Real-time Input Data Research Articles

AI-Driven Predictive Analysis for Urban Traffic Management: A Novel Approach

Urban traffic management remains one of the most complex challenges of modern cities, with congestion, inefficiencies, and accidents costing billions of dollars annually and contributing significantly to pollution and stress. Current traffic management systems are often reactive rather than predictive, responding to congestion and incidents after they occur. This paper introduces a novel AI-driven predictive analysis framework for urban traffic management that leverages advanced machine learning (ML) algorithms and real-time data inputs. The system aims to not only manage existing traffic efficiently but to predict congestion, optimize traffic flows, and enhance computer safety proactively. We explore the integration of multiple data sources—such as GPS data, traffic cameras, IoT sensors, and social media feeds—into a cohesive AI model that learns and evolves. The goal is to create a fully autonomous traffic management system that adjusts dynamically to urban changes, improving overall city mobility, sustainability, and quality of life.

Decision-Making Model for Life Cycle Management of Aircraft Components

This paper presents a novel decision-making framework for the life cycle management of aircraft components, integrating advanced data analytics, artificial intelligence, and predictive maintenance strategies. The proposed model addresses the challenges of balancing safety, reliability, and cost-effectiveness in aircraft maintenance. By using real-time health monitoring systems, failure probability models, and economic analysis, the framework enables more informed and dynamic maintenance strategies. The model incorporates a comprehensive approach that combines reliability assessment, economic analysis, and continuous re-evaluation to optimize maintenance, replacement, and life extension decisions. The optimization method on the base of genetic algorithm (GA) is employed to minimize total life cycle costs while maintaining component reliability within acceptable thresholds. The framework’s effectiveness is demonstrated through case studies on three distinct aircraft components: mechanical, avionics, and engine. These studies showcase the model’s versatility in handling different failure patterns and maintenance requirements. This study introduces a data-driven decision-making framework for optimizing the life cycle management of aircraft components, focusing on reliability, cost-effectiveness, and safety. To achieve optimal maintenance scheduling and resource allocation, a GA is employed, allowing for an effective exploration of complex solution spaces and enabling dynamic decision-making based on real-time data inputs. The GA-based optimization approach minimizes total life cycle costs while maintaining component reliability, with the framework’s effectiveness demonstrated through case studies on key aircraft components. Key findings from the case study demonstrate significant cost reductions through optimization, with mechanical components showing a 10% more reduction in total life cycle costs, avionics components achieving a 14% more cost reduction, and engine components demonstrating a 7% more decrease in total costs. The research also presents an optimized dynamic maintenance schedule that adapts to real-time component health data, extending component lifespans and reducing unexpected failures. The framework effectively addresses key industry challenges such as no fault found events while minimizing unexpected failures and enhancing the overall reliability and safety of aircraft maintenance practices. Sensitivity analysis further demonstrates the model’s robustness, showing stable performance under varying failure rates, maintenance costs, and degradation rates. The study contributes a scalable approach to predictive maintenance, balancing safety, cost, and resource allocation in dynamic operational environments.

A Dynamic System to Control the Entry of Non-Authorized Visitors and Detect Superspreader Farms in Strongly Interconnected Systems.

This study explores the critical challenges the livestock sector faces, particularly those related to biosecurity, animal welfare, and antibiotic use restrictions. It highlights the need to implement advanced information and communication technologies to enhance operational sustainability and decision-making. We introduce the Biorisk® External platform, a cloud-based visit control system designed to optimize biosecurity management by accurately tracking visitor activity through QR codes and GPS geolocation. During a 6-month study period from July to December 2023, we analyzed visits to 142 different swine production sites and 30 vehicle movement patterns. The analysis revealed trends in visitation patterns and compliance with biosecurity SOPs. The software categorized visits as authorized (A), not authorized with access (NAWA), and not authorized without access (NAWOA), providing a framework to assess biosecurity risks. Additionally, network analysis identified interconnected farms, which were classified as 'superspreaders', highlighting their considerable risk of disease transmission. This study advocates for the integration of digital systems in livestock operations to improve biosecurity measures, facilitate real-time data input, and support informed decision-making. By enhancing biosecurity protocols through technology, the livestock industry can better safeguard animal health, increase operational efficiency, and reduce potential economic losses associated with disease outbreaks.

A Digital Intervention for Capturing the Real-Time Health Data Needed for Epilepsy Seizure Forecasting: Protocol for a Formative Co-Design and Usability Study (The ATMOSPHERE Study).

