• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1998년도 하계학술대회 논문집 B
• /
• pp.417-419
• /
• 1998

In a large diesel engine actuator position servo system, it is impossible to isolate an actuator fault from a load torque with conventional fault detection isolation (FDI) schemes because they are propagated through a channel. This paper deals with a parity equation based residual generation to isolate them in the system. The actuator fault is modelled by a multiplicative type fault that can be characterized as discrepancies between the nominal and actual plant parameters, whereas the load torque is modelled by an additive disturbance. The transformation implemented in the residual generator should be determined on-line to achieve the isolation. Simulation studies show the practical applicability of the FDI scheme.

• International Journal of Control, Automation, and Systems
• /
• 제5권3호
• /
• pp.329-336
• /
• 2007

This paper considers a fault accommodation problem for inertial navigation systems (INS) that have redundant inertial sensors such as gyroscopes and accelerometers. It is wellknown that the more sensors are used, the smaller the navigation error of INS is, which means that the error covariance of the position estimate becomes less. Thus, when it is decided that double faults occur in the inertial sensors due to fault detection and isolation (FDI), it is necessary to decide whether the faulty sensors should be excluded or not. A new accommodation rule for double faults is proposed based on the error covariance of triad-solution of redundant inertial sensors, which is related to the navigation accuracy of INS. The proposed accommodation rule provides decision rules to determine which sensors should be excluded among faulty sensors. Monte Carlo simulation is performed for dodecahedron configuration, in which case the proposed accommodation rule can be drawn in the decision space of the two-dimensional Cartesian coordinate system.

• Asian-Australasian Journal of Animal Sciences
• /
• 제19권12호
• /
• pp.1742-1748
• /
• 2006

The experiment was conducted on 264 crossbred Karan-Fries (Holstein Friesian${\times}$Tharparkar) cows, over one year to explore the possibility of using milk urea (MU) concentration and milk protein content to monitor feeding adequacy under farm condition and to investigate the effects of different animal factors and season on MU concentration. Individual noon (1200 to 1300 h) milk samples were collected once in every month and analysed for urea and protein contents. Representative feed samples were also collected on the same day of milk collection and were analysed for CP content. A significant positive association (p<0.01) between MU concentration and milk yield was observed. MU concentrations (mg/dl) were found to be significantly (p<0.01) higher and lower in first lactation (44.8${\pm}$0.7) and in early lactation stage (40.7${\pm}$0.5), respectively. Average MU values were found to be significantly (p<0.01) higher in winter (50.7${\pm}$0.3) and lower in summer (32.9${\pm}$0.6). During the investigation, of the total MU observations, 50.3% were within the range of 30 to 50 mg/dl, 21.4% were <30 mg/dl and only 7.5% were >60 mg/dl. MU concentration was found to be associated significantly (p<0.05) with CP content of forages rather than concentrate. A close positive association (p<0.01) between MU level and daily milk protein (DMP) yield was observed during the investigation. The regression equation, DMP yield (g) = -24.6+33.5 daily milk yield (kg) +0.9 MU (mg/dl) was developed to establish the reference level of DMP yield. The result indicates that the effect of parity and stage of lactation may be ignored while interpreting MU values. However, reference MU values may be standardised separately for high milk yielders as level of milk yield contributes significantly to the variation of MU. The study revealed that the MU values together with DMP yield and milk protein content could be used as a potential non-invasive pointer to monitor feeding adequacy in dairy cows under farm conditions.

• 대한수의학회지
• /
• 제41권2호
• /
• pp.243-252
• /
• 2001

Milk constituents and somatic cell count (SCC) were analysed for 4,059 milk samples from 178 dairy farms from April to December in 1999. Correlations among each milk constituents, milk urea (MU) concentration and SCC in association with lactation stage, milk yield and parity, and balancing status of nutrients' supplies were analysed, and the results are summarized as follows; Averages of milk fat percent, total solids percent and milk yield were $3.72{\pm}0.91%$, $12.50{\pm}1.31%$ and $23.80{\pm}8.54kg$, respectively, whereas those were significantly lower during the summer season. In contrast, average of MU concentration was $0.0361{\pm}0.0006%$ which was significantly higher during the summer season. With milk yield, concentrations of fat, protein and SCC in milk decreased but concentrations of lactose and urea in milk and body condition score (BCS) were not altered. Concentrations of fat, protein, lactose, total solids, SNF, and urea in milk were significantly affected by stage of lactation ($P{\leq}0.0001$) but SCC was not changed. Parity of dairy cows had a significant effect on concentrations of fat ($P{\leq}0.02$), lactose ($P{\leq}0.0001$), total solids ($P{\leq}0.002$), and SNF ($P{\leq}0.0001$) in milk and milk yield ($P{\leq}0.0005$) but did not change concentrations of urea and protein in milk. Somatic cell count had significant positive correlationship with percentages of fat, protein and total solids ($P{\leq}0.0001$), respectively, but had negative correlationship with percentages of urea and lactose in milk and milk yield ($P{\leq}0.0001$). Milk urea concentration was negatively correlated with concentrations of protein, fat, total solids, and SNF in milk and milk yield ($P{\leq}0.0001$) and, according to regression analysis using milk urea concentration and SCC, following equation was obtained; $Y(MU)=3.688{\times}10^{-2}-4.04{\times}10^{-7}{\times}X(SCC{\times}1,000)(r^2=0.0038$, $P{\leq}0.0001$). We studied balance between protein and energy supplies to dairy cows in each farm based upon milk urea and protein concentrations, and results showed that 137 of total 178 farms fed cows unbalanced amounts of dietary protein and energy.

