BP Model Research Articles

Simultaneous identification of groundwater contaminant source and hydraulic parameters based on multilayer perceptron and flying foxes optimization.

Groundwater contaminant source identification (GCSI) has practical significance for groundwater remediation and liability. However, when applying the simulation-optimization method to precisely solve GCSI, the optimization model inevitably encounters the problems of high-dimensional unknown variables to identify, which might increase the nonlinearity. In particular, to solve such optimization models, the well-known heuristic optimization algorithms might fall into a local optimum, resulting in low accuracy of inverse results. For this reason, this paper proposes a novel optimization algorithm, namely, the flying foxes optimization (FFO) to solve the optimization model. We perform simultaneous identification of the release history of groundwater pollution sources and hydraulic conductivity and compare the results with those of the traditional genetic algorithm. In addition, to alleviate the massive computational load caused by the frequent invocation of the simulation model when solving the optimization model, we utilized the multilayer perception (MLP) to establish a surrogate model of the simulation model and compared it with the method of backpropagation algorithm (BP). The results show that the average relative error of the results of FFO is 2.12%, significantly outperforming the genetic algorithm (GA); the surrogate model of MLP can replace the simulation model for calculation with fitting accuracy of more than 0.999, which is better than the commonly used surrogate model of BP.

Mach Number Prediction for 0.6 m and 2.4 m Continuous Transonic Wind Tunnels

With the development of the design technology, more and more advanced and diverse wind tunnels have been constructed to match complex requirements. However, it is hard to design a precise physical model of a wind tunnel that can be controlled. In addition, if a new wind tunnel is designed, the experimental data may be insufficient to build a controlling model. This article reports research on the following two models: (1) for a 0.6 m continuous transonic wind tunnel supported by a large amount of historical data, the false nearest neighbor (FNN) algorithm was adopted to calculate the order of the input variables, and the nonlinear auto-regressive model with the exogenous inputs–backpropagation network (NARX-BP) was proposed to build its Mach number prediction model; (2) for a new 2.4 m continuous transonic wind tunnel with only a small amount of experimental data, the method of model migration, the input and output slope/bias correction–particle swarm optimization (IOSBC-PSO) algorithm, was developed to convert the old model of the 0.6 m wind tunnel into the new model of the 2.4 m wind tunnel, so that the new Mach number prediction could be conducted. Through simulation experiments, it was found that by introducing the NARX-BP algorithm to build the Mach number prediction model, the root-mean-square error (RMSE) of the model decreased by 44.93–77.90%, and the maximum deviation (MD) decreased by 64.05–85.32% compared to the BP model. The performance of the IOSBC-PSO migration model was also better than that of the non-migration model, as evidenced by the 82.06% decrease of the RMSE value and the 78.25% decrease of the MD value. The experiments showed the effectiveness of the proposed strategy.

Intelligent Inversion Analysis of Hydraulic Engineering Geological Permeability Coefficient Based on an RF–HHO Model

The permeability of the natural geology plays a crucial role in accurately analyzing seepage behavior in the project area. This study presents a novel approach for the inverse analysis of the permeability coefficient. The finite element model (FEM) combined with orthogonal experimental design is used to construct a sample set of permeability coefficient inversion. The established random forest (RF) algorithm surrogate model is applied to determine the optimal values of permeability parameters in the project area using the Harris hawk optimization (HHO) algorithm. This method was used to explore and verify the distribution of natural seepage fields for the P hydropower station. The results showed that the RF model outperformed the classical CART and BP models at each borehole regarding performance evaluation indices. Furthermore, the water head prediction results were more accurate, and the RF model performed admirably in terms of prediction, anti-interference, and generalization. The HHO algorithm effectively searched for the optimal permeability coefficient of the geology. The maximum value of the relative error of the borehole water head inverted was 1.11%, and the accuracy met engineering standards. The initial seepage field distribution pattern calculated followed the basic distribution pattern of the mountain seepage field.

