Real-time Prediction Method Research Articles

Real-time prediction of ROP based on GRU-Informer

Accurate ROP (rate of penetration) prediction contributes to better production task planning, ensuring efficient production line operation, and reducing production costs. ROP prediction is influenced by multiple factors, making accurate prediction challenging. Current research primarily relies on historical data for training and modeling, lacking methods for real-time ROP prediction. This paper introduces a GRU-Informer model for real-time ROP prediction. The model employs GRU (Gated Recurrent Unit) neural networks at the lower level to capture short-term correlations in drilling parameters and uses the Informer model at the top to address long-term dependencies among drilling parameters. Thus, the GRU-Informer can capture both short-term and long-term time dependencies, providing better ROP predictions. This paper constructs a dataset using historical data from a southwestern Chinese oil field for experimentation. RMSE (Root Mean Square Error), MAE (mean absolute error) and R2\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${R}^{2}$$\\end{document} (Coefficient of Determination) are employed as evaluation metrics for the model. Experimental results demonstrate that the GRU-Informer outperforms traditional recurrent neural networks like LSTM (Long Short-Term Memory), GRU neural networks and Informer in real-time ROP prediction, indicating its practical value.

Real-time analysis and prediction method of ion concentration using the effect of O-H stretching bands in aqueous solutions based on ATR-FTIR spectroscopy.

Analyzing the concentration of ions in aqueous solutions in real-time plays an important role in the fields of chemistry and biology. Traditional methods for measuring ion concentrations, such as concentration analysis by measuring electrical conductivity, inductively coupled plasma mass spectrometry, and ion chromatography, have been used in many research fields. However, these methods are limited in determining ion concentrations instantaneously. Fourier-transform infrared-attenuated total reflectance (ATR-FTIR) spectroscopy provides a new approach for determining ion concentrations in aqueous solutions. This allows for fast analysis without pretreatment and is scalable for real-time measurements. In this study, we present a method for measuring ion concentrations by examining ion-water interactions in the O-H stretching band of aqueous solutions using ATR-FTIR spectroscopy. Five aqueous solutions, namely LiCl + HCl, LiOH + HCl, LiOH, Li3PO4, and NaCl were used in the experiments and prepared at concentrations between 0.5-2 M. The ion concentrations in the prepared aqueous solutions were measured using ATR-FTIR spectroscopy. We observed that the difference in absorbance increased and decreased linearly with changes in concentration. The concentration of ions in the aqueous solution could be measured by validating the designed linear regression analysis function model. In this study, we proposed five linear regression analysis function models, all of which showed high coefficients of determination above 0.9, with the highest coefficient of determination reaching 0.9969. These results show that ATR-FTIR spectroscopy has the potential to be applied as a rapid and simple concentration analysis system.

Real-time prediction of wave-induced hull girder loads for a large container ship based on the recurrent neural network model and error correction strategy

Real-time acquisition of wave-induced hull girder loads of a sailing ship will help the captain make reasonable decisions, which is of great significance for improving the safety of the ship's navigation. This paper investigates the real-time prediction method of hull girder loads based on the Recurrent Neural Network (RNN) model and error correction strategy. Firstly, taking the vertical bending moment, horizontal bending moment, and torsional moment at the mid-ship position of a large container ship as examples, corresponding neural network prediction models are established through parameter influence analysis. Secondly, various sea state conditions are used to verify the feasibility of established network prediction models to predict the hull girder loads in real-time. The VBM prediction model performs better than the TM prediction model and HBM prediction model, and the errors of the TM prediction model and HBM prediction model are slightly larger in some cases. Lastly, an improved prediction model based on an error correction strategy is proposed to improve the prediction accuracy of the neural network prediction model, and the adequate performance of the error correction strategy is discussed.