Epilepsy is a chronic neurological disorder affecting individuals globally, marked by recurrent and apparently unpredictable seizures that pose significant challenges, including increased mortality, injuries, and diminished quality of life. Despite advancements in treatments, a significant proportion of people with epilepsy continue to experience uncontrolled seizures. The apparent unpredictability of these events has been identified as a major concern for people with epilepsy, highlighting the need for innovative seizure forecasting technologies. The ATMOSPHERE study aimed to develop and evaluate a digital intervention, using wearable technology and data science, that provides real-time, individualized seizure forecasting for individuals living with epilepsy. This paper reports the protocol for one of the workstreams focusing on the design and testing of a prototype to capture real-time input data needed for predictive modeling. The first aim was to collaboratively design the prototype (work completed). The second aim is to conduct an "in-the-wild" study to assess usability and refine the prototype (planned research). This study uses a person-based approach to design and test the usability of a prototype for real-time seizure precipitant data capture. Phase 1 (work completed) involved co-design with individuals living with epilepsy and health care professionals. Sessions explored users' requirements for the prototype, followed by iterative design of low-fidelity, static prototypes. Phase 2 (planned research) will be an "in-the-wild" usability study involving the deployment of a mid-fidelity, functional prototype for 4 weeks, with the collection of mixed methods usability data to assess the prototype's real-world application, feasibility, acceptability, and engagement. This phase involves primary participants (adults diagnosed with epilepsy) and, optionally, their nominated significant other. The usability study will run in 3 rounds of deployment and data collection, aiming to recruit 5 participants per round, with prototype refinement between rounds. The phase-1 co-design study engaged 22 individuals, resulting in the development of a mid-fidelity, functional prototype based on identified requirements, including the tracking of evidence-based and personalized seizure precipitants. The upcoming phase-2 usability study is expected to provide insights into the prototype's real-world usability, identify areas for improvement, and refine the technology for future development. The estimated completion date of phase 2 is the last quarter of 2024. The ATMOSPHERE study aims to make a significant step forward in epilepsy management, focusing on the development of a user-centered, noninvasive wearable device for seizure forecasting. Through a collaborative design process and comprehensive usability testing, this research aims to address the critical need for predictive seizure forecasting technologies, offering a promising approach to improving the lives of individuals with epilepsy. By leveraging predictive analytics and personalized machine learning models, this technology seeks to offer a novel approach to managing epilepsy, potentially improving clinical outcomes, including quality of life, through increased predictability and seizure management. DERR1-10.2196/60129.

Groundwater contamination modelling in Ayad River Basin, Udaipur

Groundwater, a vital freshwater resource catering to agricultural, domestic, and industrial needs, faces a pressing challenge of contamination due to escalating human activities. This study focuses on the Ayad River Basin in the Udaipur district of Rajasthan, employing the FEFLOW simulation code for the first time. A steady-state numerical model and a groundwater contaminant prediction model for total dissolved solids (TDS), nitrate, and fluoride were developed, simulating trends over the next five years with an accuracy exceeding 95%. The results reveal an eastward increase in TDS, nitrate, and fluoride concentrations, attributed to contamination from two waste disposal sites-Titadi and Baleecha. Titadi, operational for four decades until closure in 2010, retains residual waste over 32 thousand m2. The initiation of a new dumping ground at Baleecha by the Udaipur Municipal Corporation post-2010 exacerbates regional contamination. Nitrate contamination is particularly high in agricultural zones with excessive chemical fertilizer usage. Of the 27 scenarios tested, 23 support using the water for irrigation but would require treatment before using it for drinking. Recommendations include deploying a chemical sensor network for real-time data input into the web enabled FEFLOW model, real-time monitoring and alerts, and a mobile application providing personalized guidance on water usage and health risks in case of contamination. This study can be beneficial to decision-makers, who work on the policy and groundwater management strategies.

An intelligent airflow perception model for metal mines based on CNN-LSTM architecture

In view of the harsh underground working environment and difficulties of airflow monitoring sensors placement, implementing an intelligent ventilation strategy is crucial for ensuring ventilation safety during mining process. The key issue lies in obtaining global real-time airflow parameters for ventilation safety and intelligent control. Thus, we propose an intelligent perception approach based on artificial intelligence (AI) method for acquiring airflow parameters, which operates by leveraging partial airflow data from specific monitoring points to predict airflow at undisclosed locations. Initially, this approach is facilitated by the training of an AI database through numerical simulations of ventilation networks. Subsequently, an intelligent airflow perception model is constructed, incorporating convolution neural network (CNN), long short-term memory (LSTM), and hybrid CNN-LSTM architectures. Through iterative updates and enhancements, these models demonstrate average deviations between predicted and actual airflow parameters of less than 5% in both simulated scenarios and empirical applications. Furthermore, in the case study, the CNN-LSTM architecture model exhibits superior performance for intelligent airflow perception. This architecture combing with airflow monitoring system, and utilizing partial real-time data inputs to obtain perception point outputs, can effectively enhance employee productivity, reduce energy consumption, and prevent resource wastage.

