Abstract

A detailed analysis of 341 internally gravel-packed zones in wells from three oil fields was carried out to determine which parameters significantly influence the placement of the gravel in such a completion and the well's resulting productivity.

The database analysis demonstrates that, in addition to the well-known impairing factors, such as dirty completion fluids and viscous fluid-loss control pills, the performance of gravel packs can be improved by increasing the number of perforations - especially in 7" completions. This suggests that the perforation tunnel is the most critical area in an internally gravel-packed completion.

Introduction

Detailed production testing at various stages of the completion phase of two wells was carried out in 1989 to determine when and how the performance of gravelpacked completions was being impaired. The tests suggested that the major contribution to the final impairment took place during the actual gravel packing. Figure 1 shows how the well inflow quality indicator (WIQI) of a well, which was more than 100% prior to gravel packing, was reduced to a mere 160/0 after gravel packing. Despite many efforts it has proved difficult to achieve a radical improvement in the performance of internally gravel packed wells.

For many years the data from well completions had been recorded in a database. This data has been analysed with the objective of gaining a better understanding of which completion parameters govern the successful placement of gravel in internally gravel-pack completions and the productivity of the thus completed wells. Some statistical information about the data base is given in Table 1. The distribution of the pack factor and WIQI are given in Figs. 2 and 3.

DATA BASE ANALYSIS

The correlation structure of the data base was estimated in terms of Spearman's rank correlation coefficients. For each pair of variables in the data, a statistic was calculated that quantifies the similarity of the two variables. The statistic used here is based on the rankings of the two variables alone and not on their values. This approach has the advantage of not being overinfluenced by the nonlinearity of the relationship between the variables.

In order to ensure that the calculated correlation coefficients are reliable, the significance of the variables' similarity is also calculated in terms of a p-value. A p-value below about 0.05 implies that a significant relationship exists between the two variables. (One should bear in mind that a significant statistical correlation between two variables does not imply that there is a causal relationship between them. The causality can be justified only on the basis of extraneous, physical knowledge.)

P. 369

This content is only available via PDF.
You can access this article if you purchase or spend a download.