Research on driver posture recognition and prediction in vehicle

The statistics show that more than 90% of traffic accidents are caused by the individual factors of the driver. Therefore, real-time monitoring of the current driver’s posture in the car and further predicting of the future driving posture are of great significance for designing the Adaptive Restraint System (ARS) to improve the safety protection of drivers. In this study, a real-time vehicle driver posture recognition and prediction method is proposed. The traditional Kinect-OpenPose algorithm, which has the problems such as lack of depth value, key point pulsation and measurement error of Kinect, was optimized. The OpenPose network structure and skeleton model are modified to improve the algorithm. In this study, a key point motion trajectory prediction network based on Long Short Term Memory (LSTM) is developed. The hyperparameter optimization method is combined with multi-layer grid search to obtain the best prediction effect in braking and steering conditions, and the driver’s future posture can be obtained by predicting the subsequent movement trajectory of key points. It has been verified that the method in this study has high accuracy, high robustness and low cost in key point recognition and prediction. And the improved processing speed of the algorithm did not affect the accuracy of driving posture recognition and prediction. This method predicted the driver’s posture in future of the collision, which can be used to guide the development of controlling system of the ARS.

IoT-Enabled Fault Prediction and Maintenance for Smart Charging Piles

With the application of the Internet of Things (IoT), smart charging piles, which are important facilities for new energy electric vehicles (NEVs), have become an important part of the smart grid. Since the smart charging piles are generally deployed in complex environments and prone to failure, it’s significant to perform efficient fault diagnosis and timely maintenance for them. One of the key problems to be solved is how to conduct fault prediction based on limited data collected through IoT in the early stage and develop reasonable preventive maintenance strategies. In this paper, a real-time fault prediction method combining costsensitive logistic regression (CS-LR) and cost-sensitive support vector machine classification (CS-SVM) is proposed. CS-LR is firstly used to classify the fault data of smart charging piles, then the CS-SVM is adopted to predict the faults based on the classified data. The feasibility of the proposed model is illustrated through the case study on fault prediction of real-world smart charging piles. To demonstrate the advantage in prediction accuracy, the proposed fault prediction model is compared with the classic baseline models such as LR, SVM, decision tree (DT), K-nearest neighbor (KNN), and back propagation neural network (BPNN). Finally, based on the proposed fault prediction method, preventive maintenance based on a probability threshold with the minimum total expected cost is proposed. Simulation results show that the proposed maintenance strategy has a better performance in reducing the total maintenance cost compared with traditional periodic maintenance. This is valuable for the development of preventive maintenance strategies for repairable systems under early real-time monitoring data.

Refined self-attention mechanism based real-time structural response prediction method under seismic action

Accurate prediction of structural response under earthquake is of great significance for structural damage and performance evaluation. In order to improve the efficiency of structure time history response prediction, this paper proposes a novel SeisFormer model based on the self-attention mechanism and deep learning technology. Through autoregressive prediction, the SeisFormer can achieve real-time prediction of the response time histories of a large number of nodes in the structure under seismic action and can effectively solve the problem of data scarcity. Four case studies are performed to verify the accuracy and efficiency of the proposed methodology, including validation on datasets obtained from elastoplastic seismic analysis of a single-story structure, a three-story structure, and an eleven-story structure, and measured data of a shaking table test model. In addition, this paper further studies the prediction accuracy of the SeisFormer through ablation experiments and comparative experiments. The experimental results show that the SeisFormer can accurately predict the acceleration, velocity, and displacement time histories of numerous nodes in the structure. The prediction accuracy outperforms the LSTM model, and the prediction speed is 193–109,824 times faster than finite element method. Furthermore, with data augmentation through autoregressive prediction, the SeisFormer model can achieve efficient and accurate predictions when training data is exceptionally scarce, enabling engineering applications.

Real-time prediction models for remaining cold start time in proton exchange membrane fuel cell based on stack temperature

In order to mitigate the irreversible damage caused by cold start and preserve cell performance, this paper proposes a real-time prediction method based on the remaining cold start time of the proton exchange membrane fuel cell (PEMFC). This method can protect the cell by referencing the current cold start effect. In this work, a real-time prediction model for the remaining cold start time based on the backpropagation (BP) neural network, long short-term memory (LSTM) neural network, and nonlinear regression of logarithmic function is established. Furthermore, a combined prediction model with variable weight is established by integrating these three prediction methods. Results demonstrated that, when considering two distinct cold start experimental conditions as samples, the average prediction accuracy of the three single models across all prediction times is 78.80%, 71.76%, 67.07% and 84.72%, 84.66%, 87.03%, respectively. In comparison to the optimal single prediction model, the combined prediction model exhibits an improvement in prediction accuracy by 9.84% and 2.83% under the two different experimental conditions, respectively. In addition, the maximum average prediction error between predicted values and actual values at each time instant within a range of only 3 s, thereby providing substantial support for accurate cold start performance prediction.

