Virtual Sensor Research Articles

Tool Fracture Detection in Electromechanical Broaching Through Machine Sensor

Broaching is a machining technique that plays a crucial role in the quality and integrity of manufactured products as it demands superior surface quality and precision compared to other machining methods. Monitoring tool condition is essential, as a damaged tool can significantly affect the quality of the surface finish and integrity of the final part causing irreparable damage to both the part and the work tool.This paper presents the development of a virtual sensor designed for the real-time detection of tool fractures during the broaching process, allowing timely decision-making without generating greater damage to the part or tool. The virtual sensor is created using signals collected from the broaching machine's data, without the need for external sensors; having the potential to improve product quality, increase process efficiency, and reduce material waste as well as production costs by detecting tool fractures early and, therefore, increasing the overall sustainability of the process.

A Survey on Graph Neural Networks for Time Series: Forecasting, Classification, Imputation, and Anomaly Detection.

Time series are the primary data type used to record dynamic system measurements and generated in great volume by both physical sensors and online processes (virtual sensors). Time series analytics is therefore crucial to unlocking the wealth of information implicit in available data. With the recent advancements in graph neural networks (GNNs), there has been a surge in GNN-based approaches for time series analysis. These approaches can explicitly model inter-temporal and inter-variable relationships, which traditional and other deep neural network-based methods struggle to do. In this survey, we provide a comprehensive review of graph neural networks for time series analysis (GNN4TS), encompassing four fundamental dimensions: forecasting, classification, anomaly detection, and imputation. Our aim is to guide designers and practitioners to understand, build applications, and advance research of GNN4TS. At first, we provide a comprehensive task-oriented taxonomy of GNN4TS. Then, we present and discuss representative research works and introduce mainstream applications of GNN4TS. A comprehensive discussion of potential future research directions completes the survey. This survey, for the first time, brings together a vast array of knowledge on GNN-based time series research, highlighting foundations, practical applications, and opportunities of graph neural networks for time series analysis.

Exploring the impact of the process parameters on the thermal treatment of viscous food fluids in a tube-in-tube heat exchanger

Thermal treatments are essential to abate the microbial load in food matrices and ensure food safety. Heat exchangers (HEs) are commonly adopted for the pasteurization of liquid foods, thanks to their simple structure and ease of use. In particular, tubular HEs are often employed with viscous fluids and products containing fibres and pieces. The aim of pasteurization is to ensure that every point of the product reaches and maintains the target temperature for an adequate amount of time. To this end, it is essential to guarantee uniformity and prevent portions of fluid from reaching an insufficient temperature, and others from being overheated. To evaluate the thermal treatment inside a tube-in-tube HE, computational fluid dynamics (CFD) was used to simulate a pilot plant. Three different fluids with shear-thinning temperature-dependent non-Newtonian behaviour were evaluated. A CFD simulation campaign was carried out reproducing, for each juice, the thermal treatment under different operating conditions. Then, the results were analysed with statistical techniques to evaluate the impact of each operating parameter on the treatment, highlighting how high viscosity greatly impacts the thermal treatment, resulting in lower average temperature and temperature uniformity, while causing an increase in the pressure drop. The results of the simulation campaign were also used to generate a reduced order model (ROM) of the device, that could be exploited to estimate the outcomes of a treatment both as averaged values on a section or at a specific point like “virtual sensors”. The ROM could be also adopted to design and evaluate alternative treatments by performing numerical optimization based on given objectives and constraints. Finally, as the ROM can be implemented in real-time, it could be used to monitor and control the functioning of food plants, detecting deviations from the expected conditions and possible failure conditions.

Planning the Greenhouse Climatic Mapping Using an Agricultural Robot and Recurrent-Neural-Network-Based Virtual Sensors

Planning the Greenhouse Climatic Mapping Using an Agricultural Robot and Recurrent-Neural-Network-Based Virtual Sensors

Virtual Sensing in Steel Bridges: Time Series Deep Learning for Stress Prediction

