| Three design models for predicting pulling loads during Horizontal Directional Drilling (HDD) installations (Driscopipe, Drillpath and Petroleum Research Council International (PRCI)) were evaluated, initially comparing their predictions with recorded rig loads. It was found that the models did not accurately predict the maximum pull load, or the loading trend as the pipe was drawn through the borehole. Real-time monitoring cell technology was subsequently developed for two purposes: (i) to gain a better understanding of pulling loads and mud pressures by making measurements downhole between the reamer and pipe pull head, and (ii) as a tool to enhance quality control and quality assurance (QC/QA) in HDD installations. A database of 19 commercial non-engineered installations, 15 monitored for pull load only, and 4 for combined pull load and mud pressure was compiled using a 220 kN capacity monitoring cell. The electronic load record provided by the monitoring cell can form the basis of a robust QC/QA program. Analysis of the load cell data resulted in a proposal for a new design approach that consists of empirical design charts for three categories of construction effort---low, medium and high, for four soil groups---gravely, sandy, silty and clayey soils. Most non-engineered installations can be successfully completed without employing a high degree of construction effort (i.e. without achieving a clean open borehole). The installations monitored for this thesis were assessed to fall predominantly in the medium construction effort-clayey soils category. Current analytical design models were developed for large-scale engineered crossings that are normally associated with a high degree of construction effort. To gather data in this critical area a large-scale monitoring cell was developed with 670 kN axial capacity and 1000 kPa pressure capacity. It is expected that this monitoring cell, employed on installations involving a high degree of construction effort, will provide the data needed to accurately assess current design models, and, together with laboratory experimentation, the validity of the design parameter values that are input into these models. It is also expected that the technology developed and the data collected during the course of this research project will lead to the development of more accurate models for both engineered and non-engineered HDD crossings. |