Physics-Informed Deep Learning-Based Real-Time Structural Response Prediction Method

High-precision and efficient structural response prediction is essential for intelligent disaster prevention and mitigation in building structures, including post-earthquake damage assessment, structural health monitoring, and seismic resilience assessment of buildings. To improve the accuracy and efficiency of structural response prediction, this study proposes a novel physics-informed deep-learning-based real-time structural response prediction method that can predict a large number of nodes in a structure through a data-driven training method and an autoregressive training strategy. The proposed method includes a Phy-Seisformer model that incorporates the physical information of the structure into the model, thereby enabling higher-precision predictions. Experiments were conducted on a four-story masonry structure, an eleven-story reinforced concrete irregular structure, and a twenty-one-story reinforced concrete frame structure to verify the accuracy and efficiency of the proposed method. In addition, the effectiveness of the structure in the Phy-Seisformer model was verified using an ablation study. Furthermore, by conducting a comparative experiment, the impact of the range of seismic wave amplitudes on the prediction accuracy was studied. The experimental results show that the method proposed in this paper can achieve very high accuracy and at least 5000 times faster calculation speed than finite element calculations for different types of building structures.

Real-time energy consumption prediction method for air-conditioning system based on long short-term memory neural network

With increasing energy consumption, how to achieve the energy-saving operation of air-conditioning systems is crucial for improving the energy efficiency of buildings. The accurate and reliable energy consumption prediction of air-conditioning systems can be useful for optimizing the energy supply and equipment operation strategies. However, most existing studies focus on the prediction of the long-term energy consumption of air-conditioning systems, which usually exceeds the individual control time steps of the air-conditioning system. If the energy consumption of air-conditioning systems in the next control step can be predicted, then the least energy-consuming operating strategy can be selected at the beginning of each control step by relying on the predicted energy consumption, thereby achieving greater energy conservation. This paper proposes a real-time energy consumption prediction model for air-conditioning systems based on a long short-term memory neural network. The correlation between the energy consumption of air-conditioning system and input variables can be obtained by calculating the Spearman correlation coefficient. The predicted model results show that the normalized root mean square error of the prediction model is 0.0429 when the indoor air temperature, condenser inlet temperature, and historical energy consumption are set as inputs to predict the energy consumption of the air-conditioning system for the next 15 min. Meanwhile, the energy consumption at a certain indoor temperature setpoint can be predicted by the prediction model to aid in selecting a more appropriate indoor temperature setpoint, which can reduce energy consumption by optimizing the operation strategy of the air-conditioning system at each control step.

A real-time operational carbon emission prediction method for the early design stage of residential units based on a convolutional neural network: A case study in Beijing, China

Buildings are currently responsible for 39% of global energy related carbon emissions. Thus, carbon reduction in the construction industry is crucial to global climate issues. This study contributes a real-time operational carbon emission prediction method for the early design stage of residential units based on a convolutional neural network (CNN). The CNN model was trained based on a dataset with 2000 real residential units focusing on typical cases in northern China, particularly in Beijing, which is a significant energy consumption area. The R2 values of the operational-stage heating and electricity carbon emissions in the validation dataset reached between 0.91 and 0.98. The trained model based on this method can provide the real-time prediction of the carbon emission potential of proposed residential units to buttress architects to make scientific design decisions. The model was then validated on new cases that varied in the total area, rooms, orientation, window locations, and window sizes. The performance of the CNN model is verified by comparing the results with the output of an EnergyPlus model. While the variation trends of the output results of the two methods are found to be consistent, the CNN model results still have certain deviations from those of the EnergyPlus model. Consequently, the proposed method can reflect the variation of operational-stage carbon emissions due to the change of the floor layouts of residential units, and has a faster speed and higher convenience than the simulation-based method.