Development of a NOx Calculation Model for Low-Speed Marine Diesel Engines Based on Soft Measurement Technology

With the increasing level of intelligence of marine engines, there is an increasing demand for the online monitoring of engines, and marine NOx emissions have been of great concern. In this paper, a NOx simulation model is developed based on virtual measurement technology, which can calculate and predict NOx emissions based on the current operating state parameters of low-speed two-stroke diesel engines. First, the calibrated 3D simulation model is used to design the experiments to obtain the simulation experimental samples. Based on the NOx generation mechanism and diesel engine work-related parameters, the relevant factors were selected as alternative input parameters for the NOx emission model. The correlation analysis was then performed on the input parameters using the grey relational analysis correlation method and the Pearson correlation coefficient, and the principal component analysis method was used to reduce the dimensionality of the relevant factors by minimizing the loss of important information in reducing the complexity of the whole model. Then, the structure-related parameters of the backpropagation neural network (BPNN) were adaptively optimized using the group method of data handling (GMDH) to improve the accuracy of the NOx soft measurement model. Finally, the developed GMDH–BP model was validated with data and compared with the error evaluation index of BPNN and BPNN optimized by genetic algorithm (GA), and the developed NOx simulation model demonstrated high prediction accuracy under the same hyperparameter settings. The result provides technical support for the subsequent realization of the real-time online monitoring of NOx emissions from low-speed marine diesel engines without NOx sensors.

GA−BP Prediction Model for Automobile Exhaust Waste Heat Recovery Using Thermoelectric Generator

Thermoelectric generator (TEG) has important applications in automotive exhaust waste heat recovery. The Back propagation neural network (BP) can predict the electrical generating performance of TEG efficiently and accurately due to its advantage of being good at handing nonlinear data. However, BP algorithm is easy to fall into local optimum, and its training data usually have deviation since the data are obtained through the simulation software. Both of the problems will reduce the prediction accuracy. In order to further improve the prediction accuracy of BP algorithm, we use the genetic algorithm (GA) to optimize BP neural network by selection, crossover, and mutation operation. Meanwhile, we create a TEG for the heat waste recovery of automotive exhaust and test 84 groups of experimental data set to train the GA−BP prediction model to avoid the deviation caused by the simulation software. The results show that the prediction accuracy of the GA−BP model is better than that of the BP model. For the predicted values of output power and output voltage, the mean absolute percentage error (MAPE) increased to 2.83% and 2.28%, respectively, and the mean square error (MSE) is much smaller than the value before optimization, and the correlation coefficient (R2) of the network model is greater than 0.99.

Hybrid machine learning-based prediction model for the bond strength of corroded Cr alloy-reinforced coral aggregate concrete

This study investigated the bond–slip behavior between newly developed Cr alloy steel bars and coral aggregate concrete (CAC) on the basis of experimental and hybrid machine learning. Several variable parameters were examined: (a) corrosion ratio, (b) diameter, (c) CAC strength, (d) anchorage length, and (e) relative protective layer thickness. Results showed that bond strength is positively correlated with relative protective layer thickness and CAC strength, whereas it is negatively correlated with diameter and anchorage length. The bond strength initially rose and subsequently fell as the corrosion ratio increased, with a critical threshold of 1%. The bond–slip curve of Cr–CAC had four stages: microslip, slip, declining, and residual stages. Additionally, the empirical, particle swarm optimization backpropagation (PSO-BP), BP, and support vector machine regression (SVR) models were evaluated using various performance evaluation metrics. The PSO-BP model outperformed the empirical, BP, and SVR models in all evaluation metrics. Compared with the PSO-BP model, the BP and SVR models had a 6.12% and 2.0% reduction in R2, 87.67% and 58.90% increase in root mean square error, 52.27% and 4.54% increase in mean absolute error, 254.72% and 152.83% increase in mean square error, and 76.52% and 52.22% increase in mean absolute percentage error, respectively. The PSO-BP model established in this research can quickly and accurately predict the bond strength of corroded Cr–CAC, providing a rapid assessment method for its application.

