SPE 103211 Automated Reporting Using Rig Sensor Data Enables Superior Drilling Project Management

G. Thonhauser, University of Leoben, W. Mathis, TDE Thonhauser Data Engineering GmbH, both SPE

Copyright 2006, Society of Petroleum Engineers This paper was prepared for presentation at the 2006 SPE Annual Technical Conference and Exhibition held in San Antonio, Texas, U.S.A., 2427 September 2006. This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract An integrated process management system is the key to success of an enterprise working in the exploration business. Classic project management consists of the creation of an expectation, the well plan, appropriate process monitoring and the post analysis of the project through the comparison of the expectation and the actual process. As a result of the post analysis lessons learned are compiled. The quality of this traditional project management is strongly dependent on information flow. State-of-the-art reporting on a drilling rig is recurrent human reporting such as the classic morning report, although there are a lot more sources of information available. The presented concept of automated reporting, away from human observations, is ensured by utilizing rig sensor data. The automated reporting provides management, engineering and operations with the level of detail and granularity of objective, high quality information they need. It also offers the possibility of knowledge management and exchange at the same organizational level providing different views of the same project. This means that operations receive information on detailed process parameters, engineering on essential design variables for future wells and management on cost. Along with traditional project management this new automated reporting enables continuous, real-time project tracking and analysis even with a declining workforce because all information is bundled and under central control. Automated reporting is able to cover the life-cycle of every well construction and allows the establishment of high quality benchmarks for future planning and project evaluation in realtime. Moreover, it enhances the learning and experience level of every single person involved in the process. Finally, it offers the possibility to manage multiple concurrent projects such as a companys fleet of drilling rigs to be able to optimize resource allocation and project economics.

Introduction The successful construction and operation of oil and gas wells depends to a high degree on the experience of the people involved and their evaluation and management of more or less unknown factors (e.g. geology). The aim of this paper is to outline a concept which provides a platform that allows the optimal use of the knowledge of an organization and to support all decisions and actions during the life-cycle of a well construction project as much as possible. Providing the possibility of a continuous and real-time comparison of plan and actual at the proper technical and commercial level of detail is the key for a successful project execution. All data collected for the purpose of analysis is completely integrated into a domain model. The result is the description of the Drilling Process over time. The key element is a high level of automation to increase efficiency and improve the quality of conventional morning reporting, e.g. using rig sensor data. The derivation of a highly accurate drilling operations plan forms the basis for further performance improvement. Existing plans are often a self-fulfilling prophecy in terms of performance. The current level of reporting fulfills its purpose as an administrative tool, but is not the ideal basis for the planning of future wells. Another point is the distribution of the right data at the proper level of detail to different organizational levels. The requirements vary from the driller, who essentially needs information about the use of the brake-handle of the drawworks, to upper management who want overall cost and time performance. In order to fulfill these requirements a new way of reporting, processing of data and analysis, together with visualization has to be developed. The challenge is the integration of subjective experience, the measured state of the system and the financial aspects of a project. A key element is to increase the usability of information with less data entry effort. Data quality management is also of central importance. Management Concept for Well Construction Projects The concept discussed in the paper focuses on the application of a novel project management strategy as an integral approach considering data management, technical and cost aspects. The scope of this concept was already published by Thonhauser [5]. Summarizing, the key elements of this concept are defined as follows:

SPE 103211

Expectation (Well Plan) It is proposed to extend the concept of the conventional well plan to providing targeted information to all organizational levels (from drillers on-the-brake to managers). As part of this work it was recognized that it was not the classical well plan that determined the execution of a well construction project, but rather the expectation of the individual. In other words, whatever the driller does on the brake will be the result of the plan and his experience. As a consequence, the success of a project will depend on the quality of the plan and the experience of the personnel executing the plan. The conventional well plan, especially in terms of operations planning, is primarily focused on providing a time versus depth curve and a cost estimate. A more detailed plan, such as drilling the well on paper, is typically not generated due to time constraints and a lack of available personnel. Furthermore, benchmarks can only be integrated at a very coarse level, such as rate-of-penetration expectation per hole phase. The result of such a conventional plan is sufficient to administrate a well, but not to improve the performance. To improve performance the driller would have to know weighton-bit and bit rotational speed recommendations for a certain bit per geological layer etc. If we consider the fact that the well is drilled by the driller and that he/she is the key enabler of higher performance, we have to generate a plan for the driller, not for the engineer who works based on morning reports. The concept discussed in this paper suggests automating this task, so that a rigorous plan can be developed for any well. The conventional Technical Drilling Program is replaced by a project model, which combines project scope with the required resources and processes. The result is a project plan, which combines the well construction (with elements such as open holes, casings, cement, perforations etc.) with the process description (tasks) and the required resources (costs). The generation of this plan is supported by utilizing the results of the monitoring proposed as part of this concept (automated reporting and benchmarking). The successful generation of the model is dependent on the seamless integration of subsurface model (geology), the subsurface well construction (the well) and the well construction process (sequence of tasks to be performed) and the resources used (cost and logistics). This integration is a challenge as all domains have a different View of things, but still use the same information. Monitoring The monitoring component is responsible for the description of the actual state of the system. This may be split into several components: Structured Data This set of information defines all technical aspects of the resources used for the project. It includes information such as well construction elements, trajectory, drill string data, geological data, cost etc. Measurement Data This kind of information addresses all data which can

