QMC DATA MINING PROGRAM for Windows
ANN MODEL PROGRAM
The ANN Model is based on Artificial Neural Network (ANN) technology. The ANN Model is a multi-variable, linear and non-linear modeling tool that determines relationships between data for any type of process. The ANN Model is a prediction tool used to establish a model for the desired process and allows the user to pick the variables to model from and to. Five proven types of algorithms are included: linear, sine, sinh, tangent and tanh. These different technologies combine the best features neural nets have to offer with the highest accuracy and speed.
A full graphical interface is provided to remove data, view results and select model independent and dependent variables.
The ANN Model proves to be highly accurate over a wide range of data, vital in prediction of the distillation operation at different operating conditions. With this model, plant personnel can verify proper operation and view disturbances in the process.
Provided in the ANN Model are:
- ANN Model Linear
- ANN Model Sine
- ANN Model Sinh
- ANN Model Tangent
- ANN Model Tanh
RESULTS:
QMC Results Screen provides graphics showing the predicted versus actual and current numerical values from the data set modeled. The QMC Results Screen allows the user to set the input values to view desired outcomes of changes in process variables. Let the results run automatically or use the step action.
QMC Sensitivity Screen provides graphics showing the predicted versus actual and sensitivities in graphical formats. This allows the user to set the input values to see desired outcomes of changes in process variables and sensitivity changes. Let the results run automatically or use the step action.
QMC Optimization Screen – Provides a way to determine the settings to reach a goal. The user can select the result requirements and the optimization routine will adjust the dependent variable values to meet the requirements. If needed, the user can also constrain the independent values to a selected value.
QMC ModelSolver Results - Provides all equations needed to calculate results. This can be provided as computer code for interfacing with the DCS or historian computer. Specified templates can be provided.
Excel Results Summary - Used to evaluate different modeling results and keep a record of corresponding accuracies.
Excel Add Predicted – Provides the predicted value in a spreadsheet for future use in the graphic packages or for model comparisons.
Excel Calculations – The QMC Program provides interfacing to Excel to display all model results in spreadsheet form. Neural net weights are given for the hidden layers; weights for all layers are provided as well as all calculation results. Once in Excel form, all Excel functions are available for sensitivity analysis and optimization.
System Requirements
Available for: Win95/98/NT
Minimum CPU Requirements: 486
Minimum Ram: 8 MB
Version: 6
Published: 2000