Apparatus and method for softsensing fuel cell system
AssigneeKorea Institute of Energy Research
InventorsMinjin KIM, Young-Jun SOHN, Gu-Gon PARK, Byungchan BAE, Sung-Dae YIM, Young-Woo CHOI, Seok-Hee PARK, Tae-Hyun YANG, Won-Yong LEE, Chang-Soo KIM
Life of this patent
Estimating the term…
Abstract
Provided are an apparatus and a method for soft-sensing a vehicle or stationary fuel cell system capable of predicting a quality variable based on a process variable and providing the predicted quality variable in real time.
Claims20
An apparatus for soft-sensing a fuel cell system, comprising:
- a connecting unit detachable from a control unit for being connected to the outside of a stationary fuel cell system;
- a collecting unit connected to the connecting unit and receiving data of the stationary fuel cell system;
- a quality variable predicting unit connected to the collecting unit and predicting a quality variable of the stationary fuel cell system based on the received data;
- and
- a monitoring unit connected to the quality variable predicting unit and outputting the predicted quality variable.
A method for soft-sensing a fuel cell system by an apparatus for soft-sensing a stationary fuel cell system detachable from a control unit for being connected to the outside of the stationary fuel cell system and configured to include a connecting unit, a collecting unit, a quality variable predicting unit, and a monitoring unit, comprising:
- a process variable selecting step of selecting a process variable having an effect on a quality variable;
- a modeling data collecting step of collecting measurement results of a flooding index of a hydrogen electrode of a stationary fuel cell stack or a flooding index of an oxygen electrode of the stationary fuel cell stack, which is the quality variable, depending on a variation of the process variable selected in the process variable selecting step;
- a model building-up step of building-up an experiment model representing a correlation between the process variable and the quality variable collected in the modeling data collecting step;
- a model accuracy verifying step of verifying a coefficient of determination (R2:
- R-square) depending on a regression analysis of the model built-up in the model building-up step;
- and
- a model applying step of applying the model verified in the modeling accuracy verifying step in the case in which the coefficient of determination depending on the regression analysis of the model in the modeling accuracy verifying step is a predetermined allowable value or more.
An apparatus for soft-sensing a fuel cell system, comprising:
- a connecting unit detachable from a control unit for being connected to the outside of the vehicle fuel cell system;
- a collecting unit connected to the connecting unit and receiving data of the vehicle fuel cell system;
- a quality variable predicting unit connected to the collecting unit and predicting a quality variable of the vehicle fuel cell system based on the received data;
- and
- a monitoring unit connected to the quality variable predicting unit and outputting the predicted quality variable.
A method for soft-sensing a fuel cell system by an apparatus for soft-sensing a fuel cell system detachable from a control unit in the vehicle fuel cell system and configured to include a connecting unit, a collecting unit, a quality variable predicting unit, and a monitoring unit, comprising:
- a process variable selecting step of selecting a process variable having an effect on a quality variable;
- a modeling data collecting step of collecting measurement results of a flooding index of a hydrogen electrode of a vehicle fuel cell stack or a flooding index of an oxygen electrode of the vehicle fuel cell stack, which is the quality variable, depending on a variation of the process variable selected in the process variable selecting step;
- a model building-up step of building-up an experiment model representing a correlation between the process variable and the quality variable collected in the modeling data collecting step;
- a model accuracy verifying step of verifying a coefficient of determination (R2:
- R-square) depending on a regression analysis of the model built-up in the model building-up step;
- and
- a model applying step of applying the model verified in the modeling accuracy verifying step in the case in which the coefficient of determination depending on the regression analysis of the model in the modeling accuracy verifying step is a predetermined allowable value or more.
Dependencies are inferred from the claim text.