Analisis Perancangan Sistem Jabatan Fungsional Pada UPT. Perlindungan Tanaman Pangan Dan Holtikultura Sumatera Utara

This scientific work aims to provide understanding and knowledge to readers and researchers about how important it is to carry out system analysis and warfare in a field, especially in the Functional field in order to improve or improve the quality of a company or the field itself. System design is the depiction, planning, sketching of several separate elements in one complete and functioning unit. This research uses a qualitative approach with descriptive analysis method. The results of the research are that from several weaknesses obtained in the process of observing the system running in the UPT Finance Section - Food Crop Protection and Horticulture, a system development plan is needed. One way is by optimizing computer use, such as designing a desktop-based system to input data in real time and then improving data validation mechanisms and storage processes. A development plan that can be carried out using a computer will simplify the process of sending reports from the field easily and quickly.

Super Real-Time Forecast of Wildland Fire Spread by A Dual-Model Deep Learning Method

Driven by climate change, more frequent and extreme wildfires have brought a greater threat to humans globally. Fastspreading wildfires endanger the safety of residents in the wildland-urban interface. To mitigate the hazards of wildfires and facilitate early evacuation, a rapid and accurate forecast of wildfire spread is critical in emergency response. This study proposes a novel dualmodel deep learning approach to achieve a super real-time forecast of 2-dimensional wildfire spread in different scenarios. The first model utilizes the U-Net technique to predict the burnt area up to 5 hours in advance. The second model incorporates ConvLSTM layers to refine the forecasted results based on real-time updated input data. To evaluate the effectiveness of this methodology, we applied it to Sunshine Island, Hong Kong, and generated a numerical database consisting of 210 cases (12,600 samples) to train the deep learning models. The simulated wildfire spread database has a fine resolution of 5 m and a time step of 5 minutes. Results show that both models achieve an overall agreement of over 90% between numerical simulation and AI forecast. The real-time wildfire forecasts by AI only take a few seconds, which is 102 ~ 104 times faster than direct simulations. Our findings demonstrate the potential of AI in offering fast and high-resolution forecasts of wildfire spread, and the novel contribution is to leverage two models which can work in tandem and be utilized at various stages of wildfire management.

An IoT based smart water quality assessment framework for aqua-ponds management using Dilated Spatial-temporal Convolution Neural Network (DSTCNN)

Assuring the quality of water is crucial for the growth and survival of fish in aquaculture ponds. Traditional methods of water quality monitoring can be inefficient which makes real-time monitoring and decision is a challenging one. Some deep learning techniques have shown apparent in improving water quality monitoring and assessment process, but encounter some limitations like data-overfitting, interpretability, and finds difficulties in capturing complex spatial and temporal dynamics that have hindered their effectiveness. To overcome these challenges, we propose an enhanced Dilated Spatial-temporal Convolution Neural Network (DSTCNN) for water quality monitoring in aquaculture, which uses an IoT system setup for capturing real-time data inputs from aqua ponds. The water quality data captured through the IoT sensors is labeled as per the water quality index (WQI) standards for analysis. This labeled data is effectively classified into two categories by the proposed DSTCNN model based on their suitability for fish growth or potential to cause fish mortality. By the leveraging power of dilated convolutions, the DSTCNN architecture accurately handles the intricacies of both spatial and temporal data, enabling it to capture essential features and patterns across multiple snapshots. This capability empowers the model to truly comprehend the complex relationships inherent in spatiotemporal data. Furthermore, to address the concerns like overfitting due to complexity of data and enhance generalization, the proposed model employs a hybrid activation function that synergistically combines ReLU and sigmoid during the activation process. The proposed DSTCNN model has been implemented on real-time and public datasets and obtained 99.28% and 99.02% accuracy respectively, whereas the state-of-the-art PCR-GB model obtains 96.97% and 97.11% accuracy on real-time and public datasets respectively.

