{"id":{"repo_id":"rgu","oai_identifier":"oai:rgu-repository.worktribe.com:3235199"},"canonical_url":"https://search.dev.ndltd.org/etd/rgu/oai:rgu-repository.worktribe.com:3235199","repository":{"repo_id":"rgu","name":"Robert Gordon University","base_url":"https://rgu-repository.worktribe.com/oaiprovider"},"display":{"title":"A drilling forensic framework and algorithm for the analysis and diagnoses of drilling dysfunction.","abstract":"In the oil and gas industry, drilling wells can be time-consuming, largely influenced by the rate of penetration of the drill bit. Drilling dysfunctions result in wasted energy, reduced penetration rates, and potential damage to the bottom hole assembly or drill bit which can significantly impact overall drilling performance, costs, and emissions. These dysfunctions often stem from bit-rock interactions or drill string vibrations. The recent availability of digital data and real-time information has enabled instantaneous measurement of drilling parameters and loads, enhancing forensic assessments of drilling dysfunctions. However, consistent approaches to drilling dysfunction identification at the rig site are still lacking. Over the past decade, machine learning algorithms have been developed to enhance drilling performance, yet there are limited applications of artificial intelligence for real-time drilling dysfunction identification. This thesis aims to improve drilling performance by evaluating drilling data and developing an advanced forensic workflow to diagnose drilling dysfunctions. In this study, a workflow for drilling forensics was created, identifying the necessary information for effective assessments. Data from five wellbores were used in a case study to test this methodology, applied actively during the drilling of these wells. The raw data files were reprocessed using physics-based drilling mechanics equations, and a comparative analysis was conducted to correlate variables linked to drilling dysfunctions. A novel algorithm was developed for post-run analysis, detecting and categorizing drilling dysfunctions, and verified through manual post-run analysis. The impacts of drilling dysfunctions on machine learning capabilities were then analysed. The rate of penetration was predicted using the full data set and compared to predictions from a data set with dysfunctions removed, highlighting the effects of removing dysfunctions prior to machine learning training. The novelty of this study lies in the proposed drilling dysfunction forensics framework and the unique algorithm that combines physics-based and machine learning models to create an automated driller's roadmap. This includes the innovative use of resistivity to detect interfacial severity. Additionally, the study provides new insights into the impacts of drilling dysfunctions on machine learning predictions. Following the execution of the five 17.5-inch hole sections analysed in this study, implementing practices to tackle drilling dysfunctions resulted in the well campaign being delivered within 309 days, 79 days ahead of plan, with 40 of these days attributable to the performance drilling aspects described in this thesis. The approach for generating the algorithm demonstrated that utilizing this novel process can create an automated, integrated driller's roadmap in a fraction of the time required for manual methods. The machine learning models created showed a coefficient of determination ranging from 0.94 to 0.98 on training data and 0.60 to 0.86 on test data, indicating a favourable overall model fit. The forensics framework introduced in this study has formed the basis of a supporting industry guideline as part of an ongoing global effort by the International Association of Drilling Contractors (IADC) to upgrade the current drill bit and bottom hole assembly dull code manual. This study also aligns with 10 out of the 17 United Nations Sustainable Development Goals, promoting a sustainable future.","abstract_html":"In the oil and gas industry, drilling wells can be time-consuming, largely influenced by the rate of penetration of the drill bit. Drilling dysfunctions result in wasted energy, reduced penetration rates, and potential damage to the bottom hole assembly or drill bit which can significantly impact overall drilling performance, costs, and emissions. These dysfunctions often stem from bit-rock interactions or drill string vibrations. The recent availability of digital data and real-time information has enabled instantaneous measurement of drilling parameters and loads, enhancing forensic assessments of drilling dysfunctions. However, consistent approaches to drilling dysfunction identification at the rig site are still lacking. Over the past decade, machine learning algorithms have been developed to enhance drilling performance, yet there are limited applications of artificial intelligence for real-time drilling dysfunction identification. This thesis aims to improve drilling performance by evaluating drilling data and developing an advanced forensic workflow to diagnose drilling dysfunctions. In this study, a workflow for drilling forensics was created, identifying the necessary information for effective assessments. Data from five wellbores were used in a case study to test this methodology, applied actively during the drilling of these wells. The raw data files were reprocessed using physics-based drilling mechanics equations, and a comparative analysis was conducted to correlate variables linked to drilling dysfunctions. A novel algorithm was developed for post-run analysis, detecting and categorizing drilling dysfunctions, and verified through manual post-run analysis. The impacts of drilling dysfunctions on machine learning capabilities were then analysed. The rate of penetration was predicted using the full data set and compared to predictions from a data set with dysfunctions removed, highlighting the effects of removing dysfunctions prior to