A novel structural damage detection method using a hybrid IDE–BP model

Recently, artificial intelligence has been widely applied in the field of structural damage detection. In this study, a back-propagation neural network (BPNN) hybrid with an improved differential evolution algorithm (IDE) is proposed to advance the development of structural damage detection. Generally, most engineering structures’ dynamic characteristics can be directly obtained by modal analysis, and the corresponding vibration features can provide information on structural damage. Thus, based on the results of modal analysis, artificial intelligent methods such as BP neural network find their way to locate the damage and quantify the damage severity. However, the existing research shows that the original BP has insufficient accuracy in structural damage detection, especially cannot effectively deal with complex structural damage. In this study, an improved differential evolution algorithm is proposed, introducing a group sampling strategy and an adaptive mutation strategy. Then, the proposed IDE and BP neural networks are hybridized into IDE–BP to promote the neural network’s data classification and regression ability. Subsequently, based on the IDE–BP model, a method for detecting the location and severity of multi-damage scenarios is established. The decision tree algorithm (DT) is utilized to identify the number of damaged elements. To verify the effectiveness of the proposed damage detection method, the IDE–BP model is applied to the damage detection for a steel pipe and a cantilever plate. To study the anti-noise ability and robustness of the IDE–BP model, the noise and incomplete modes are introduced. Compared with the original BP, SVM, KNN, PSO–BP,​ and DE–BP, the numerical results show that the damage detection method based on the IDE–BP model performs better for all single and multiple damage scenarios.

Construction Schedule of Medium Voltage Overhead Distribution Network Optimization Based on Neural Network Algorithm

INTRODUCTION: The medium voltage overhead distribution network is a complex planning problem, and monitoring the construction progress of the medium voltage overhead distribution network is an important aspect of the planning problem. Regarding the structure of the medium voltage overhead distribution network, it is one of the important components in China's distribution network system. If the construction progress of the medium voltage overhead distribution network is optimized, it will affect the normal operation of the entire power grid system in a certain region. Optimizing the construction progress of medium voltage overhead distribution networks using modern development related technologies or algorithms is one of the practical research topics. OBJECTIVES: To better predict various issues that may arise during the construction progress of the distribution network. Complex geographical conditions lead to construction difficulties, and lack of technology leads to frequent stoppages of construction, further leading to frequent power outages in various regions. In terms of the overall power supply system, the reliability of power supply is not high. METHODS: By comparing the BP neural network algorithm with the CNN network algorithm, the actual operation effect is evaluated by experts. RESULTS: This article predicts the implementation progress of the medium voltage overhead distribution network, and the results show that the accuracy of the BP model for predicting the construction progress of the medium voltage overhead distribution network can reach 88%; The accuracy of the CNN medium voltage overhead distribution network construction progress prediction model reaches 77%. CONCLUSION: The use of neural network algorithms to optimize the construction schedule of medium voltage overhead distribution networks has been evaluated by experts in practical applications, and the medium voltage overhead distribution network in power supply systems has potential application value.

Ultra-Short-Term Wind Power Prediction Based on LSTM with Loss Shrinkage Adam

With the rapid increase in wind power, its strong randomness has brought great challenges to power system operation. Accurate and timely ultra-short-term wind power prediction is essential for the stable operation of power systems. In this paper, an LsAdam–LSTM model is proposed for ultra-short-term wind power prediction, which is obtained by accelerating the long short-term memory (LSTM) network using an improved Adam optimizer with loss shrinkage (LsAdam). For a specific network topology, training progress heavily depends on the learning rate. To make the training loss of LSTM shrink faster with standard Adam, we use the past training loss-changing information to finely tune the next learning rate. Therefore, we design a gain coefficient according to the loss change to adjust the global learning rate in every epoch. In this way, the loss change in the training process can be incorporated into the learning progress and a closed-loop adaptive learning rate tuning mechanism can be constructed. Drastic changes in network parameters will deteriorate learning progress and even make the model non-converging, so the gain coefficient is designed based on the arctangent function with self-limiting properties. Because the learning rate is iteratively tuned with past loss-changing information, the trained model will have better performance. The test results on a wind turbine show that the LsAdam–LSTM model can obtain higher prediction accuracy with much fewer training epochs compared with Adam–LSTM, and the prediction accuracy has significant improvements compared with BP and SVR models.