Unraveling the Impact of Fitness Trackers: A Comprehensive Review

In the era of digital transformation, health and fitness have become paramount concerns for individuals seeking a balanced and active lifestyle. The rise of fitness trackers and technology-driven solutions has revolutionized the way people approach their well-being. Utilizing HTML, CSS, JavaScript, React, and database management systems, this research paper develops and implements an innovative fitness tracker website.The fitness tracker website outlined in this study offers a user-centric experience, allowing individuals to set personalized fitness goals and access tailored resources, exercises, and guidance. Through interactive interfaces and real-time data input, users can monitor their progress, receive expert advice, and stay motivated on their fitness journey. The platform's holistic approach encompasses user authentication, goal selection, exercise recommendations, and a comprehensive home page featuring fitness-related content. This research paper delves into the technical aspects of building a fitness tracker website, emphasizing the integration of frontend and backend technologies. The paper discusses the significance of user experience design, data validation, and dynamic content delivery. Furthermore, it explores the practical application of the fitness tracker, highlighting its potential to enhance user engagement, provide valuable fitness insights, and promote a healthier lifestyle.The findings presented in this research paper shed light on the evolution of fitness tracking platforms and their role in empowering individuals to achieve their fitness aspirations. By combining technological innovation with personalized fitness solutions, this fitness tracker website exemplifies a new paradigm in the realm of health and well- being. The study not only showcases the technical intricacies involved in its development but also underscores the broader impact of such platforms on promoting a culture of fitness and wellness in our digital age. KEYWORDS: Machine Health and Wellness , Digital Transformation , Technology-driven Solutions ,Fitness Tracker

A novel distributed approach for event-triggered economic dispatch of energy hubs under ramp-rate limits integrated with sustainable energy networks

This paper investigates a new consensus-oriented distributed approach for the event-triggered (ET) economic dispatch problem (EDP) over a smart grid under ramp-rate limits (RRLs) integrated with green power sources (GPSs) such as solar and wind energy for demand response strategies over hybrid energy power systems. To address the RRL condition, the authors have transformed the RRLs as minimum and maximum bounds on the derivative of generation for a generator. Then, a Karush–Kuhn–Tucker (KKT) condition and a more practical approximate KKT condition are developed for determining the optimality conditions. A practical ET protocol is proposed over a topology between generators by application of the proposed approximate KKT condition. In contrast to existing distributed optimization methods, this paper provides both optimally condition and distributed optimization scheme for dealing with RRLs integrated with sustainable hybrid energy systems. In addition, a computationally efficient ET mechanism, eliminating Zeno behaviour, has been considered for dealing with the efficient utilization of communication resources. This study incorporates the real-time input data from thermal production plants and GPSs for experimental analysis. RETScreen software having data-set of over 6,700 local meteorological stations is applied to obtain input data for GPSs. Furthermore, Weibull and Beta distribution functions have been applied for dealing with uncertainty in wind and solar energy sources. Two case studies are examined (with and without GPSs), and simulation results demonstrated the suitable performance of the proposed distributed ET EDP approach.

Data-Driven Active Learning Control for Bridge Cranes

For positioning and anti-swing control of bridge cranes, the active learning control method can reduce the dependence of controller design on the model and the influence of unmodeled dynamics on the controller’s performance. By only using the real-time online input and output data of the bridge crane system, the active learning control method consists of the finite-dimensional approximation of the Koopman operator and the design of an active learning controller based on the linear quadratic optimal tracking control. The effectiveness of the control strategy for positioning and anti-swing of bridge cranes is verified through numerical simulations.

Using Machine Learning to Predict the Density of Solar Winds

Solar winds are a variable flow of ions originating from the Sun’s corona that have the ability to travel to Earth and disrupt its magnetic field, leading to unpredictable space weather storms which can disturb communication systems and power grids. Prediction and therefore preparation for these solar winds can ensure stability in everyday life; however, an accurate predictor for solar wind features has yet to be developed. Previous models for solar wind prediction have been inhibited by incomplete datasets, correlated features, and over-simplification of solar winds. In this study, a linear regression model was fit to predict the density of solar winds using machine learning. The greater the density of a certain solar wind event, the greater the chance of collision with Earth’s magnetosphere, and therefore ramifications on Earth, due to the presence of more charged particles per unit space within the solar wind. The dataset used in this study was taken from a public NASA/NOAA dataset which included 1,048,575 solar wind events. After data preprocessing was performed, the working dataset was composed of 792,587 solar winds events. The model was then trained using the Scikit-Learn Python library to determine a linear regression equation. Following three different training strategies, the model had a relatively low MAE and RMSE score, demonstrating low variance between the true and predicted values of solar wind density. In the future, this model could be applied to real-time input data to warn against concerningly high solar wind density which may negatively impact life on Earth.