This study introduces an innovative approach for predicting stress responses in steel bridges, specifically focusing on a railway bridge in Vänersborg, Sweden. Four deep learning models have been evaluated: Multilayer Perceptron (MLP), Long Short-Term Memory (LSTM), Temporal Convolutional Network (TCN), and a hybrid LSTM-TCN. Training on stress history data from a multiscale Finite Element (FE) model and a validation with real-world data from a bridge monitoring system revealed high prediction accuracy near sensor locations, surpassing an R-squared score of 0.9, comparable to the polynomial local response function method.The comparative analysis provides critical insights into the great potential of deep learning-based sequence models for identifying intricate, temporally dependent stress patterns across the bridge, including predictions at points distant from direct sensor measurements. These models demonstrate a notable capability for capturing highly non-linear relationships between stress histories. While sequence models (LSTM, TCN, and hybrid LSTM-TCN) tended to provide conservative estimates impacting fatigue life predictions, the MLP model occasionally underestimated critical stress cycles.This research emphasizes the potential of deep learning techniques for time series to enhance bridge monitoring systems, improve virtual sensing, and enable real-time monitoring capabilities. Our proposed methodology provides a comprehensive understanding of stress data in steel bridges, which is crucial for ensuring their maintenance and safety.

Enhancing Battery Thermal Management with Virtual Temperature Sensor Using Hybrid CNN-LSTM

Enhancing Battery Thermal Management with Virtual Temperature Sensor Using Hybrid CNN-LSTM

CROMOSim: A Deep Learning-Based Cross-Modality Inertial Measurement Simulator

With the prevalence of wearable devices, inertial measurement unit (IMU) data has been utilized in monitoring and assessing human mobility such as human activity recognition (HAR) and human pose estimation (HPE). Training deep neural network (DNN) models for these tasks require a large amount of labelled data, which are hard to acquire in uncontrolled environments. To mitigate the data scarcity problem, we design CROMOSim, a cross-modality sensor simulator that simulates high fidelity virtual IMU sensor data from motion capture systems or monocular RGB cameras. It utilizes a skinned multi-person linear model (SMPL) for 3D body pose and shape representations to enable simulation from arbitrary on-body positions. Then a DNN model is trained to learn the functional mapping from imperfect trajectory estimations in a 3D SMPL body tri-mesh due to measurement noise, calibration errors, occlusion and other modelling artifacts, to IMU data. We evaluate the fidelity of CROMOSim simulated data and its utility in data augmentation on various HAR and HPE datasets. Extensive empirical results show that the proposed model achieves a 6.7% improvement over baseline methods in a HAR task.

Enhanced Nonlinear Current State Observer Based Virtual Current Sensors for Permanent Magnet Synchronous Motor Drives

Enhanced Nonlinear Current State Observer Based Virtual Current Sensors for Permanent Magnet Synchronous Motor Drives

A method for implementing a ship navigation simulator for the generation and utilization of virtual data

Active research is underway to utilize various sensors for situational awareness, but validating these studies requires real sensor data, which is challenging to acquire due to the time and space constraints inherent in ships. Additionally, the diverse formats of sensor data and the need for postprocessing pose further hurdles. By developing a ship navigation simulator that faithfully replicates real-world environments, these limitations can be overcome, and ample virtual sensor data can be generated through immersive simulations. In this study, we propose a method for implementing such a simulator. A life-sized terrain model was placed within a virtual reality environment, representing maritime environments, including port structures. To simulate ship motions according to maritime weather conditions, we applied various laws of physics within virtual reality. Finally, to provide validation data for critical navigation technologies, such as collision avoidance and ship detection, we generated a virtual automatic identification system and camera data. We implemented and operated a ship navigation simulator using this approach, and simulations conducted across various environments generated real-time sensor data efficiently, affirming the effectiveness of the proposed method.

USING Sensor: User-Smartphone Decoupling Detection as a Mobile-Based Virtual Sensor

USING Sensor: User-Smartphone Decoupling Detection as a Mobile-Based Virtual Sensor

Toward safer Human-Robot collaboration in MR environment

This paper presents a Mixed Reality (MR) approach to extend the Safe Move tool from ABB Robot Studio (RS) to view and intuitively edit the safety configuration, allowing the import of existing configurations and the export of modified ones to RS for certification by the expert. The added virtual sensor can detect the user's position allowing collision detection and avoidance. The robot's motion is not only adjusted to the safety zone in which the robot is but also to its relative position to the operator by monitoring in real-time positions of the operator's hands as well as head and triggering appropriate action in the robot.