A Real-Time Vehicle Speed Prediction Method Based on a Lightweight Informer Driven by Big Temporal Data

At present, the design of modern vehicles requires improving driving performance while meeting emission standards, leading to increasingly complex power systems. In autonomous driving systems, accurate, real-time vehicle speed prediction is one of the key factors in achieving automated driving. Accurate prediction and optimal control based on future vehicle speeds are key strategies for dealing with ever-changing and complex actual driving environments. However, predicting driver behavior is uncertain and may be influenced by the surrounding driving environment, such as weather and road conditions. To overcome these limitations, we propose a real-time vehicle speed prediction method based on a lightweight deep learning model driven by big temporal data. Firstly, the temporal data collected by automotive sensors are decomposed into a feature matrix through empirical mode decomposition (EMD). Then, an informer model based on the attention mechanism is designed to extract key information for learning and prediction. During the iterative training process of the informer, redundant parameters are removed through importance measurement criteria to achieve real-time inference. Finally, experimental results demonstrate that the proposed method achieves superior speed prediction performance through comparing it with state-of-the-art statistical modelling methods and deep learning models. Tests on edge computing devices also confirmed that the designed model can meet the requirements of actual tasks.

Intelligent real-time prediction of multi-region thrust of EPB shield machine based on SSA-LSTM

During the tunneling process of shield machine, the accurate determination of tunneling parameters is the guarantee of safe operation of shield machine. Through the real-time construction data analysis and mining of shield machine, the tunneling parameters at the next moment are obtained in real-time. Based on this, this paper proposes an intelligent real-time prediction method for multi-region thrust of EPB shield machine based on Sparrow Search algorithm-Long and Short-term Memory (SSA-LSTM). By correlation analysis of the construction big data, the data features with a great correlation of shield machine thrust are obtained. And it is used as the input of LSTM prediction model to explore the nonlinear relationship between inputs and output. SSA is used to optimize LSTM prediction model to establish a more accurate nonlinear relationship, and then the multi-region thrust of shield machine at the next moment is accurately obtained. The simulation results show that SSA-LSTM model can accurately predict the thrust of each region of shield machine at the next moment, and the prediction performance is better than other models. The method provides a reference basis for shield machine to implement accurate thrust regulation and provides a guarantee for effective control of Earth pressure balance (EPB) in sealed cabin to ensure construction safety, which has engineering application value.

Real-time detection of stealthy IoT-based cyber-attacks on power distribution systems: A novel anomaly prediction approach

IoT-enabled appliances are essential components in the development of modern smart homes and cities as they offer energy-efficient solutions that bring comfort to customers, resulting in a better quality of life. Despite the advantages, these smart appliances also present new privacy and security challenges that create vulnerabilities in both smart homes and Power Distribution Systems (PDSs), ultimately affecting their efficiency and reliability. Hence, identifying the security risks linked with behind-the-meter (BTM) IoT-enabled appliances is crucial to develop appropriate countermeasures for safeguarding the PDSs. Previous studies have primarily focused on developing countermeasures for attacks on utility-scale domains, while the problem of attack detection has been largely neglected in the literature. This paper fills the gap by presenting a stealthy attack model applied to a group of compromised IoT-enabled appliances manipulating their operation to deteriorate the power quality of a PDS. The presented attack minimizes the chance of getting detected by maintaining the operating conditions of such appliances within the consumers’ comfort levels. Then, a novel real-time anomaly detection and prediction method is proposed based on demand power spectrum analysis and Long-short-term memory (LSTM) neural networks. Further, a Feeder Loading Abnormal Power Spectrum (FLAPS) index is introduced and utilized to estimate the attack onset time, which is then reported to the PDS’s operator for further investigation and action. The evaluation of the stealthy attack model and the anomaly prediction approach is carried out on a residential community modeled using the IEEE-13 bus power distribution feeder, which exhibits various power consumption patterns and appliance operation schedules. The effects of the proposed stealthy attack on the PDS are analyzed quantitatively by examining the occurrences of voltage violations and the wear and tear of the line voltage regulator (LVR) of the system. Numerical results show that compromising only 5% of the PDS’s load, the LVR operates 1.8 more times than during normal conditions. This impact increases to 8 times at 50% compromised load with many voltage violation cases. Also, the simulation results demonstrate the superiority of the proposed prediction technique to detect stealthy attacks 50% faster than threshold-based approaches, thereby enabling the system’s operator to take faster actions.

