An improved compact propulsion system model based on batch normalize deep neural network

2021 ◽  
Vol 0 (0) ◽  
Author(s):  
Juan Fang ◽  
Qiangang Zheng ◽  
Haibo Zhang ◽  
Chongwen Jin
Keyword(s):  
Neural Network ◽  
Steady State ◽  
Real Time ◽  
Data Storage ◽  
High Accuracy ◽  
Propulsion System ◽  
System Model ◽  
Support Vector ◽  
Baseline Model ◽  
Model Based

Abstract Aero-engine on-board steady state model is an important part of many advanced engine control algorithms. In order to build a high accuracy and real-time steady-state onboard model in a large envelope, an ICPSM (improved compact propulsion system model) based on batch normalize neural network is proposed in this paper. Compared with piecewise linearization model and support vector machine model, conventional CPSM which is mainly composed of baseline model and nonlinear sub model has the advantages of high real-time performance and small data storage. However, as the similarity conversion error increases with the distance from the design point, the cumulative error of the conventional baseline model also increases, which makes the model unable to maintain high accuracy in the full envelope. Thus, a high accuracy baseline model in full envelope based on batch normalize neural network is proposed in this paper. The simulation result shows that compared with the conventional compact propulsion system model, the percentage error of parameters of the improved compact propulsion system model based on the batch neural network is reduced by two times, the single step operation time is reduced by 18%, and the data storage of the onboard model is reduced as well.

scholarly journals Augmented Reality Maintenance Assistant Using YOLOv5

2021 ◽  
Vol 11 (11) ◽  
pp. 4758
Author(s):  
Ana Malta ◽  
Mateus Mendes ◽  
Torres Farinha
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Object Recognition ◽  
Augmented Reality ◽  
Real Time ◽  
Recognition System ◽  
High Accuracy ◽  
Video Streams ◽  
The Neural Network ◽  
Deep Learning Neural Network

Maintenance professionals and other technical staff regularly need to learn to identify new parts in car engines and other equipment. The present work proposes a model of a task assistant based on a deep learning neural network. A YOLOv5 network is used for recognizing some of the constituent parts of an automobile. A dataset of car engine images was created and eight car parts were marked in the images. Then, the neural network was trained to detect each part. The results show that YOLOv5s is able to successfully detect the parts in real time video streams, with high accuracy, thus being useful as an aid to train professionals learning to deal with new equipment using augmented reality. The architecture of an object recognition system using augmented reality glasses is also designed.

Neural Network Based Control of Preform Permeation in Resin Transfer Molding Processes With Real-Time Permeability Estimation

2000 ◽  
Author(s):  
David Nielsen ◽  
Ranga Pitchumani
Keyword(s):  
Neural Network ◽  
Real Time ◽  
Resin Transfer Molding ◽  
Local Pressure ◽  
Transfer Molding ◽  
Model Based ◽  
Rtm Process ◽  
Virtual Sensing ◽  
Intelligent Model ◽  
Injection Pressures

Abstract Variabilities in the preform structure in situ in the mold are an acknowledged challenge to effective permeation control in the Resin Transfer Molding (RTM) process. An intelligent model-based controller is developed which utilizes real-time virtual sensing of the permeability to derive optimal decisions on controlling the injection pressures at the mold inlet ports so as to track a desired flowfront progression during resin permeation. This model-based optimal controller employs a neural network-based predictor that models the flowfront progression, and a simulated annealing-based optimizer that optimizes the injection pressures used during actual control. Preform permeability is virtually sensed in real-time, based on the flowfront velocities and local pressure gradient estimations along the flowfront. Results are presented which illustrate the ability of the controller in accurately steering the flowfront for various fill scenarios and preform geometries.

scholarly journals Random Forest Regressor-Based Approach for Detecting Fault Location and Duration in Power Systems

Sensors ◽  
2022 ◽  
Vol 22 (2) ◽  
pp. 458
Author(s):  
Zakaria El Mrabet ◽  
Niroop Sugunaraj ◽  
Prakash Ranganathan ◽  
Shrirang Abhyankar
Keyword(s):  
Neural Network ◽  
Random Forest ◽  
Power Systems ◽  
Real Time ◽  
Fault Location ◽  
State Of The Art ◽  
Economic Consequences ◽  
Support Vector ◽  
Detection Accuracy ◽  
Data Driven Approach