Active BIM system for optimized multi-project ready-mix-concrete delivery

PurposeReady-mix concrete delivery problem (RMCDP), a specific version of the vehicle routing problem (VRP), is a relevant supply-chain engineering task for construction management with various formulations and solving methods. This problem can range from a simple scenario involving one source, one material and one destination to a more challenging and complex case involving multiple sources, multiple materials and multiple destinations. This paper presents an Internet of Things (IoT)-supported active building information modeling (BIM) system for optimized multi-project ready-mix concrete (RMC) delivery.Design/methodology/approachThe presented system is BIM-based, IoT supported, dynamic and automatic input/output exchange to provide an optimal delivery program for multi-project ready-mix-concrete problem. The input parameters are extracted as real-time map-supported IoT data and transferred to the system via an application programming interface (API) into a mixed-integer linear programming (MILP) optimization model developed to perform the optimization. The obtained optimization results are further integrated into BIM by conventional project management tools. To demonstrate the features of the suggested system, an RMCDP example was applied to solve that included four building sites, seven eligible concrete plants and three necessary RMC mixtures.FindingsThe system provides the optimum delivery schedule for multiple RMCs to multiple construction sites, as well as the optimum RMC quantities to be delivered, the quantities from each concrete plant that must be supplied, the best delivery routes, the optimum execution times for each construction site, and the total minimal costs, while also assuring the dynamic transfer of the optimized results back into the portfolio of multiple BIM projects. The system can generate as many solutions as needed by updating the real-time input parameters in terms of change of the routes, unit prices and availability of concrete plants.Originality/valueThe suggested system allows dynamic adjustments during the optimization process, andis adaptable to changes in input data also considering the real-time input data. The system is based on spreadsheets, which are widely used and common tool that most stakeholders already utilize daily, while also providing the possibility to apply a more specialized tool. Based on this, the RMCDP can be solved using both conventional and advanced optimization software, enabling the system to handle even large-scale tasks as necessary.

An automated system of soil sensor-based site-specific seeding for silage maize: A proof of concept

Despite recent studies of map-based site-specific seeding (SSS) revealing improved agronomic and economic outcomes over uniform rate seeding (URS), to our best knowledge no visible and near infrared spectroscopy (vis-NIRS) sensor-based SSS studies exist to date. This study aimed to develop and evaluate an automated sensor-based SSS technology for silage maize (Zea mays L.) production. An on-line visible and near-infrared reflectance spectroscopy (vis-NIRS) sensor was installed in front of a tractor to provide real-time input data to control the seed rate using a precision seeding machine mounted at the back of the tractor. A LabVIEW-based software was developed and used to predict soil fertility index using on-line vis-NIR spectra, which was then used to calculate the seed rate and transfer it to the controller of the seeding machine. The agronomic and economic benefits of SSS were compared with URS under a one-site-year experiment. Results showed that the proposed sensor-based SSS technology was 87.5% efficient in controlling the desired seed rates, according to the soil fertility status. In parallel with the observed spatial similarity between predicted soil fertility index and actual seed rates, a strong linear association (R2 = 0.80) was also achieved. As a result, SSS improved silage yield by 1.4 t ha−1 (4.4%), while sowing a lower seed rate (86.4 kSeeds ha−1) than the URS (90 kSeeds ha−1). This improved gross margin by 91 € ha−1, of which only 7 € ha−1 was attributed to savings on seed cost. The proposed sensor-based SSS technology was technically sound and thus transforms within-field fertility variations into agro-economic benefits effectively.

Study on Queue Length in the Whole Process of a Traffic Accident in an Extra-Long Tunnel

To investigate traffic accident patterns in undersea tunnels and quantify the relationship among various factors and traffic accidents in undersea tunnels, we analyzed the rules of evolution during the entire process of vehicle congestion queuing. Additionally, we built a maximum queue-length estimation model based on shock-wave theory and a whole-process queue-length estimation model based on real-time data input. The results demonstrate that the model is most precise when the data are not smoothed and the time interval is 30 s. The maximum accuracy of the model is not improved by data smoothing processing, but it is substantially improved when the time interval is between 5 and 15 s. Various movable window lengths have no discernible effect on the results. Maximum queue-length estimation model accuracy is 92.34%, while real-time whole-process queue-length estimation model accuracy is 83.50%. The accuracy of the proposed model is greater than that of the input–output model, indicating that the proposed model can support timely and reasonable control measures.