Development and Evaluation of Artificial Neural Networks for Real-World Data-Driven Virtual Sensors in Vehicle Suspension

Development and Evaluation of Artificial Neural Networks for Real-World Data-Driven Virtual Sensors in Vehicle Suspension

Design and simulation of a smart energy metering system

Power theft has been one of the main problems facing the power industry in Nigeria which had led to huge economic and infrastructural loss to the government and utility companies. Hence the need to proffer solution to this challenge, this research identified the various method of stealing energy which is meter by passing, connecting directly to the supply cable just before it gets to the meter terminal, resetting of meter reading, and deliberate non-payment of electricity bills, vandalism and unfaithful act of the utility employees. The causes and effect of energy theft where also identified to be customer taking electricity as part of social service from the government and also the inability of the utility companies to meter their customers. In finding solution to this problem, this thesis adopted the simulation of anti-tamper sensor (accelerometers, light, tilt, magnetic) using Proteus to create a virtual sensor model that can mimic the behavior of the physical sensors, SMS algorithm was programmed in such that it assumes the micro-controller (ADUINO) and a GSM module from the Proteus library was used, the connection was done using the appropriate serial communication pins. The result obtained shown that smart metering was successfully designed and simulation model accurately replicated the behavior of the energy meter, including measurements of voltage, current, and power consumption. The tampering energy meter seal using highly sensitive sensors to mitigate against energy theft by provide an additional layer of security against meter tampering with incorporated of sensor output information that can sent via by GSM Module to the cloud, which help to mitigate against meter tampering. The complete system software implementation and validation was done using Proteus and Arduino.

CFD Analysis of Sloshing in the Fuel Tank of a Heavy Vehicle with Emergency Braking System

Liquid sloshing, which occurs in all accelerating and liquid-carrying vehicles, is of great importance, especially in the automotive and aerospace industries. Large-scale fluid sloshing causes both operational and safety problems in vehicles. In this study, the fuel tank of a heavy vehicle with an emergency braking system is designed in three dimensions, and liquid sloshing in the fuel tank is investigated by CFD analysis meth-od. VOF solution method, k-ԑ turbulence model, and PISO solution algorithm are used in the study. In the analysis of liquid sloshing, it is assumed that the vehicle is traveling at a certain speed, decelerates and stops with emergency braking, and remains station-ary for a while. The braking scenario and boundary conditions are based on test data from a heavy vehicle manufacturer. The designed fuel tank with a capacity of 207.6 li-ters was analyzed at 25%, 50%, and 60% diesel fuel filling levels in 6 different cases with and without anti-slosh baffles. Four virtual sensors were placed on the side wall of the fuel tank in the direction of vehicle movement, and time-dependent pressure changes were analyzed for all cases. In addition, the fuel volume ratio in all cases is visualized and presented for specific time steps. With the use of anti-slosh baffles, the maximum pressure, the rate of pressure increase, and the liquid sloshing were reduced by a factor of 2-3 for different cases. With the design of the fuel tank using anti-slosh baffles, instantaneous interruptions in the fuel system are prevented. Reducing the im-pact pressures on the tank walls is expected to positively affect noise, vibration, and stability problems.

LSTM-Based Virtual Load Sensor for Heavy-Duty Vehicles.

In this paper, a special recurrent neural network (RNN) called Long Short-Term Memory (LSTM) is used to design a virtual load sensor that estimates the mass of heavy vehicles. The estimation algorithm consists of a two-layer LSTM network. The network estimates vehicle mass based on vehicle speed, longitudinal acceleration, engine speed, engine torque, and accelerator pedal position. The network is trained and tested with a data set collected in a high-fidelity simulation environment called Truckmaker. The training data are generated in acceleration maneuvers across a range of speeds, while the test data are obtained by simulating the vehicle in the Worldwide harmonized Light vehicles Test Cycle (WLTC). Preliminary results show that, with the proposed approach, heavy-vehicle mass can be estimated as accurately as commercial load sensors across a range of load mass as wide as four tons.

Analysis of Sensors in VR Equipment to 3D printer in IoT System

Virtual reality (VR) technology, artificial intelligence (AI), sensors and three-dimensional (3D) printing technology are all advanced and promising technologies. And these technologies are changing people’s life and social development. Besides, Internet of Things (IoT) is a product of the development of modern communications, which is gradually applied to real life. However, there is no current system including VR, AI, sensors, and 3D printing at the same time. Therefore, this article presents an IoT system to integrate VR, AI, sensor, and 3D printer. By using VR equipment to implement human-computer interaction, using AI technology to assist users with operation and optimize 3D printing, using sensors to add functions and using 3D printer to obtain real-time manufacturing, this IoT system can simplify and optimize users’ manipulation and it improves 3D printing technology. Because of the support of existing technology, this article explores the feasibility of IoT system and finds that it is a realistic system. This IoT system has positive effects on many industries like education, architectural design, and advanced manufacturing.