Prediction of Chemical Corrosion Rate and Remaining Life of Buried Oil and Gas Pipelines in Changqing Gas Field

Green synthesis and metal oxide composites have attracted much attention from researchers of industry and academia. As a typical application of green synthesis and metal oxide composites, the continuous change of industrial technology and the continuous improvement of the social and economic level, the demand for oil and gas are also increasing. However, the spatial gap between the place of origin and the place of demand for oil and gas resources is large, so the long-distance oil and gas pipeline came into being. However, under the action of time, coupled with the corrosion effect of the soil due to deep burial, some pipelines have serious aging and corrosion phenomena. Therefore, in order to give corresponding guarantees for economic development, we need to conduct in-depth research and analysis of the corrosion of oil and gas long-distance pipelines and give effective solutions. In this paper, the corrosion rate prediction of buried oil and gas pipelines is studied in Changqing gas field. By improving the inertial weights and learning factors of the traditional particle swarm algorithm, the parameters of the generalized regression neural network are optimized and selected, and the corrosion rate prediction model of buried pipelines is finally constructed. Comparative analysis with other swarm intelligence algorithms shows that the improved particle swarm algorithm has stronger convergence ability and higher prediction accuracy than the BP model and SVM model. In addition, based on the detection data collected at the site of the gathering and transportation pipeline in Changqing gas field, this paper uses the extreme value distribution theory and the local corrosion progress formula to establish a prediction model for the residual life of corrosion of buried pipelines. The model established in this paper can effectively determine the risk pipe segment of buried pipeline and provide a decision-making basis for pipeline management departments. The work provides an important application guidance to green synthesis and metal oxide composites.

Research on the development and application of CNN model in mobile payment terminal and blockchain economy

As the most widely used payment method at this stage, mobile payment is more and more closely related to the blockchain economy. Traditional methods lack a certain degree of accuracy. This research proposes a feature-based and sequential-based Bilateral AM (BAM) and Convolutional Neural Network (CNN)-gated recurrent unit for the development and application of mobile payment and blockchain economy (Gated Recurrent Unit, GRU) hybrid model (BAM-CNN-GRU), select 5 feature parameters with high correlation with the blockchain for multivariate prediction. The introduction of BAM can automatically quantify the correlation between the input variables and the blockchain, and strengthen the expression of historical key information on the predicted output; the introduction of CNN can extract high-dimensional features that reflect the non-stationary dynamic changes of the blockchain. The proposed hybrid model achieves good results in both single-step and multi-step long-term series and multivariate input blockchain prediction. Compared with the other six methods, MAE is reduced by 75.45%, 64.74%, 62.84%, respectively. 59.41%, 45.54%, 44.16%. Compared with the BAM-GRU model, the CNN-GRU model, the GRU model, the LSTM model, the support vector machine SVM model and the BP model, the prediction accuracy of the hybrid model has been greatly improved, and it has a broader application prospect.

An Ultra-Short-Term PV Power Forecasting Method for Changeable Weather Based on Clustering and Signal Decomposition

Photovoltaic (PV) power shows different fluctuation characteristics under different weather types as well as strong randomness and uncertainty in changeable weather such as sunny to cloudy, cloudy to rain, and so on, resulting in low forecasting accuracy. For the changeable type of weather, an ultra-short-term photovoltaic power forecasting method is proposed based on affinity propagation (AP) clustering, complete ensemble empirical mode decomposition with an adaptive noise algorithm (CEEMDAN), and bi-directional long and short-term memory network (BiLSTM). First, the PV power output curve of the standard clear-sky day was extracted monthly from the historical data, and the photovoltaic power was normalized according to it. Second, the changeable days were extracted from various weather types based on the AP clustering algorithm and the Euclidean distance by considering the mean and variance of the clear-sky power coefficient (CSPC). Third, the CEEMDAN algorithm was further used to decompose the data of changeable days to reduce its overall non-stationarity, and each component was forecasted based on the BiLSTM network, so as to obtain the PV forecasting value in changeable weather. Using the PV dataset obtained from Alice Springs, Australia, the presented method was verified by comparative experiments with the BP, BiLSTM, and CEEMDAN-BiLSTM models, and the MAPE of the proposed method was 2.771%, which was better than the other methods.

