing, predicting the strength of fractured rock masses is crucial for stability analysis of the surrounding rock in coal mine strata. In this study, rock samples were obtained from construction sites in deep coal mine strata and intact, as well as fissured, rock specimens were prepared and subjected to triaxial compression tests. A numerical model based on the discrete element method was then established and the micro-parameters were calibrated. A total of 288 triaxial compression tests on the rock specimens under different conditions of confining pressure, loading rate, fissure dip angle, and fissure length, were conducted to obtain the triaxial compressive strength of the fractured rock specimens under different conditions. To address the limitations of traditional back propagation (BP) neural networks in solving stochastic problems, a modified BP neural network model was developed using a random factor and an interlayer mean square error corrected network model evaluation function. The traditional and modified BP neural network models were then employed to predict the triaxial compressive strength of the fractured rock specimens. Through comparative analysis, it was found that the modified BP neural network prediction model exhibited smaller errors and significantly reduced overfitting, making it an effective tool for predicting the strength of fractured rocks in deep coal mine strata.

A Double-Layer Vehicle Speed Prediction Based on BPNN-LSTM for Off-Road Vehicles.

The accurate prediction of vehicle speed is crucial for the energy management of vehicles. The existing vehicle speed prediction (VSP) methods mainly focus on road vehicles and rarely on off-road vehicles. In this paper, a double-layer VSP method based on backpropagation neural network (BPNN) and long short-term memory (LSTM) for off-road vehicles is proposed. First of all, considering the motion characteristics of off-road vehicles, the VSP problem is established and the relationship between the variables in the problem is carefully analyzed. Then, the double-layer VSP framework is presented, which consists of speed prediction and information update layers. The speed prediction layer established by using LSTM is to predict vehicle speed in the horizon, and the information update layer built by BPNN is to update the prediction information. Finally, with the help of mining truck and loader operation scenarios, the proposed VSP method is compared with the analytical method, BPNN prediction method, and recurrent neural network (RNN) prediction method in terms of speed prediction accuracy. The results show that, under the premise of ensuring the real-time prediction performance, the average prediction error of the proposed BPNN-LSTM prediction method under two operation scenarios reduces by 48.14%, 35.82% and 30.09% compared with the other three methods, respectively. The proposed speed prediction method provides a new solution for predicting the speed of off-road vehicles, effectively improving the speed prediction accuracy.