• Journal of Electrical Engineering and Technology
• /
• v.9 no.3
• /
• pp.908-925
• /
• 2014

Induction Motor (IM) has several desirable features for high performance adjustablespeed operation. This paper presents the design of a robust controller for vector control induction motor drive performances improvement. Proposed predictive speed controller, which is aimed to guarantee the stability of the closed loop, is based on the Poisson-Laguerre (PL) models for the association vector control drive and the induction motor; without necessity of any mechanical parameter, and requires only two control parameters to ensure implicitly the integrator effect on the steady state error, load torque disturbances rejection and anti-windup effect. In order to improve robustness, insensitivity against external disturbances and preserve desired performance, adaptive control is added with the aim to ensure an online identification of controller parameters through an online PL models identification. The proposed control is compared with the conventional approach using PI controller. Simulation with MATLAB/SIMULINK software and experimental results for a 1kW induction motor using a dSPACE system with DS1104 controller board are carried out to show the improvement performance.

• 한국전산유체공학회:학술대회논문집
• /
• 2010.05a
• /
• pp.444-445
• /
• 2010

Predictive modeling schemes have been developed to characterize the heat Transfer and thermo-mechanical behavior for the plasma display panel (PDP) in operation. The inverse approach was adopted to predict the accurate temperature distribution and deformation in PDP. The predictive models were validated with the measurements from real panel. The developed models could be utilized to predict and/or improve the product quality of PDP.

• Korean Journal of Remote Sensing
• /
• v.33 no.1
• /
• pp.25-35
• /
• 2017

Spatial downscaling with fine resolution auxiliary variables has been widely applied to predict precipitation at fine resolution from coarse resolution satellite-based precipitation products. The spatial downscaling framework is usually based on the decomposition of precipitation values into trend and residual components. The fine resolution auxiliary variables contribute to the estimation of the trend components. The main focus of this study is on quantitative analysis of impacts of trend component estimates on predictive performance in spatial downscaling. Two regression models were considered to estimate the trend components: multiple linear regression (MLR) and geographically weighted regression (GWR). After estimating the trend components using the two models,residual components were predicted at fine resolution grids using area-to-point kriging. Finally, the sum of the trend and residual components were considered as downscaling results. From the downscaling experiments with time-series Tropical Rainfall Measuring Mission (TRMM) 3B43 precipitation data, MLR-based downscaling showed the similar or even better predictive performance, compared with GWR-based downscaling with very high explanatory power. Despite very high explanatory power of GWR, the relationships quantified from TRMM precipitation data with errors and the auxiliary variables at coarse resolution may exaggerate the errors in the trend components at fine resolution. As a result, the errors attached to the trend estimates greatly affected the predictive performance. These results indicate that any regression model with high explanatory power does not always improve predictive performance due to intrinsic errors of the input coarse resolution data. Thus, it is suggested that the explanatory power of trend estimation models alone cannot be always used for the selection of an optimal model in spatial downscaling with fine resolution auxiliary variables.

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10b
• /
• pp.156-159
• /
• 1996

An adaptive predictive control for steam-reforming plant which consist of a steam-gas reformer and a waste heat steam-boiler was studied by using MIMO bilinear model. The simulation experiments of the process identification were performed by using linear and bilinear models. From the simulation results it was found that the bilinear model represented the dynamic behavior of a steam-reforming plant very well. ARMA model was used in the process identification and the adaptive predictive control. To verify the performance and effectiveness of the adaptive predictive controller proposed in this study the simulation results of steam-reforming plant control based on bilinear model were compared to those of linear model. The simulation results showed that the adaptive predictive controller based on bilinear model provides better performance than those of linear model.

• ICROS
• /
• v.2 no.1
• /
• pp.59-71
• /
• 1996

In the last decade, the model predictive control methods have enjoyed many industrial applications with successful results. Although the general predictive control methods for nonlinear chemical processes are not yet formulated, the promising features of the model predictive control methods attract attentions of many researchers who are involved with difficult but important nonlinear process control problems. Recently, the class of bilinear model has been introduced as an useful tool for examining many nonlinear phenomena. Since their structural properties are similar to those of linear models, it is not difficult to develop a robust adaptive model predictive control method based on bilinear model. We expect that the model predictive control method based on bilinear model will expand its region in the world of nonlinear systems.