A Deep Learning Framework for Day Ahead Wind Power Short-Term Prediction

Due to the increasing proportion of wind power connected to the grid, day-ahead wind power prediction plays a more and more important role in the operation of the power system. This paper proposes a day-ahead wind power short-term prediction model based on deep learning (DWT_AE_BiLSTM). Firstly, discrete wavelet transform (DWT) is used to denoise the data, then an autoencoder (AE) technology is used to extract the data features, and finally, bidirectional long short-term memory (BiLSTM) is used for prediction. To verify the effectiveness of the proposed DWT_AE_BiLSTM model, we studied three different power stations and compared their performance with the shallow neural network model. Experimental analysis shows that this model is more competitive in forecasting accuracy and stability. Compared with the BP model, the proposed model has increased by 3.86%, 3.22% and 3.42% in three wind farms, respectively.

A Data-Driven Indirect Approach for Predicting the Response of Existing Structures Induced by Adjacent Excavation

A data-driven indirect approach for predicting the response of existing structures induced by excavation is hereby proposed based on making full use of monitoring data during excavation, which can predict the deformation history of the research object during excavation. In this article, a machine-learning-based model framework for implementing the proposed approach is constructed and the treatment of key issues in the design and implementation of the proposed method is described in detail including the theoretical framework, the implementation mode of the method, the dimensionality reduction of the model parameters, and the normalization of data for model. On this basis, three models are provided to predict the settlement of buildings induced by adjacent excavation, namely the SVM model, BP model, and BP–SVM model. Relying on an excavation project for a subway in Xuzhou, Jiangsu Province, China, the proposed method is verified, and some conclusions are obtained.

A data driven model for estimating the fatigue life of 7075-T651 aluminum alloy based on the updated BP model

pitting corrosion is an important factor, which leads to stress concentration. A sharp stress concentration will have a big impact on the fatigue life of 7075- T651 aluminum alloy. Most of the studies have been conducted to investigate its fatigue performance and proposed some estimated models. However, these models only considered the influence of the depth, the length, and the width of pits on the fatigue life and neglected their coupling effects. In order to predict the fatigue life of 7075- T651 aluminum alloy accurately. First, some numerical models were established to explore the stress concentration factor (SCF) affected by pit sizes. Second, some data driven models were proposed to investigate it. The SCF was regarded as the output and the length, the width, and the depth of pits were acted as the inputs. Third, the fatigue performance of the aluminum alloy 7075-T651 was explored, based on the fracture mechanics theory. The results indicate that the neural network optimized by the sparrow search algorithm not only has a relatively small error but also has stronger robustness, which can predict the SCF well. In addition, the SCF increases, as the length and the width of the pits increase, and decreases, as the pit depth increases. The pit depth can be treated as an important factor to influence the fatigue life of 7075- T651 aluminum alloy.

Low cycle fatigue life prediction of titanium alloy using genetic algorithm-optimized BP artificial neural network

This paper proposes a novel approach for estimating titanium alloy structural parts’ low cycle fatigue (LCF) life based on the continuous damage mechanics (CDM) model. Using a genetic algorithm-optimized back-propagation artificial neural network (GABP-ANN), the LCF life of titanium alloy structural parts is accurately predicted. Firstly, experimental data and finite element simulation data are combined to constitute a fatigue life database, and the relative error is taken as the training target for the BP neural network. The relationship is established between the relative error and the initial parameters (such as hidden layer nodes and learning rate). Secondly, the GABP-ANN model improves the shortage of the BP model at randomly selected initial training parameters and further enhances the predictive performance. Then, the accuracy and stability of five titanium alloy structural parts under two machine learning (ML) models are analyzed. The experimental values, finite element model (FEM) simulated values, and predicted values of the two ML models are compared and analyzed. It is found that the GABP-ANN model has a significant advantage in predicting the fatigue life of titanium alloy structural parts. Finally, it is verified that the proposed ML models have better data learning ability. The above results indicate that the GABP-ANN technique provides a highly accurate and stable method for predicting the LCF life of titanium alloy structural parts.

