Your Selections

Neural networks
Show Only

Collections

File Formats

Content Types

Dates

Sectors

Topics

Authors

Publishers

Affiliations

Committees

Events

Magazine

   This content is not included in your SAE MOBILUS subscription, or you are not logged in.

A Machine Learning based Multi-objective Multidisciplinary Design Optimization (MMDO) for Lightweighting the Automotive Structures

Mahindra and Mahindra, Ltd.-Ranga Srinivas Gunti
  • Technical Paper
  • 2019-28-2424
To be published on 2019-11-21 by SAE International in United States
The present work involves Machine Learning (ML) based Multi-objective Multidisciplinary Design Optimization (MMDO) for lightweighting the automotive structures. The challenge in deployment of MMDO algorithms in solving real-world automotive structural design problems is the enormous time involved in solving full vehicle finite element models that involve large number of design variables and multiple performance constraints pertaining to vehicle dynamics, durability, crash and NVH domains. With the availability of powerful workstations and using the advanced Computer Aided Engineering (CAE) tools, it has become possible to generate huge sets of simulation data pertaining to multiple domains. In the present work, lightweigting of the vehicle structure is achieved, considered the vehicular hardpoint locations and the gages of the vehicle structures as the design variables and performance parameters pertaining to vehicle dynamics, structural durability, front-end intrusions during an IIHS offset impact test and the modal frequencies of few critical structural members as the constraint variables. Artificial Neural Networks (ANN) based algorithms were used for developing the predictive models of various performance parameters. The predictive models were then used to…
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.

Digital Twins for Prognostic Profiling

Altair Engineering India Pvt Ltd.-Painuri Thukaram, Sreeram Mohan
  • Technical Paper
  • 2019-28-2456
To be published on 2019-11-21 by SAE International in United States
Digital Twins for Prognostic Profiling Authors: Sreeram Mohan*, Painuri Thukaram**, Panduranga Rao*** Objective / Question: Ability to have least failures in products on the field with minimum effort from the manufacturers is a major area of focus driven by Industry 4.0 initiatives. Amidst traditional methods of performing system / subsystem level tests often does not enable the complete coverage of a machine health performance predictions. This paper highlights a workable workflow that could be used as a template while considering system design especially employing Digital Twins that help in mimicking real-life scenarios early in the design cycle to increase product’s reliability as well as tend to near zero defects. Methodology: With currently available disruptive technologies , systems are integrated multi-domain 'mechatronics' systems operating in closed-loop/close-interaction. This poses great challenges to system health monitoring as failure of any component can trigger catastrophic system failures. It may be the reason that component failures, as per some aerospace reports, are found to be major contributing factors to aircraft loss-of-control. Essentially, it is either too expensive or impossible to…
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.

APPLICATION ORIENTED HYBRIDIZED DYNAMIC MODELS OF POWERTRAIN CONTROL FOR CONNECTED VEHICLES – A CASE STUDY ON TURBOCHARGER CONTROL

Continental Germany-Richard Kopold
Continental India-Vivek Venkobarao
  • Technical Paper
  • 2019-28-2443
To be published on 2019-11-21 by SAE International in United States
In a connected vehicle environment, the engine drive cycles operate in synchronized and regulated manner. This requires smooth transitions for improved CO_2 footprint. To arrive at this, there is need for intelligent and faster airpath control at transients. Authors aim to model and control every actuator of a coupled system in a synchronized manner with faster dynamic response. The turbocharger control is vital and forms heart of the system; This demands accurate position prediction of VTG. Deriving a control law for turbocharger is challenging due to the hybridized nature of turbocharger models in engine management system. It becomes extremely critical to estimate accurately, the position of VTG without introduction of any sensing devices. The control engineer always need to solve the trade-off between the controller performance KPI’s – rise time, transient response, controllability, observability and capability – stability and dynamics response etc. Author propose a model which improve the performance and capability of VTG control. Author presents a novel technique to model VTG position. A neural network based supervised learning model is derived. The model…
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.
new

