Objetivo The main objective of this project was to establish and quantify the causes of apparently anomalous logs from gamma-ray density tools. To do this required the development and validation of a special version of the McBEND code, called McDUFF, which enables the efficient prediction of individual detector count rates to a very high accuracy. A full range of experimental benchmark data was provided for the validation of McDUFF ranging from the determination of basic detector response functions in simplified geometries to experimental logging tool measurements in full scale test-blocks. The latter have also be used to simulate variations in the borehole environment paying particular attention to rugosity of different amplitudes and frequency.The objectives of the project were met by the production and use of :The McDUFF Monte Carlo Code.The ARCAS Rig - a simplified experimental logging tool/formation rig, having a variable geometry.The SPARTAN Gamma-ray Density Calibration Facility.The RODENT reference gamma-ray density logging tool.A large data base of high quality benchmark logging data.The McDUFF code developed during the project is essentially a cut-down version of the general purpose code McBEND with a consequent simplification of the manual and an improved user image.The small diameter collimators on both the source and the near detector crystal effectively preclude the application of the unique form of splitting and Russian Roulette used in McBEND. The forced flight method was developed to improve the efficiency of calculations for configurations with collimated detectors. The principle of the method is to define a geometric interface, such as a disc which is positioned in the collimator, and to assume that any particle that reaches the detector must have previously crossed this surface. For each gamma-ray scattering event which occurs in the formation there is a possibility that the scattered photon will travel uncollided through the interface and onto the detector. The photons are forced to follow this path and the weight of the particle is reduced according to the probability of this outcome following the scattering. With the saving of a factor of about 50 attributable to the forced flight option, execution times to achieve an accuracy of +/- 1% for a typical short spaced Nal detector will be reduced to about 80 hours on a medium powered workstation.Two approaches were developed for calculating the pulse height distribution (PHD) for the detector.a) The gamma flux spectrum calculated within the detector is multiplied by an experimentally determined response function to obtain the PHD.b) A special algorithms used to directly score the photon energy deposition within the crystal.Both methods have been satisfactorily compared with measurements for simple benchmark conditions. The second approach was adopted for use as it involves less data manipulation.The experimental validation data was gained from the ARCAS Rig and the SPARTAN development calibration facility using the RODENT tool.The validation data obtained in this project is some of the most exhaustive available for the modelling of gamma-ray density sondes. The validation studies have progressed from simple studies through to full tool geometry in a perturbed borehole. In doing so the C/M ratios for integrated pulse height have moved from 1.0 to 0.8. The cause of the under prediction in the SPARTAN studies remains unresolved. As most analysts use these techniques for relative studies the loss of the absolute calculation in the short term is not of major consequence. For the longer term it is essential that it is resolved so that all the mechanisms that affect the tool response can be fully understood.The project was planned in several phases :i) Computer Code DevelopmentEarly applications of the Monte Carlo method for modelling tool response pointed to the need for substantial improvements in speed for the prediction of near detector method responses in typical density tool configurations.The basic measured output from logging tool detectors is the energy deposited in the detector crystal by each gammay-ray which passes through it. This output is known as an energy dependent Pulse Height Distribution. A model of this mechanism was required to be incorporated into the Monte Carlo method to enable direct cmparison of measurement and results calculated using a computer model.For any tool a prime characteristic of interest is the sensitivity of the detector output to changes in formation density. This should be as high as possible. A perturbation theory option was required for use in the McDUFF code.ii) Code ValidationThe validation programme provided a full range of benchmark data ranging from a single detector in air ranging through progressively more complicated configurations (in 7 stages) to a laboratory sonde in a perturbed borehole environment.iii) Study of Environmental EffectsThe computer model was used to predict the effects of variations in experimental boreholes including mudcake thickness, borehole size and shape etc. Rugosity was simulated in the test blocks by means of liners with castellated variations along the length of the borehole.iv) Demonstration Using Operational SondesThe computer code McDUFF is available to participating companies under preferential terms for a period of one year following completion of the project.v) Management and TimescaleThe work was undertaken within the Radiation Physics, Shielding and Criticality Group at Winfrith through AEA Petroleum Services over a period of two years from 1 February 1989. Programa(s) ENG-HYDROCARB 2C - Programme (EEC) of support for technological development in the hydrocarbons sector, 1986-1989 Tema(s) 1.6 - MISCELLANEOUS Convocatoria de propuestas Data not available Régimen de financiación DEM - Demonstration contracts Coordinador AEA Technology Aportación de la UE Sin datos Dirección Winfrith Technology Centre DT2 8DH Dorchester Reino Unido Ver en el mapa Coste total Sin datos
