Development of a Neural Network-Based Real-Time Fatigue Monitoring System for the Heavy Load Carrying Facility

2006 ◽  
pp. 201-212
Author(s):  
Jong Choon Kim ◽  
Sung Wook Jung ◽  
Jae Boong Choi ◽  
Yoon Suk Chang ◽  
Young Jin Kim ◽  
...  
Keyword(s):  
Neural Network ◽  
Real Time ◽  
Monitoring System ◽  
Heavy Load ◽  
Load Carrying ◽  
Fatigue Monitoring

Development of a Neural Network-Based Real-Time Fatigue Monitoring System for the Heavy Load Carrying Facility

2006 ◽  
Vol 110 ◽  
pp. 201-212
Author(s):  
Jong Choon Kim ◽  
Sung Wook Jung ◽  
Jae Boong Choi ◽  
Yoon Suk Chang ◽  
Young Jin Kim ◽  
...  
Keyword(s):  
Fatigue Life ◽  
Real Time ◽  
Monitoring System ◽  
Life Time ◽  
Fatigue Loading ◽  
Heavy Load ◽  
Web Based ◽  
Plant Maintenance ◽  
Load Carrying ◽  
Fatigue Monitoring

The heavy load carrying facility, such as ladle crane, is operating under severe working environment. It usually carries melted iron to the furnace, and thus, the accident due to crane failure may cause detrimental damage to the entire steel making factory. While the ladle crane is designed for 20 years of safe operation in a steel making company, several critical cracks due to fatigue loading have been reported during the maintenance process. In order to prevent fatal failure due to crack growth, ladle crane has been periodically inspected and maintained. However, the inspection and maintenance including repair and replacement cause the whole manufacturing line to stop, it is critical to set the appropriate inspection interval and replacement criteria. For this reason, the importance of plant maintenance (PM) has been highly raised to provide efficient plant operation. Recently, a number of engineering methodologies, such as fitness for service guidelines (FFS) and plant lifecycle management (PLM) system, have been applied to improve the plant operation efficiency. Also, a network-based business operation system, which is called ERP (Enterprise Resource Planning), has been introduced in the field of plant maintenance. However, there hasn’t been any attempt to connect engineering methodologies to the ERP PM(Plant Maintenance) system. In this paper, an engineering methodology which provides life time evaluation under fatigue loading has been implemented to the web-based ERP PM system along with real-time fatigue monitoring system. In order to monitor the real time loading, a web-based fatigue monitoring system for ladle crane has been developed and installed inside the ladle crane. For the estimation of fatigue life, 3-dimensional finite element (FE) analyses were conducted for actual transients. Finally, the fatigue life time estimation program is developed by integrating FE analysis results and real-time monitoring data. For the direct calculation of remained fatigue life, an artificial neural network (ANN) algorithm has been applied. The proposed system is expected to play a great role in determining appropriate inspection and maintenance schedule which has become critical issue for the efficient plant maintenance.

A Method for Determining Quasi-Static and Dynamic Riser Inclination Using Collocated Accelerometers and Angular Rate Sensors

2014 ◽  
Author(s):  
Scot McNeill ◽  
Paul Angehr ◽  
Dan Kluk ◽  
Tomokazu Saruhashi ◽  
Ikuo Sawada ◽  
...  
Keyword(s):  
Real Time ◽  
Monitoring System ◽  
Body Force ◽  
Angular Rate ◽  
Low Frequency ◽  
Monitoring Systems ◽  
Rate Data ◽  
Fatigue Monitoring ◽  
Inclination Angles ◽  
Force Components

A method is described for determining quasi-static and dynamic riser angles using measured data typically found in a riser fatigue monitoring system, specifically acceleration and angular rate data. Quasi-static riser inclination and orientation of the inclination plane are determined from the low frequency triaxial accelerations, containing measurement of the gravitational body force. Components of the gravitational body force along the accelerometer axes vary slowly with the riser quasi-static response. The slowly varying Euler angles are determined from the components of gravity along the three axes. Dynamic riser inclination along and transverse to the quasi-static inclination plane are determined by integration of the angular rates, followed by transformation into a coordinate system aligned with the quasi-static inclination plane. The quasi-static and dynamic inclination angles are combined to arrive at the time trace of riser inclination angles. Following implementation of the method in Matlab®, the procedure was validated and the program verified using laboratory test data. A double-gimbaled platform was constructed, on which were mounted a triaxial accelerometer, biaxial angular rate and biaxial inclinometer (reference sensor). A battery of static and dynamic tests was carried out on the platform. Machinists’ levels and angle gauges were used to set the inclination in the various tests. The angles derived from the acceleration and angular rate data were compared to those of the reference inclinometer. Angle estimates were shown to match the reference angles with negligible error. The method was then implemented into the real-time Riser Fatigue Monitoring System (RFMS) aboard the Chikyu drillship. The algorithm was run using data from an emergency disconnect event that occurred in November, 2012. Quasi-static riser inclination angles were quite large due to high currents near the sea surface. Dynamic riser inclination angles proved to be significant due to Vortex Induced Vibration of the lower portion of the riser that immediately followed the disconnect event. It is important to note that the method uses data that is typically already included in real-time riser monitoring systems. Therefore only a software update is required to provide real-time riser angle information. If the method is built into data processing routines for real-time riser monitoring systems, the need for additional instrumentation, such as inclinometers near flex joints, may be circumvented. On the other hand, if inclinometers already exist, the method serves as an independent source of riser angle information at several locations on the riser. The method can also be used to calculate riser and Blow out Preventer (BOP) stack angles from data recorded using stand-alone, battery-powered loggers.

scholarly journals Monitoring Air Quality and Estimation of Personal Exposure to Particulate Matter Using an Indoor Model and Artificial Neural Network

2020 ◽  
Vol 12 (9) ◽  
pp. 3794 ◽  
Author(s):  
Hyeon-Ju Oh ◽  
Jongbok Kim
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Air Quality ◽  
Real Time ◽  
Monitoring System ◽  
Personal Exposure ◽  
Operating Conditions ◽  
Real Time Monitoring ◽  
Potential Threat ◽  
Artificial Neural

Exposure to particulate materials (PM) is known to cause respiratory and cardiovascular diseases. Respirable particles generated in closed spaces, such as underground parking garages (UPGs), have been reported to be a potential threat to respiratory health. This study reports the concentration of pollutants (PM, TVOC, CO) in UPGs under various operating conditions of heating, ventilation and air-conditioning (HVAC) systems using a real-time monitoring system with a prototype made up of integrated sensors. In addition, prediction of the PM concentration was implemented using modeling from vehicle traffic volumes and an artificial neural network (ANN), based on environmental factors. The predicted PM concentrations were compared with the level acquired from the real-time monitoring. The measured PM10 concentrations of UPGs were higher than the modeled PM10 due to short-term sources induced by vehicles. The average inhalable and respirable dosage for adult was calculated for the evaluation of health effects. The ANN predicted PM concentration showed a close correlation with measurements resulting in R2 ranging from 0.69 to 0.87. This study demonstrates the feasibility of the use of the air quality monitoring system for personal-exposure to vehicle-induced pollutant in UPGs and the potential application of modeling and ANN for the evaluation of the indoor air quality.

Sign in / Sign up

Export Citation Format

Share Document