Modelling the upwelling radiance detected in a seawater column for oil-in-water emulsion tracking

This study concerns changes in the solar light transfer in seawater as a result of the appearance of oil substances in the form of oil-in-water emulsions. The expected effect of the studies is to gather knowledge that would be useful in designing an optical sensor for monitoring oil substances penetrating the seawater column. The paper presents the process of the Monte Carlo modelling of the upwelling radiance detected by a virtual underwater sensor. Moreover, this article discusses the predicted difference between the intensity of the signal from the upwelling radiance meter and the signal from the downwelling radiance meter. The modelling results suggest that in order to obtain a strong signal indicating the presence of oil substances, the radiance meter should receive light running upwards in the sea depth.

Deep long short-term memory neural networks as virtual sensors for marine diesel engine NOx prediction at transient conditions

Virtual Sensors based on deep-learning models for predicting the NOx emissions of a Diesel Engine under transient conditions were developed and verified. Raw data from laboratory experimental measurements, under marine transient loading cycles, were used for training and evaluation of the developed models. NOx prediction under transient conditions is often inaccurate by implementing conventional methods since they fail to capture the dynamic behavior of internal combustion engines. The proposed model is based on Long Short-Term Memory (LSTM) Networks. A Deep Feed-forward Neural Network (DFNN) was also developed to validate the LSTM. The LSTM input is a time sequence of past measurements of the inputs while the DFNN only uses the most recent measurements. The Bayesian Hyberband Optimization (BOHB) algorithm determined the structure and parameters of each network. Each model uses the same inputs and is directly derived from the engine ECU. The LSTM validation showed that the model can generalize and accurately predict the NOx emissions under transient loading compared to the DFNN.

An AutomationML extension towards interoperability of 3D virtual commissioning software applications

ABSTRACT To achieve interoperability between different 3D virtual commissioning software, a generic virtual commissioning data model is required. AutomationML is a standard neutral format for interoperability in the engineering phase. However, the current AutomationML standard is not sufficient for full-scope 3D-based virtual commissioning data exchange, as attributes and modeling method of 3D virtual commissioning-related sensors, actuators and signal connections are not standardized in AutomationML. To fill this gap, the authors suggest extending AutomationML for interoperability in 3D virtual commissioning. In this paper, a case-driven iterative approach is introduced to evolve towards an AutomationML extension. This extension is gradually developed by taking the union of all virtual commissioning-related functions and attributes of 3D virtual commissioning software. During the iteration, naming rules are applied when a new attribute is added to the extension. With this approach, an initial AutomationML extension is created by implementing a first iteration. The interoperability performance of this extension is subsequently evaluated by conducting data exchange of a representative set of 3D emulation models between two 3D virtual commissioning software, namely Siemens NX and Visual Components, via self-developed ‘Import’ and ‘Export’ plug-ins. It shows that AutomationML extension-based data exchange converts 70% more attributes than that only based on AutomationML.

Fault detection and diagnosis for variable refrigerant flow systems by using virtual sensors and deep learning

The primary objective of this work was to implement and evaluate a fault detection and diagnostic system for variable refrigerant flow (VRF) systems with multiple indoor units. To achieve this goal, virtual sensors and deep learning models are developed to predict the system performance whose direct measurement could be expensive or difficult. The primary challenges in implementing a diagnostic and performance monitoring system are the high initial costs of additional physical sensors. However, the combination of virtual sensor and the deep learning model can identify and isolate a condenser fouling fault. The virtual sensors were used to determine the condenser air flow rate in and the performance of a VRF system equipped with a variable-speed compressor and fans. An important step in automated fault detection and diagnostics is evaluating whether a condenser fouling fault is adequately significant to warrant service based on various criteria, including the potential to degrade comfort, increase in energy usage that can affect the overall system performance operation, and operating conditions that can shorten VRF system life. The deep learning model is developed to determine the performance degradation ratio between the performance estimated using the outputs of virtual sensors and that estimated using deep learning-based expected performance models. The fault impact models can help determine whether a fault is sufficiently serious to warrant a service recommendation.