• Transactions of the Korean Society of Machine Tool Engineers
• /
• v.11 no.4
• /
• pp.89-96
• /
• 2002

This paper aims to develop an intelligent model for predicting top-bead width for the robotic GMA(Gas Metal Arc) welding process using BP(Back-propagation) neural network and multiple regression analysis. Firstly, based on experimental data, the basic factors affecting top-bead width are identified. Then BP neural network model and multiple regression models of top-bead width are established. The modeling methods and procedure are explained. The developed models are then verified by data obtained from the additional experiment and the predictive behaviors of the two kind of models are compared and analysed. Finally the modeling methods, predictive behaviors md the advantages of each models are discussed.

• Journal of Integrative Natural Science
• /
• v.4 no.4
• /
• pp.266-272
• /
• 2011

Focal adhesion kinase (FAK) is a potential target for the treatment of primary cancers as well as prevention of tumor metastasis. To understand the structural and chemical features of FAK inhibitors, we report comparative molecular field analysis (CoMFA) for the series of 7H-pyrrolo(2,3-d)pyrimidines. The CoMFA models showed good correlation between the actual and predicted values for training set molecules. Our results indicated the ligand-based alignment has produced better statistical results for CoMFA ($q^2$ = 0.505, $r^2$ = 0.950). Both models were validated using test set compounds, and gave good predictive values of 0.537. The statistical parameters from the generated 3D-QSAR models were indicated that the data are well fitted and have high predictive ability. The contour map from 3D-QSAR models explains nicely the structure-activity relationships of FAK inhibitors and our results would give proper guidelines to further enhance the activity of novel inhibitors.

• Journal of information and communication convergence engineering
• /
• v.21 no.1
• /
• pp.82-89
• /
• 2023

Various machine-learning models may yield high predictive power for massive time series for time series prediction. However, these models are prone to instability in terms of computational cost because of the high dimensionality of the feature space and nonoptimized hyperparameter settings. Considering the potential risk that model training with a high-dimensional feature set can be time-consuming, we evaluate a feature-importance-based feature selection method to derive a tradeoff between predictive power and computational cost for time series prediction. We used two machine learning techniques for performance evaluation to generate prediction models from a retail sales dataset. First, we ranked the features using impurity- and Local Interpretable Model-agnostic Explanations (LIME) -based feature importance measures in the prediction models. Then, the recursive feature elimination method was applied to eliminate unimportant features sequentially. Consequently, we obtained a subset of features that could lead to reduced model training time while preserving acceptable model performance.

• Journal of Korea Artificial Intelligence Association
• /
• v.1 no.2
• /
• pp.21-25
• /
• 2023

The purpose of this study was to compare the performance using multiple regression models to predict the energy consumption of steel industry. Specific independent variables were selected in consideration of correlation among various attributes such as CO2 concentration, NSM, Week Status, Day of week, and Load Type, and preprocessing was performed to solve the multicollinearity problem. In data preprocessing, we evaluated linear and nonlinear relationships between each attribute through correlation analysis. In particular, we decided to select variables with high correlation and include appropriate variables in the final model to prevent multicollinearity problems. Among the many regression models learned, Boosted Decision Tree Regression showed the best predictive performance. Ensemble learning in this model was able to effectively learn complex patterns while preventing overfitting by combining multiple decision trees. Consequently, these predictive models are expected to provide important information for improving energy efficiency and management decision-making at steel industry. In the future, we plan to improve the performance of the model by collecting more data and extending variables, and the application of the model considering interactions with external factors will also be considered.

• Journal of Information Technology Applications and Management
• /
• v.11 no.1
• /
• pp.117-135
• /
• 2004

Since human decision making behavior is likely to follow nonlinear strategy, it is conjectured that the human decision making behavior can be modeled better by nonlinear models than by linear models. All that linear models can do is to approximate rather than model the decision behavior. This study attempts to test this conjecture by analyzing human decision making behavior and combining the results of the analysis with predictive performance of both linear models and nonlinear models. In this way, this study can examine the relationship between the predictive performance of models and the existence of valid nonlinear strategy in decision making behavior. This study finds that the existence of nonlinear strategy in decision making behavior is highly correlated with the validity of the decision (or the human experts). The second finding concerns the significant correlations between the model performance and the existence of valid nonlinear strategy which is detected by Lens Model. The third finding is that as stronger the valid nonlinear strategy becomes, the better nonlinear models predict significantly than linear models. The results of this study bring an important concept, validity of nonlinear strategy, to modeling human experts. The inclusion of the concept indicates that the prior analysis of human judgement may lead to the selection of proper modeling algorithm. In addition, lens Model Analysis is proved to be useful in examining the valid nonlinearity in human decision behavior.