Power system failures or outages due to short-circuits or “faults” can result in long service interruptions leading to significant socio-economic consequences. It is critical for electrical utilities to quickly ascertain fault characteristics, including location, type, and duration, to reduce the service time of an outage. Existing fault detection mechanisms (relays and digital fault recorders) are slow to communicate the fault characteristics upstream to the substations and control centers for action to be taken quickly. Fortunately, due to availability of high-resolution phasor measurement units (PMUs), more event-driven solutions can be captured in real time. In this paper, we propose a data-driven approach for determining fault characteristics using samples of fault trajectories. A random forest regressor (RFR)-based model is used to detect real-time fault location and its duration simultaneously. This model is based on combining multiple uncorrelated trees with state-of-the-art boosting and aggregating techniques in order to obtain robust generalizations and greater accuracy without overfitting or underfitting. Four cases were studied to evaluate the performance of RFR: 1. Detecting fault location (case 1), 2. Predicting fault duration (case 2), 3. Handling missing data (case 3), and 4. Identifying fault location and length in a real-time streaming environment (case 4). A comparative analysis was conducted between the RFR algorithm and state-of-the-art models, including deep neural network, Hoeffding tree, neural network, support vector machine, decision tree, naive Bayesian, and K-nearest neighborhood. Experiments revealed that RFR consistently outperformed the other models in detection accuracy, prediction error, and processing time.

scholarly journals Deep Learning-Based Human Activity Real-Time Recognition for Pedestrian Navigation

Sensors ◽  
2020 ◽  
Vol 20 (9) ◽  
pp. 2574 ◽  
Author(s):  
Junhua Ye ◽  
Xin Li ◽  
Xiangdong Zhang ◽  
Qin Zhang ◽  
Wu Chen
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Real Time ◽  
Learning Algorithms ◽  
Support Vector ◽  
Classification Methods ◽  
Pedestrian Navigation ◽  
Mode Recognition ◽  
Posture Recognition ◽  
Motion Mode

Several pedestrian navigation solutions have been proposed to date, and most of them are based on smartphones. Real-time recognition of pedestrian mode and smartphone posture is a key issue in navigation. Traditional ML (Machine Learning) classification methods have drawbacks, such as insufficient recognition accuracy and poor timing. This paper presents a real-time recognition scheme for comprehensive human activities, and this scheme combines deep learning algorithms and MEMS (Micro-Electro-Mechanical System) sensors’ measurements. In this study, we performed four main experiments, namely pedestrian motion mode recognition, smartphone posture recognition, real-time comprehensive pedestrian activity recognition, and pedestrian navigation. In the procedure of recognition, we designed and trained deep learning models using LSTM (Long Short-Term Memory) and CNN (Convolutional Neural Network) networks based on Tensorflow framework. The accuracy of traditional ML classification methods was also used for comparison. Test results show that the accuracy of motion mode recognition was improved from 89.9 % , which was the highest accuracy and obtained by SVM (Support Vector Machine), to 90.74 % (LSTM) and 91.92 % (CNN); the accuracy of smartphone posture recognition was improved from 81.60 % , which is the highest accuracy and obtained by NN (Neural Network), to 93.69 % (LSTM) and 95.55 % (CNN). We give a model transformation procedure based on the trained CNN network model, and then obtain the converted . t f l i t e model, which can be run in Android devices for real-time recognition. Real-time recognition experiments were performed in multiple scenes, a recognition model trained by the CNN network was deployed in a Huawei Mate20 smartphone, and the five most used pedestrian activities were designed and verified. The overall accuracy was up to 89.39 % . Overall, the improvement of recognition capability based on deep learning algorithms was significant. Therefore, the solution was helpful to recognize comprehensive pedestrian activities during navigation. On the basis of the trained model, a navigation test was performed; mean bias was reduced by more than 1.1 m. Accordingly, the positioning accuracy was improved obviously, which is meaningful to apply DL in the area of pedestrian navigation to make improvements.