Forecasting Intrusion in Critical Power Systems Infrastructure Using Advanced Autoregressive Moving Average (AARMA) Based Intrusion Detection for Efficacious Alert System

Cyber intrusions into critical infrastructure inflict economic and physical damage. Extensive research is needed to identify and mitigate intrusions in power grid infrastructure. The modern solution is to use a data science time-series approach to identify the intrusion based on the electric grid data collected from the sensors. This paper addresses the new vision of the data science time-series modelling approach to integrate it with the existing power system security system. In this paper, the Advanced Autoregressive Moving Average (AARMA) model is designed to detect the possible intrusion of the given data set. An attack forecast is a model to predict possible cyber intrusions using real-time data input from sensors. By investigating the statistical properties of the sensors’ data set, intrusion detection is possible with a high accuracy of about 90%. Using AARMA, the operators have the benefit of an effective alert system to adjust their configuration and other resource allocation to tackle intrusions with low impact. MATLAB software is used to monitor the IEEE 9-bus and IEEE 33-bus test system against possible cyber-attacks using the proposed AARMA model.

Energy Demand Model of Battery E-Buses for LPT: Implementation, Validation and Scheduling Optimization

The progressive conversion of conventional bus fleets into full-electric fleets have gained focus in recent years, instilled by awareness about the environment and significant trends of urbanization. Public transport operators in major cities worldwide have put efforts into fulfilling this change. However, an efficient electrification process is still a challenge for most operators. This paper aims to propose an E-Bus vehicle model that estimates the actual energy consumption. The proposed model is implemented on the case study of a real bus line for Local Public Transport (LPT) and considers all technical characteristics of the vehicle. Real-time input data are represented by real driving cycles of the actual bus fleet and slope profile of the line. Real-time input data allow to establish directly the influence on the energy consumed during real operations. As simulation results, the global energy consumption and battery State of Charge (SOC) are then computed for the whole daily service operations. The simulation results are validated with the real data available and several scenarios are then considered within the simulations. Based on the results obtained, further improvements are proposed and discussed aiming to optimize the utilization of bus fleet, regarding both vehicles scheduling and new charging solutions.

Learning-Based Nonlinear Model Predictive Controller for Hydraulic Cylinder Control of Ship Steering System

The steering mechanism of ship steering gear is generally driven by a hydraulic system. The precise control of the hydraulic cylinder in the steering mechanism can be achieved by the target rudder angle. However, hydraulic systems are often described as nonlinear systems with uncertainties. Since the system parameters are uncertain and system performances are influenced by disturbances and noises, the robustness cannot be satisfied by approximating the nonlinear theory by a linear theory. In this paper, a learning-based model predictive controller (LB-MPC) is designed for the position control of an electro-hydraulic cylinder system. In order to reduce the influence of uncertainty of the hydraulic system caused by the model mismatch, the Gaussian process (GP) is adopted, and also the real-time input and output data are used to improve the model. A comparative simulation of GP-MPC and MPC is performed assuming that the interference and uncertainty terms are bounded. Consequently, the proposed control strategy can effectively improve the piston position quickly and precisely with multiple constraint conditions.

PID-based CNN-LSTM for accuracy-boosted virtual sensor in battery thermal management system

Battery thermal management is essential to achieve good performance and a long battery system lifespan in electric vehicles and stationary applications. Such a thermal management system is dependent on temperature monitoring, which is frequently hampered by the limited sensor measurements. The virtual sensor is brought forward to overcome this physical restriction and provide broader access to the battery’s temperature distribution. Through leveraging the combined convolutional neural network (CNN) and long short-term memory (LSTM) networks to extract both spatial and temporal information from the data, this paper proposes a novel virtual sensing platform. A PID compensator is included to offer auxiliary correction to the inputs and drive the prediction error to zero over time in a feedback loop. Off-line and online modes of this CNN-LSTM virtual sensor are considered. The network, which is trained off-line, will work with the PID compensator in the online mode with real-time sensor data. With the PID-based accuracy-boosted virtual sensor, the performance of the trained CNN-LSTM prediction on real-time data inputs is improved. Besides, this PID compensator reduces the number of hyper-parameters to be tuned. Based on control theory, the design of PID and its analysis are presented as well. With generated battery thermal data, numerical simulations show that the CNN-LSTM-PID virtual sensing framework can achieve the real-time prediction error reduction rate to 35.52% on average with 18.78% less online calculation time.