machine learning training. The novelty of this study lies in the proposed drilling dysfunction forensics framework and the unique algorithm that combines physics-based and machine learning models to create an automated driller's roadmap. This includes the innovative use of resistivity to detect interfacial severity. Additionally, the study provides new insights into the impacts of drilling dysfunctions on machine learning predictions. Following the execution of the five 17.5-inch hole sections analysed in this study, implementing practices to tackle drilling dysfunctions resulted in the well campaign being delivered within 309 days, 79 days ahead of plan, with 40 of these days attributable to the performance drilling aspects described in this thesis. The approach for generating the algorithm demonstrated that utilizing this novel process can create an automated, integrated driller's roadmap in a fraction of the time required for manual methods. The machine learning models created showed a coefficient of determination ranging from 0.94 to 0.98 on training data and 0.60 to 0.86 on test data, indicating a favourable overall model fit. The forensics framework introduced in this study has formed the basis of a supporting industry guideline as part of an ongoing global effort by the International Association of Drilling Contractors (IADC) to upgrade the current drill bit and bottom hole assembly dull code manual. This study also aligns with 10 out of the 17 United Nations Sustainable Development Goals, promoting a sustainable future.","abstract_has_math":false,"creators":["Watson, William"],"institution":"Robert Gordon University","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["M. Amish and G. Oluyemi"],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025","date_published":"2025","updated_at":"2026-07-24T04:10:12Z","subjects":["Drilling dysfunctions","Rate of penetration","Drilling forensics","Machine learning","Bit–rock interaction","Drill string vibrations","Real‑time drilling data"],"languages":["en"],"rights":[],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["oai:rgu-repository.worktribe.com:3235199","https://doi.org/10.48526/rgu-wt-3235199"],"render_values":[{"text":"oai:rgu-repository.worktribe.com:3235199","href":null,"code":true},{"text":"https://doi.org/10.48526/rgu-wt-3235199","href":"https://doi.org/10.48526/rgu-wt-3235199","code":true}]}]},"links":{"outbound_url":"https://rgu-repository.worktribe.com/3235199/1/WATSON%202025%20A%20drilling%20forensic%20framework.","outbound_label":"Repository record","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["M. Amish and G. Oluyemi"]},{"key":"dc:contributor.sponsor","label":"Sponsor","values":["No Funder Acknowledged (Outputs)"]},{"key":"dc:creator","label":"Author","values":["Watson, William"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2025-04-30"]},{"key":"dc:date.issued","label":"Date","values":["2025"]},{"key":"dc:publisher.institution","label":"Dc Publisher Institution","values":["Robert Gordon University"]},{"key":"dc:relation.isreferencedby","label":"Dc Relation Isreferencedby","values":["https://rgu-repository.worktribe.com/output/3235199"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Drilling dysfunctions","Rate of penetration","Drilling forensics","Machine learning","Bit–rock interaction","Drill string vibrations","Real‑time drilling data"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["oai:rgu-repository.worktribe.com:3235199","https://doi.org/10.48526/rgu-wt-3235199"]},{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://rgu-repository.worktribe.com/3235199/1/WATSON%202025%20A%20drilling%20forensic%20framework."]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["In the oil and gas industry, drilling wells can be time-consuming, largely influenced by the rate of penetration of the drill bit. Drilling dysfunctions result in wasted energy, reduced penetration rates, and potential damage to the bottom hole assembly or drill bit which can significantly impact overall drilling performance, costs, and emissions. These dysfunctions often stem from bit-rock interactions or drill string vibrations. The recent availability of digital data and real-time information has enabled instantaneous measurement of drilling parameters and loads, enhancing forensic assessments of drilling dysfunctions. However, consistent approaches to drilling dysfunction identification at the rig site are still lacking. Over the past decade, machine learning algorithms have been developed to enhance drilling performance, yet there are limited applications of artificial intelligence for real-time drilling dysfunction identification. This thesis aims to improve drilling performance by evaluating drilling data and developing an advanced forensic workflow to diagnose drilling dysfunctions. In this study, a workflow for drilling forensics was created, identifying the necessary information for effective assessments. Data from five wellbores were used in a case study to test this methodology, applied actively during the drilling of these wells. The raw data files were reprocessed using physics-based drilling mechanics equations, and a comparative analysis was conducted to correlate variables linked to drilling dysfunctions. A novel algorithm was developed for post-run analysis, detecting and categorizing drilling dysfunctions, and verified through manual post-run analysis. The impacts of drilling dysfunctions on machine learning capabilities were then analysed. The rate of penetration was predicted using the full data set and compared to predictions from a data set with dysfunctions removed, highlighting the effects of removing dysfunctions prior to machine learning training. The novelty of this study lies in the proposed drilling dysfunction forensics framework and the unique algorithm that combines physics-based and machine learning models to create an automated driller's roadmap. This includes