The practice of measuring the formation density with a standard pulsed-neutron system has been described in the literature1,2,3,4,5 , and has been successfully applied in many reservoirs. This success leads to the obvious question: Can a system be developed that is optimized for this measurement?

To understand which parameters should be measured, characterized, or require improved response, a more exact analytical model is required. The pulsed-neutron density measurement is based on the transport of gamma rays created by inelastic scattering of fast neutrons. The current density algorithm uses a diffusion model for the gamma ray transport process. This simple model does not include the effect of variations in fast neutron parameters on the process of gamma ray production. As a result, an ambiguity is introduced in the density log that is associated with the dynamic nature of the gamma source. The initial distribution of gamma rays is a function of atom density and hydrogen content and must be properly described as a first step toward a more exact transport model.

The pulsed-neutron density measurement is analyzed and optimization is achieved using theoretical models, responses from laboratory formations and test wells, and through computer modeling. Along with data supporting an improved analytical model of the measurement, data is shown for incorporating fast neutron detectors and for optimizing spacings of the gamma ray detectors. As a result of this study, a prototypical detector array is described for improved density response.

Keywords:
atom density, count rate, machine learning, contribution, porosity, formation density, neural network, fast neutron detector, gamma, spe 71742
Subjects:
Formation Evaluation & Management, Open hole/cased hole log analysis
Copyright 2001, Society of Petroleum Engineers
You can access this article if you purchase or spend a download.