Air Quality—Meteorology Correlation Modeling Using Random Forest and Neural Network

Under the global warming trend, the diffusion of air pollutants has intensified, causing extremely serious environmental problems. In order to improve the air quality–meteorology correlation model’s prediction accuracy, this work focuses on the management strategy of the environmental ecosystem under the Artificial Intelligence (AI) algorithm and explores the correlation between air quality and meteorology. Xi’an city is selected as an example. Then, the theoretical knowledge is explained for Random Forest (RF), Backpropagation Neural Network (BPNN), and Genetic Algorithm (GA) in AI. Finally, GA is used to optimize and predict the weights and thresholds of the BPNN. Further, a fusion model of RF + BP + GA is proposed to predict the air quality and meteorology correlation. The proposed air quality–meteorology correlation model is applied to forest ecosystem management. Experimental analysis reveals that average temperature positively correlates with Air Quality Index (AQI), while relative humidity and wind speed negatively correlate with AQI. Moreover, the proposed RF + BP + GA model’s prediction error for AQI is not more than 0.32, showing an excellently fitting effect with the actual value. The air-quality prediction effect of the meteorological correlation model using RF is slightly lower than the real measured value. The prediction effect of the BP–GA model is slightly higher than the real measured value. The prediction effect of the air quality–meteorology correlation model combining RF and BP–GA is the closest to the real measured value. It shows that the air quality–meteorology correlation model using the fusion model of RF and BP–GA can predict AQI with the utmost accuracy. This work provides a research reference regarding the AQI value of the correlation model of air quality and meteorology and provides data support for the analysis of air quality problems.

Russian housing price forecast based on BP neural Network

Contemporarily, Russia's real estate market has huge potential for investors. The purpose of this paper is to investigated the impact of various factors on the real estate market price based on various statistical models. We collected past housing price data from the Internet. Subsequently, PCA model was adopted to reduce the dimension of the data to simplify the calculation. Finally, the BP model is implemented to predict the price changes in the real estate market. After calculation, our forecast data and the reality of housing prices are slightly different, but the forecast of the real estate market is consistent with the reality. In order to further find the influence of various factors on housing price, we also calculated the characteristic importance of different factors and found the most important influencing factor, i.e., the region. These results can provide investors with an idea to analyze the real estate market prices, and help policy makers find the real estate market prices in line with the market rules to avoid the occurrence of bubbles.

Research on Fault Diagnosis of HMCVT Shift Hydraulic System Based on Optimized BPNN and CNN

There are some problems in the shifting process of hydraulic CVT, such as irregularity, low stability and high failure rate. In this paper, the BP neural network and convolutional neural network are used for fault diagnosis of the HMCVT hydraulic system. Firstly, through experiments, 120 groups of pressure and flow data under normal and four typical fault modes were obtained and preprocessed; they were divided into 80 groups of training samples and 40 groups of test samples via random extraction, using the BP neural network model and convolutional neural network model for fault classification. The results show that compared with BP, PSO-BP and other models, the fault diagnosis rate of the BAS-BP neural network model can reach 92.5%, and the average diagnosis accuracy rate of the convolutional neural network can reach 97.5%, which can be effectively applied to the fault diagnosis of the HMCVT hydraulic system and provide some reference for the shifting reliability of hydraulic CVT.

Stock index time series prediction based on ensemble learning model

The accuracy of stock index prediction is of great significance to national economic development. However, because of the nonlinearity and long-term dependence of stock index data, effective prediction of future stock index price becomes a challenge. In order to solve the above problems, this paper proposes a research method of stock index time series prediction based on ensemble learning model. This method first uses an Adaboost.R2 algorithm to iteratively train multiple LSTM models and then integrates these LSTM models based on the parameters obtained by iterative training. Finally, it uses the ensemble model to predict stock index time series data. This paper uses the Shanghai Composite index, CSI 300 index and Shenzhen Composite index as experimental data sets, and uses the BP model, CNN model and LSTM model as comparative models to conduct an experimental analysis. The experimental results show that the new ensemble learning model proposed in this paper has certain advantages in the research of stock index time series prediction.