Memory pulsed neutron-neutron logging

Abstract

This paper describes the MINK tool which is based on pulsed neutron-neutron (PNN) logging technology, its key principles and differences from other logging and data interpretation methods, and illustrates its applications to estimating oil saturation in wells.

The principal feature of the MINK tool is its data recording technique in which all neutron-count decays are saved to memory and processed separately. In conventional PNN logging, 100 decays are initially accumulated and then averaged before processing, which results in information loss.

Theoretical considerations and available experimental data have indicated that the probability distribution of neutrons is governed by Poisson's law [1] and allows the use of the maximum likelihood method (MLM) for experimental data fitting [2] with a substantial advantage over the conventional least-square method (LSM). This was particularly important for reservoir characterisation, as the rock's response is mainly recorded at late times. Neutron count rates at late times are small, and the least square method becomes a hit-and-miss technique. The authors have developed algorithms and software for the efficient and reliable determination of single- and double-exponential approximation parameters using the maximum likelihood method.

The processing of data from three selected wells has shown that Sigma profiles determined by the least square method are noisier than those determined by the maximum likelihood method. Moreover, some thin reservoir units were not seen in Sigma profiles obtained by the least square method, in contrast to those obtained by the maximum likelihood method.

Statistical modelling has been performed to validate the developed algorithms. Decays with experimentally determined decay times and Poisson's distribution of neutron count decays have been generated. Modelling data processing has shown that the maximum likelihood method provides 1.7 to 2 times higher accuracy than the least square method under equal conditions or 3 to 4 times smaller data collection volumes.

The MINK technology is expected to determine the oil content in dense reservoirs and low-salinity wellbore fluids.

Introduction

Hydrocarbon development monitoring requires the evaluation of the current and residual oil and gas content of reservoirs to be performed in cased wells. This objective can be achieved by pulsed neutron logging. The advantage of PNN logging over pulsed neutron gamma (PNG) logging is that the former has no natural neutron background (only several neutrons per hour) or induced neutron radiation during tool operation. For this reason, PNN logging can be employed in zones where PNG signals are masked by the gamma-ray background. Gamma-ray anomalies with gamma radiation levels higher than in the open hole greatly impede water saturation analysis, and pulsed neutron gamma logging produces incorrect readings. In such cases, PNN logging can be advantageously used instead of PNG logging.

This paper describes the MINK pulsed neutron-neutron logging technology in memory mode. Section 1 describes the MINK memory PNN logging tool and its operating principles. The technical specifications of the MINK tool are given in Appendix. Section 2 describes the distribution of neutron counts. General physical considerations together with the analysis of real and modelled data have shown that this distribution is governed by Poisson's law. The maximum likelihood method has been described in terms of its application to processing PNN logging data for single- and double-exponential decay approximation. Uncertainty in the determination of adjustable parameters (reservoir Sigma values) have been compared for the least square method applied to averaged and accumulated decays and the maximum likelihood method applied to unaveraged decays. It has been shown that the latter method is 1.7 to 2 times more accurate for equal numbers of decays. Section 3 presents log processing results and shows how the above methods can be used for reservoir characterisation. The key findings are briefly given in the Conclusions.