CapRadar: Real-time adaptive bandwidth prediction for dynamic wireless networks

With the emergence of 4G/5G cellular networks, mobile Internet is becoming increasingly popular and numerous mobile applications have emerged, which puts higher demands on the accuracy of predicting available bandwidth. However, compared to WiFi and wired networks, the links of 4G/5G cellular networks are highly dynamic, and the unstable nature of their network environments makes the bandwidth trajectories vary. Traditional bandwidth prediction models rarely consider the bandwidth characteristics of various network scenarios, and a single prediction model is difficult to be applied to all scenarios, which makes it challenging to achieve high accuracy of available bandwidth prediction. To solve the above problems, we propose a real-time bandwidth prediction method called CapRadar. The method classifies bandwidth into scenarios and matches the optimal prediction model for each type of scenario, i.e., switches the prediction model in real time according to the changes of the scenario. Specifically, we first extract the statistical characteristics of the bandwidth using statistical method, and based on the extracted characteristics, a SVM classifier is used to detect different network scenarios. After that, the best prediction model is matched for them in the algorithm library based on the scenario classification results. The experimental results show that CapRadar can reduce the root mean square error (RMSE) by about 18.9% and the mean error (MAE) by about 21.5%. For practical applications, we use a pre-trained SVM model for real-time scenario detection, and then we can dynamically switch the prediction model.

A real-time temperature rise prediction method for PM motor varying working conditions based on the reduced thermal model

Real-time and accurate temperature rise calculation can play a preventive role in preventing overheating motor damage. This article proposes a temperature rise calculation method for varying working conditions PM motors based on a reduced order model. The calculation model of this method consists of a thermal conductance matrix and a heat source matrix. The thermal conductance matrix is obtained by reducing the order of the detailed lumped parameter thermal network (DLPTN) conductance matrix, which significantly reduces the model order while retaining computational accuracy. A loss prediction function is introduced in acquiring the heat source matrix, which is constructed based on the relationship between motor loss and speed and power. It can predict the stator iron loss and PM loss with different speeds and powers in real-time, ensuring the accuracy of heat source loading. The experimental results show that this method can predict the temperature rise of critical nodes of the PM motor in real-time under varying working conditions. Its calculation results have high accuracy, providing an effective method for overheating PM motor protection.

Near real-time spatial prediction of earthquake-induced landslides: A novel interpretable self-supervised learning method

ABSTRACT Near real-time spatial prediction of earthquake-induced landslides (EQILs) can rapidly forecast the occurrence position of widespread landslides just after a violent earthquake; thus, EQIL prediction is very crucial to the 72-hour ‘golden window’ for survivors. This work focuses on a series of earthquake events from 2008 to 2022 occurring in the Tibetan Plateau, a famous seismically-active zone, and proposes a novel interpretable self-supervised learning (ISeL) method for the near real-time spatial prediction of EQILs. This new method innovatively introduces swap noise at the unsupervised mechanism, which can improve the generalization performance and transferability of the model, and can effectively reduce false alarm and improve accuracy through supervised fine-tuning. An interpretable module is built based on a self-attention mechanism to reveal the importance and contribution of various influencing factors to EQIL spatial distribution. Experimental results demonstrate that the ISeL model is superior to the excellent state-of-the-art machine learning and deep learning methods. Furthermore, according to the interpretable module in the ISeL method, the critical controlling and triggering factors are revealed. The ISeL method can also be applied in other earthquake-frequent regions worldwide because of its good generalization and transferability.

Real-time monitoring and prediction method of commercial building fire temperature field based on distributed optical fiber sensor temperature measurement system

In order to accomplish the task of real-time monitoring and prediction of fire temperature in commercial buildings, this study proposes a sensor arrangement method based on sensor and fire field characteristics to better reflect the fire temperature changes of building space, and conducts high temperature sensor measurement experiments based on this method. Then the study carries out fire dynamics simulations to establish a temperature change dataset containing fire and sensor temperature changes. The study establishes several prediction models using different algorithms based on the dataset, and the models built by artificial neutral network (ANN) and long short-term memory (LSTM) respectively have the best performance. The results show that the percentage error between the output temperature of the ANN mapping model and the actual temperature of the fire key plane is 1.94%. The percentage error between the output temperature of the LSTM temperature prediction model and the actual temperature is 1.03%, which satisfies the demand for early warning of the fire temperature hazard. This study also validates the generalization of the two models for fire temperature prediction in different spatial areas, and the method can provide a good solution for real-time monitoring and early warning of fire temperature in commercial buildings.