the innovative use of resistivity to detect interfacial severity. Additionally, the study provides new insights into the impacts of drilling dysfunctions on machine learning predictions. Following the execution of the five 17.5-inch hole sections analysed in this study, implementing practices to tackle drilling dysfunctions resulted in the well campaign being delivered within 309 days, 79 days ahead of plan, with 40 of these days attributable to the performance drilling aspects described in this thesis. The approach for generating the algorithm demonstrated that utilizing this novel process can create an automated, integrated driller's roadmap in a fraction of the time required for manual methods. The machine learning models created showed a coefficient of determination ranging from 0.94 to 0.98 on training data and 0.60 to 0.86 on test data, indicating a favourable overall model fit. The forensics framework introduced in this study has formed the basis of a supporting industry guideline as part of an ongoing global effort by the International Association of Drilling Contractors (IADC) to upgrade the current drill bit and bottom hole assembly dull code manual. This study also aligns with 10 out of the 17 United Nations Sustainable Development Goals, promoting a sustainable future."]},{"key":"dc:title","label":"Title","values":["A drilling forensic framework and algorithm for the analysis and diagnoses of drilling dysfunction."]}]}],"canonical_facts":{"dc:contributor.advisor":["M. Amish and G. Oluyemi"],"dc:contributor.sponsor":["No Funder Acknowledged (Outputs)"],"dc:creator":["Watson, William"],"dc:date":["2025-04-30"],"dc:date.issued":["2025"],"dc:description.abstract":["In the oil and gas industry, drilling wells can be time-consuming, largely influenced by the rate of penetration of the drill bit. Drilling dysfunctions result in wasted energy, reduced penetration rates, and potential damage to the bottom hole assembly or drill bit which can significantly impact overall drilling performance, costs, and emissions. These dysfunctions often stem from bit-rock interactions or drill string vibrations. The recent availability of digital data and real-time information has enabled instantaneous measurement of drilling parameters and loads, enhancing forensic assessments of drilling dysfunctions. However, consistent approaches to drilling dysfunction identification at the rig site are still lacking. Over the past decade, machine learning algorithms have been developed to enhance drilling performance, yet there are limited applications of artificial intelligence for real-time drilling dysfunction identification. This thesis aims to improve drilling performance by evaluating drilling data and developing an advanced forensic workflow to diagnose drilling dysfunctions. In this study, a workflow for drilling forensics was created, identifying the necessary information for effective assessments. Data from five wellbores were used in a case study to test this methodology, applied actively during the drilling of these wells. The raw data files were reprocessed using physics-based drilling mechanics equations, and a comparative analysis was conducted to correlate variables linked to drilling dysfunctions. A novel algorithm was developed for post-run analysis, detecting and categorizing drilling dysfunctions, and verified through manual post-run analysis. The impacts of drilling dysfunctions on machine learning capabilities were then analysed. The rate of penetration was predicted using the full data set and compared to predictions from a data set with dysfunctions removed, highlighting the effects of removing dysfunctions prior to machine learning training. The novelty of this study lies in the proposed drilling dysfunction forensics framework and the unique algorithm that combines physics-based and machine learning models to create an automated driller's roadmap. This includes the innovative use of resistivity to detect interfacial severity. Additionally, the study provides new insights into the impacts of drilling dysfunctions on machine learning predictions. Following the execution of the five 17.5-inch hole sections analysed in this study, implementing practices to tackle drilling dysfunctions resulted in the well campaign being delivered within 309 days, 79 days ahead of plan, with 40 of these days attributable to the performance drilling aspects described in this thesis. The approach for generating the algorithm demonstrated that utilizing this novel process can create an automated, integrated driller's roadmap in a fraction of the time required for manual methods. The machine learning models created showed a coefficient of determination ranging from 0.94 to 0.98 on training data and 0.60 to 0.86 on test data, indicating a favourable overall model fit. The forensics framework introduced in this study has formed the basis of a supporting industry guideline as part of an ongoing global effort by the International Association of Drilling Contractors (IADC) to upgrade the current drill bit and bottom hole assembly dull code manual. This study also aligns with 10 out of the 17 United Nations Sustainable Development Goals, promoting a sustainable future."],"dc:identifier":["oai:rgu-repository.worktribe.com:3235199","https://doi.org/10.48526/rgu-wt-3235199"],"dc:identifier.uri":["https://rgu-repository.worktribe.com/3235199/1/WATSON%202025%20A%20drilling%20forensic%20framework."],"dc:language":["en"],"dc:publisher.institution":["Robert Gordon University"],"dc:relation.isreferencedby":["https://rgu-repository.worktribe.com/output/3235199"],"dc:subject":["Drilling dysfunctions","Rate of penetration","Drilling forensics","Machine learning","Bit–rock interaction","Drill string vibrations","Real‑time drilling data"],"dc:title":["A drilling forensic framework and algorithm for the analysis and diagnoses of drilling dysfunction."],"dc:type":["Thesis"]},"updated_at":"2026-07-24T04:10:12Z"}