Research on Modeling of Propeller in a Turboprop Engine

2015 ◽  
Vol 32 (2) ◽  
Author(s):  
Jiaqin Huang ◽  
Xianghua Huang ◽  
Tianhong Zhang
Keyword(s):  
Approximate Method ◽  
Real Time ◽  
Integrated Control ◽  
Computing Time ◽  
High Accuracy ◽  
Surface Theory ◽  
Model Based ◽  
Lifting Surface ◽  
Strip Theory ◽  
Lift And Drag

AbstractIn the simulation of engine-propeller integrated control system for a turboprop aircraft, a real-time propeller model with high-accuracy is required. A study is conducted to compare the real-time and precision performance of propeller models based on strip theory and lifting surface theory. The emphasis in modeling by strip theory is focused on three points as follows: First, FLUENT is adopted to calculate the lift and drag coefficients of the propeller. Next, a method to calculate the induced velocity which occurs in the ground rig test is presented. Finally, an approximate method is proposed to obtain the downwash angle of the propeller when the conventional algorithm has no solution. An advanced approximation of the velocities induced by helical horseshoe vortices is applied in the model based on lifting surface theory. This approximate method will reduce computing time and remain good accuracy. Comparison between the two modeling techniques shows that the model based on strip theory which owns more advantage on both real-time and high-accuracy can meet the requirement.

scholarly journals Comparison of Physics-Based, Semi-empirical and Neural Network-based Models for Model-based Combustion Control in a 3.0 L Diesel Engine

Energies ◽  
2019 ◽  
Vol 12 (18) ◽  
pp. 3423 ◽  
Author(s):  
Hu ◽  
d’Ambrosio ◽  
Finesso ◽  
Manelli ◽  
Marzano ◽  
...  
Keyword(s):  
Neural Network ◽  
Steady State ◽  
Diesel Engine ◽  
Effective Pressure ◽  
Combustion Control ◽  
Ann Model ◽  
Control Applications ◽  
Model Based ◽  
Brake Mean Effective Pressure ◽  
Semi Empirical

A comparison of four different control-oriented models has been carried out in this paper for the simulation of the main combustion metrics in diesel engines, i.e., combustion phasing, peak firing pressure, and brake mean effective pressure. The aim of the investigation has been to understand the potential of each approach in view of their implementation in the engine control unit (ECU) for onboard combustion control applications. The four developed control-oriented models, namely the baseline physics-based model, the artificial neural network (ANN) physics-based model, the semi-empirical model, and direct ANN model, have been assessed and compared under steady-state conditions and over the Worldwide Harmonized Heavy-duty Transient Cycle (WHTC) for a Euro VI FPT F1C 3.0 L diesel engine. Moreover, a new procedure has been introduced for the selection of the input parameters. The direct ANN model has shown the best accuracy in the estimation of the combustion metrics under both steady-state/transient operating conditions, since the root mean square errors are of the order of 0.25/1.1 deg, 0.85/9.6 bar, and 0.071/0.7 bar for combustion phasing, peak firing pressure, and brake mean effective pressure, respectively. Moreover, it requires the least computational time, that is, less than 50 s when the model is run on a rapid prototyping device. Therefore, it can be considered the best candidate for model-based combustion control applications.

The Realization of Portable Ultrasonic TOFD Imaging and Detecting System

2013 ◽  
Vol 798-799 ◽  
pp. 651-655
Author(s):  
Si Yuan Wang ◽  
Xing Qun Zhao ◽  
Ling Xia
Keyword(s):  
Embedded System ◽  
Real Time ◽  
Data Storage ◽  
Weld Joint ◽  
High Accuracy ◽  
Field Structure ◽  
Joint Detection ◽  
Application Software ◽  
Quantitative Calculation ◽  
Tofd Image

This paper introduced the design of a portable ultrasonic TOFD imaging and detecting system, and developed the hardware and application software based on embedded system. We accomplished multi-channel ultrasonic transmission and acquisition, real time A-scan signal or TOFD image display, ultrasonic data storage and analysis of large capacity and automatic quantitative calculation of cracks in this system. This system offered rapid speed and high accuracy that could be used to locate and size cracks in both weld joint detection and field structure examination.

Active Vibration Control for Lathe Based Neural Network

2011 ◽  
Vol 84-85 ◽  
pp. 183-187 ◽  
Author(s):  
Jin Hua Wang ◽  
Wen Juan Huang ◽  
Hong Yan Zhang ◽  
Yao Gang Li
Keyword(s):  
Neural Network ◽  
Control System ◽  
Vibration Control ◽  
Active Vibration Control ◽  
System Model ◽  
Matlab Simulation ◽  
Model Based ◽  
Active Vibration ◽  
Simulation Results ◽  
Control System Model

In this paper, we took lathe as the research object, and established an active vibration control system model based on neural network AVC (Active Vibration Control) system, and the Matlab simulation results showed that the AVC system can reduce vibration effectively and improve the lathe’s accuracy.

Sign in / Sign up

Export Citation Format

Share Document