Study on Robust Motion Planning Method for Automatic Parking Assist System Based on Neural Network and Tree Search

Tongji University-Fengwei Hu, Hui Chen, Jiren Zhang
Published 2019-11-04 by SAE International in United States
Automatic Parking Assist System (APAS) is an important part of Advanced Driver Assistance System (ADAS). It frees drivers from the burden of maneuvering a vehicle into a narrow parking space. This paper deals with the motion planning, a key issue of APAS, for vehicles in automatic parking. Planning module should guarantee the robustness to various initial postures and ensure that the vehicle is parked symmetrically in the center of the parking slot. However, current planning methods can’t meet both requirements well. To meet the aforementioned requirements, a method combining neural network and Monte-Carlo Tree Search (MCTS) is adopted in this work. From a driver’s perspective, different initial postures imply different parking strategies. In order to achieve the robustness to diverse initial postures, a natural idea is to train a model that can learn various strategies. As artificial neural network has outstanding potential in representing and learning knowledge, a neural network is utilized to provide prior knowledge, which is trained through supervised learning by a novel method that imitates human learning style. However, the training accuracy…
This content contains downloadable datasets
Annotation ability available
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.
new

Assessing the Combined Outcome of Rice Husk Nano Additive and Water Injection Method on the Performance, Emission and Combustion Characters of the Low Viscous Pine Oil in a Diesel Engine

Anna University Chennai-Mebin Samuel P, Devaradjane Gobalakichenin
University College of Engineering Villupuram-Gnanamoorthi V
Published 2019-10-22 by SAE International in United States
The research work intends to assess the need and improvement by using a low viscous bio oil, RH (rice husk) nano particles and water injection method in enhancing the performance, emission and combustion characters of a diesel engine. One of the major setbacks for using biodiesel is its higher viscosity. Hence, a low viscous oil (pine oil) which does not need transesterification process was used as a biofuel in this study. Further, to improve its characteristics a non-metallic nano additive produced from rice husk was added at 3 proportions (50, 100, 200 ppm) and the optimal quantity was found as 100 ppm based on the BTE (brake thermal efficiency) value of 30.2% at peak load condition. This efficiency value was accompanied by a considerable decrease in pollutants like HC (hydrocarbon)-34.8%, Smoke-31.6%, CO (carbon monoxide)-43.7%. On the contrary, NOx (oxides of nitrogen) emission was found to be increased for all load values. At peak load, when compared with diesel, pine oil with RH has 19.3% increased NOx emission. To reduce this increased NOx emission, water was…
This content contains downloadable datasets
Annotation ability available
new

Study of Advanced Control Based on the RBF Neural Network Theory in Diesel Engine Speed Control

SAE International Journal of Engines

China-Li-Ping Yang
Harbin Engineering University, China-Guo-Feng Zhao, Yun Long, En-Zhe Song, Xiu-Zhen Ma
  • Journal Article
  • 03-13-01-0005
Published 2019-10-14 by SAE International in United States
Based on radial basis function (RBF) neural network (NN) theory, RBF-Proportional Integral Derivative (PID) diesel engine speed control is proposed. The algorithm has strong self-learning ability and strong adaptive ability, and is able to optimize the control parameters of the speed loop controller in real time. A series of simulations are carried out with different initial weights. Simulation results reveal that initial weights have little effect on RBF-PID control performance. A STM32 MCU-based controller is developed according to the calculation requirement. Experiments are carried out on a D6114 diesel engine generator to verify the proposed speed control algorithm. The simulation results are in good agreement with the experimental results. The results show that the influence of initial weights on RBF-PID control algorithm is smaller than that on BP-PID control algorithm. When RBF-PID control algorithm is adopted, the steady speed fluctuation rate is 0.4%. When sudden load is carried out, the speed recovery time is 2.1 s and the instantaneous adjustment rate is 4.93%. When sudden unload simulation is carried out, the speed recovery time is…
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.
new

