NNLAP� Neural Network Log Analysis Program
NNLAP� uses neural network technology to generate synthetic wireline log and core properties from the available curve data. Neural networks learn the nature of the dependency between the curves you have abundantly available and the curves you need through a carefully selected and representative set of training examples. NNLAP� is an empirical method which offers significant advantages over traditional methods in that it: 1) does not require a mathematical model describing the predictive relationships, 2) yields robust solutions with only a few, well-chosen training examples, 3) preserves original data variability in the neural network-constructed mathematical model, 4) will not over-predict mean values (as will linear regression techniques), 5) is interactive, and allows the operator to use his/her knowledge and experience in training and validating the neural network solution, and 6) is fast, accurate, intuitive, and easy to learn to use. NNLAP� will generate high-quality synthetic curves (compressional or shear wave acoustic logs, porosity logs, NMR, core permeability, etc.), but can also be used to predict rock types, facies, mineralogy, and production data.
PC (Pentium or better recommended)