Digital twin-long short-term memory (LSTM) neural network based real-time temperature prediction and degradation model analysis for lithium-ion battery

Real-time temperature prediction is essential to circumvent thermal safety issues for lithium-ion batteries (LIB). However, its industrial applications are challenging due to operating temperature, voltage range, capacity degradation, and current rate (C-rate) variations. This study proposes a digital twin (DT) technology and long short-term memory (LSTM) neural network-based method for real-time temperature prediction and degradation pattern analysis. The DT model is established based on lumped thermal equivalent circuits to describe the dynamic thermal behavior of LIB and identify the parameters by calculations. Furthermore, the real-time temperature prediction framework considering voltage, current, and operating temperature is further designed following identifying results, which consists of indicators extraction, correlation analysis, LSTM neural network, and DT model four parts. In addition, the incremental capacity analysis (ICA) method is used to analyze LIB degradation patterns. The experimental results indicate that the primary degradation pattern of the experimental sample is loss of lithium inventory (LLI), and the loss of active material (LAM) appears after 600 cycles. Moreover, the maximum error and root mean square error (RMSE) of the temperature prediction framework is 0.31 °C and 0.17 °C under the constant current (CC) discharging condition, 0.85 °C and 0.47 °C under the dynamic discharging condition, respectively. The results demonstrate that the proposed real-time temperature prediction framework delivers acceptable accuracy for CC discharging and dynamic discharging conditions, which can complete the requirements of practical applications. This study dramatically reduces the response time of temperature prediction and guides optimizing battery thermal management systems (BTMS).

Real-Time Trajectory Prediction Method for Intelligent Connected Vehicles in Urban Intersection Scenarios

Intelligent connected vehicles (ICVs) have played an important role in improving the intelligence degree of transportation systems, and improving the trajectory prediction capability of ICVs is beneficial for traffic efficiency and safety. In this paper, a real-time trajectory prediction method based on vehicle-to-everything (V2X) communication is proposed for ICVs to improve the accuracy of their trajectory prediction. Firstly, this paper applies a Gaussian mixture probability hypothesis density (GM-PHD) model to construct the multidimension dataset of ICV states. Secondly, this paper adopts vehicular microscopic data with more dimensions, which is output by GM-PHD as the input of LSTM to ensure the consistency of the prediction results. Then, the signal light factor and Q-Learning algorithm were applied to improve the LSTM model, adding features in the spatial dimension to complement the temporal features used in the LSTM. When compared with the previous models, more consideration was given to the dynamic spatial environment. Finally, an intersection at Fushi Road in Shijingshan District, Beijing, was selected as the field test scenario. The final experimental results show that the GM-PHD model achieved an average error of 0.1181 m, which is a 44.05% reduction compared to the LiDAR-based model. Meanwhile, the error of the proposed model can reach 0.501 m. When compared to the social LSTM model, the prediction error was reduced by 29.43% under the average displacement error (ADE) metric. The proposed method can provide data support and an effective theoretical basis for decision systems to improve traffic safety.

Penetration Overload Prediction Method Based on a Deep Neural Network with Multiple Inputs

In the process of high-speed penetration, penetrating ammunition is prone to problems such as penetration overload signal vibration and mixings and projectile attitude deflection. It is easy to misjudge if a fuze relies only on the overload data from the ground or the utilized program, and the actual penetration overload measured under actual launch conditions cannot be taken as the dynamic judgement basis. Therefore, a real-time penetration overload prediction method based on a deep neural network is proposed, which can predict overload values according to the projectile parameter settings, the real-time collection of overload information, and the calculation speed and assist the fuze in judging the target layer and projectile attitude. In this paper, we adopt a deep learning model with multiple time series inputs and modify the input coding mode so that the model can output a 48 μs overload curve within 20 μs, meeting the real-time signal processing requirements of the high-speed missile penetration process. The mean squared error between the predicted curve and the actual curve is 0.221 for the prediction of multilayer penetrating targets and 0.452 for the prediction of thick penetrating targets. A penetration overload prediction function can be realized.