Tool Condition Monitoring in Face Milling Process Using Decision Tree and Statistical Features of Vibration Signal

B S Abdur Rahman Crescent Institute of Science & Technology-Pradeep Kumar Durairaj, Muralidharan Vaithiyanathan
Published 2019-10-11 by SAE International in United States
In milling process, the quality of the machined component is highly influenced by the condition of the tool. Hence, monitoring the condition of the tool becomes essential. A suitable mechanism needs to be devised in order to monitor the condition of the tool. To achieve this, condition monitoring of milling tool is taken up for the study. In this work, the condition of the tool is classified as good tool and tool with common faults in face milling process such as flank wear, worn out and breakage of the tool based on machine learning approach using statistical feature and decision tree technique. Vibration signals of the milling tool are obtained during machining of mild steel. Statistical features are extracted from the obtained signal, in which the important features are selected using decision tree. The selected features are given as the input to the same algorithm. The output of the algorithm is utilized for classifying the different conditions of the tool. The experimental results show that the accuracy of decision tree technique is at the acceptable…
This content contains downloadable datasets
Annotation ability available
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.
new

Fault Detection in Single Stage Helical Planetary Gearbox Using Artificial Neural Networks (ANN) and Decision Tree with Histogram Features

BSACIST-Syed Shaul Hameed, Muralidharan Vaithiyanathan, Mahendran Kesavan
Published 2019-10-11 by SAE International in United States
Drive train failures are most common in wind turbines. Lots of effort has been made to improve the reliability of the gearbox but the truth is that these efforts do not provide a lifetime solution. Majority of failures are caused by bearing and gearbox. It also states that wind turbine gearbox failure causes the highest downtime as the repair has to be done at Original Equipment Manufacturer [OEM]. This work aims to predict the failures in planetary gearbox using fault diagnosis technique and machine learning algorithms. In the proposed method the failing parts of the planetary gearbox are monitored with the help of accelerometer sensor mounted on the planetary gearbox casing which will record the vibrations. A prototype has been fabricated as a miniature of single stage planetary gearbox. The vibrations of the healthy gearbox, sun defect, planet defect and ring defect under loaded conditions are obtained. The signals show the performance characteristics of the gearbox condition. These characteristics and their number of occurrences were plotted in a histogram graph. Predominant statistical features which represent…
This content contains downloadable datasets
Annotation ability available
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.
new

Neural Network Based Virtual Sensor for Throttle Valve Position Estimation in a SI Engine

SITAMS-Chellappan Kavitha
VIT University-Bragadeshwaran Ashok, Sathiaseelan Denis Ashok, Chidambaram Ramesh Kumar
Published 2019-10-11 by SAE International in United States
Electronic throttle body (ETB) is commonly employed in an intake manifold of a spark ignition engine to vary the airflow quantity by adjusting the throttle valve in it. The actual position of the throttle valve is measured by means of a dual throttle position sensor (TPS) and the signal is feedback into the control unit for accomplishing the closed loop control in order handle the nonlinearities due to friction, limp-home position, aging, parameter variations. This work aims presents a neural networks based novel virtual sensor for the estimation of throttle valve position in the electronic throttle body. Proposed neural network model estimates the actual throttle position using three inputs such as reference throttle angle, angular error and the motor current. In the present work, the dynamic model of the electronic throttle body is used to calculate the current consumed by the motor for corresponding throttle valve movement. Proposed virtual sensor is tested for the sinusoidal and random driving cycle throttle angle input using a Bosch DVE5 electronic throttle body. Estimated throttle valve angle using the…
This content contains downloadable datasets
Annotation ability available
new

Technology Reveals How AI Makes Decisions

  • Magazine Article
  • TBMG-35339
Published 2019-10-01 by Tech Briefs Media Group in United States

Researchers have developed a technology that reveals the criteria AI systems use when making decisions. The innovative Spectral Relevance Analysis (SpRAy) method based on Layer-wise Relevance Propagation technology provides a first peek inside the “black box”.