• 대한기계학회논문집B
• /
• 제41권11호
• /
• pp.735-742
• /
• 2017

본 연구에서는 PEM 연료전지 온도 센서의 고장을 감지 및 판별할 수 있는 모델 기반 센서 고장 감지 방법이 적용된다. 연료전지 차량이 작동하는 과정에서 스택 온도는 연료전지의 내구성에 영향을 미친다. 따라서 고장 진단 알고리즘이 고장 신호를 감지하는 것은 중요하다. 센서 고장 감지의 주요 목적은 연료전지 시스템의 안정적인 작동을 보장하여 고온과 저온으로부터 스택을 보호하는 것이다. 상태 공간에 기반한 패러티 방정식이 스택 온도와 냉각수 입구 온도와 같은 센서 고장을 감지하는데 적용되며, 잔차는 정상적인 온도 신호와 비교된다. 그리고 잔차는 현재의 센서 고장을 감지하는 다양한 고장 시나리오에 의해 평가된다. 결론적으로, 본 연구에서 설계된 고장 알고리즘이 고장 신호를 감지할 수 있다.

• 재활간호학회지
• /
• 제2권2호
• /
• pp.193-205
• /
• 1999

This study was carried out to identify the important risk factors for reduced bone mass of postmenopausal RA patients and to develop discriminant function which can classify postmenopausal RA patients with either reduced or normal bone mass. Through the literature review, individual characteristics such as age, body weight, height, age of menarche, duration of menopause, gravity, parity, and breast feeding period and factors of life style such as milk consumption exercise habit, alcohol intake, cigarette smoking, coffee consumpt ion , disease activity, corticosteroid therapy were identified as influencing factors of reduced bone mass in RA patients Sixty eight postmenopausal women with rheumatoid arthritis aged between 42 and 76 were selected among those who checked bone mineral density in lumbar spine and femur from october, 1998 to Apr il, 1999 at Dong-a university hospital in Pusan. Assessment of disease activity, duration of disease and corticosteroid therapy were made by the same rheumatologist and included Ritchie articular index, erythrocyte sedimentation rate, and C-reactive protein on measuring bone mineral density. Cumulative steroid dosage was calculated from the daily dosage multiplied by t h e number of days received. The information of other risk factor including health assessment score, individual characteristics and life style factors were collected by questionnaire. Bone mineral density(BMD) was measured using DXA at lumbar spine and femoral Ward's triangle. Discriminant function(regression equation) was developed for estimating the likelihood of the presence or absence of reduced bone mass. The results are as follows: Among the subjects, thirteen(19.1%) exhibited osteoporosis in lumbar spine and twenty four(35.3%) exhibited osteoporosis in femoral Ward's triangle. For lumbar spine, the variables significant were age, body weight, health assessment score, while for femoral Ward's triangle, age, body weight, duration of disease. But disease activity and corticosteroid therapy were not signigicant to distinguish reduced bone mass from normal bone mass. When the discriminant function was evaluated by comparing the observed out come with predicted out come, the discriminant function correctly classified 85.4% of patients with reduce bone mass and 63.0% of patients with normal bone mass in the lumbar spine and 100% of patients with reduced bone mass and 9.1% of patients with normal bone mass in the femoral Ward's triangle. In summary, we found that osteoporosis in postmenopausal women with RA is more evident at the femur than the lumbar spine. Also the important discriminant factors of reduced bone mass postmenopausal women with RA were age, body weight , duration of disease and health disability. In nursing situation, the efforts to improve of functional capacity of postmenopausal women with rheumatoid arthritis should be considered to prevent osteoporosis and fractures. Also we recommend those postmenopausal women with RA who are classified as a group of the reduced bone mass in the discriminant function should examine the bone mineral density to further examine the usefulness of this discriminant function.