> ,+ \pJava Excel API v2.6 Ba==h\:#8X@"1Arial1Arial1Arial1Arial + ) , * `=2DC, title[*]contributor[author]contributor[advisor]keywords[*]date[issued] publisher citationsidentifier[uri]identifier[doi]abstractrelation[journal]relation[volume]relation[no]relation[page]qModel based control and fault diagnosis of a molten carbonate fuel cell plant;
5а ̸ ՜X xD \ ȴ t;Kim, Tae Youngl2019-02\ՑYPwhttps://repository.hanyang.ac.kr/handle/20.500.11754/99842;
http://hanyang.dcollection.net/common/orgView/200000434618;̸ ¤\@ ̰X (x D Xp ֬ pŴ Xt <\ T ǵȲ. 5 а ̸ (MCFC) x ̸ Dt <\ @ % (D 1\. ̸ X 8Ĭ Ɣ ι@ | \ 1t T. h 0| X l mĳ ιD . MCFC Ȉ \ X l| q¤0 t X MCFC Ȍ Ɛǔ Ȍ X 1D ¤$ x%X . tǔ MCFC Ȍ | |0 ` ǔ ȴ D HXՌ l X䲔 D X\. ȴ D <\ X$t % ȴ ƌX ! xD XՔ t Dt. MCFCX 1 1D X0 t, MCFC X ٳ (ĳ ٳ ̹ | Xp, t ȴ | 0 t (x ȴ 0 t | \. X ȴ ȵX @ \8 xD 0<\ X̹ MCFC \8X 0t x@ ȴ ¤\ ȩ D J. l MCFC ¤\ \ 0 tٳ (ARMA) x, \͌ ǹ 0 8 (LS-SVM) x x ½ݹ (ANN) xD ǜ% pt0| 0<\ X. t x ARMA x@ \X 1D . H ARMA xD 0<\ MCFC \8X ٳD \ x ! ȴ (MPC) )D X. DP| t T Rigorous xD 0<\ \ MPCxD 0 X. ¬tXX ARMA xD 0<\ \ x ! ȴ 0x T Rigorous x \ ȴ 1D İD . ^ \ @ t H xD \ 5 а ̸ (MCFC) ȴ xĳ X l| q¤0 t $X| \͌T ` DՔ . hX D ļX !` t @ 1D tL . 5 а ̸ (MCFC) ȌŔ \ X\̹ ǔ ɷ LŌ ¤\t |<\ ¤\X D ptD ļXՔ p . ɷ ¤\ Ȳ0 D D t hD XՔ )<\ ٳXՔ ٳHt \ Ȳ0 ¤\@ U֥ ¤\ X0 4 . 1 (Principal Component Analysis, PCA)@ Ȍ X hD X X0 t |<\ ( 0 t. kW MCFC ȌX ɷ Ȳ0 )@ t PCA| 0<\ H. PCA Ĭ \8 Ȳ0 p pt0 4X \. 췘 MCFC pt0є X |(X Ĭ| и. DPCA (Dynamic Principal Component Analysis) Ĭ pt0 \ tհE<\ H. DPCA| t MCFCX ٳ ଥ X ٳD UxX0 t PCA ĬD ȩ X. |8 0tX PCA@ H DPCA )D DP X. pt0є 1 D t . ¬tX DPCA T @ 1D .; Fuel cell systems offer clean and efficient energy production and are currently under intensive commercialization by several manufacturers. Molten carbonate fuel cell (MCFC) is achieving relatively high output and efficiency compared to other fuel cells. The global demand of fuel cell has been growing vigorously in many industrial fields. With the increase of demand, the requirements of the customers have also been getting sophisticated and strict. In order to meet customer demands for molten carbonate fuel cell products, most of molten carbonate fuel cell power plant operators are trying to enhance the flexibility of plant operation. That means they are forced to have frequent control changes that cause to produce fault of molten carbonate fuel cell power plant trip. To carry out control changes effectively, it is essential to develop a prediction model for prediction of power and control factor. To improve availability and performance of molten carbonate fuel cells, the operating temperature of a molten carbonate fuel cells (MCFC) stack should be strictly maintained within a specified operation range and an efficient control technique should be employed to meet this objective. While most of modern control strategies are based on process models, many existing models for a MCFC process are not ready to be applied in synthesis and operation of control systems. In this study, auto-regressive moving average (ARMA) model, least square support vector machine (LS-SVM) model and artificial neural network (ANN) model for the MCFC system are developed based on input output operating data. Among these models, the ARMA model showed the best tracking performance. A model predictive control (MPC) method for the operation of a MCFC process is developed based on the proposed ARMA model. For the purpose of comparison, a MPC scheme based on the linearized rigorous model for a MCFC process is developed. Results of numerical simulations <}
show that model predictive control (MPC) based on the auto-regressive moving average (ARMA) model exhibits better control performance than that based on the linearized rigorous model. As mentioned, besides Molten Carbonate Fuel Cell (MCFC) control through the proposed model, it is necessary to minimize the fault to satisfy the customer's demand. If the types of faults can be identified and predicted in advance, control can be provided. In a Molten Carbonate Fuel Cell (MCFC) power plant, a univariate alarm system that has only upper and lower limits is usually employed to identify abnormal conditions in the system. While univariate alarms are adopted for system monitoring and analysts are operating in a way to diagnose faults through alarm analysis, this simple monitoring and diagnostic system is limited for use in extended systems. Principal Component Analysis (PCA) is the most commonly used dimensionality reduction technique for detecting and diagnosing faults in power plant processes. A multivariate monitoring method for kW MCFC power plant has already been proposed based on PCA. PCA is used for statistical process monitoring it relies on the assumption that data are time independent. However, MCFC operating data represents a series of correlations with time. Dynamic Principal Component Analysis (DPCA) has been suggested as a remedy for high-dimensional and time-dependent data. The time series of PCA has been applied to confirm the normal operation of MCFC and operation at fault occurrence through DPCA. In this paper, we compared the existing PCA and the suggested DPCA method. Actual operating data was used for performance verification. Simulation results show that DPCA shows better performance.&HQsj&)KLngk
24
dMbP?_*+%" ,,??U
>@
!"$%&'()*Root EntryWorkbook6SummaryInformation(DocumentSummaryInformation8#