be defined as a data stream, e.g. time-based sensor data or depth-based geophysical well logs. Human Observation This set of data is the description of the processes and information about the state of resources from the view of the people reporting.

The monitoring has the greatest potential for improvement. Utilizing conventional reporting in combination with automated operation-recognition substantially increases the usability of the process description. Tracking Tracking is defined as the comparison of Plan or Expectation and Monitoring data. In conventional systems it is hard, if not impossible to determine deviations between planned and actual to the required level of detail. Typically, this comparison is performed on the basis of overall project budget or a small number of cost groups. An additional challenge is the separation of accurate cost information in accounting systems and the drilling process information. On the basis of planning and monitoring the presented system allows the generation of the following tracking output, which is provided to all stakeholders in the proper level of detail: Deviation in time (the project or individual tasks took longer than planned) Deviation in cost (the resources were more expensive than expected, or more resources had to be used) Deviation in project scope (e.g. a sidetrack had to be drilled) Key Enablers for the Project Management Concept The Data Model The first key feature is that the same model is used to store and view the data of the expectation (well plan) and the monitoring part (reporting). If we plan at a phase level and monitor on automated operation recognition results (with very detailed information) a comparison does not make sense. This is conventionally solved by lumping information together (morning report), so that the monitoring matches the plan rather than doing the opposite. If we want to manage a project for performance improvement and not only administrate it, we have to make sure that plan and monitoring is performed at the same level of detail. Doing so, two versions of the same project (a planned, and an actual) can be compared as we progress through the project. Both versions are stored as different scenarios. A project can have multiple scenarios like: Planned Scenario (Alternative 1) Planned Scenario (Alternative 2) Planned Scenario (Alternative 3) Actual Scenario (Monitoring Scenario) In the planning and decision phase it is possible to directly compare the different alternatives by selecting all three planned scenarios. Once an alternative is determined to be executed the user can define it as approved plan. From then on this plan is compared to the scenario holding the reporting data

SPE 103211

(monitoring). An example is shown in Figure 2. This chart displays time versus depth and the cost versus time curves of planned and actual scenarios. It can be added to the daily report, so tracking of the construction progress and cost developments can be done on a regular basis. In a very similar way this allows the possibility of directly comparing scenarios (monitoring) of different wells. This can be very useful if a new well has to be drilled and several offset wells are available. In this case the actual scenarios of all existing offset wells would be displayed to be able to compare them. Data Collection Process The basis for any analysis is consistent and accurate data. It has to be emphasized that there is a big difference between collecting data for administration or for performance improvement purposes. Here it is assumed that data is collected to improve the drilling process and to eliminate hidden lost time etc. To achieve the above mentioned quality, the current data flow on the rig and further to the office has to be redesigned. Being more precise, the data flow from the original data provider (service company) to the operating company (company man) has to be optimized. Several features are the basis for this (see Figure 1): The focus of data collection should be on objective measurements rather than subjective observations. Measurement data should play an important role in any analysis. Standards have to be defined in terms of contents and format of data supplied by service companies. Collecting useless data should be avoided, e.g., an average weight on bit per day or complete BHA run does not provide any useful information. The same is true of using a daily footage as basis for a rate of penetration over a 24 hour interval. The focus of the human data entry should be to describe differences between expected behavior (planned) and actual. This should be done as structured Feedback. The work of the company man has to be shifted from retyping data to quality controlling and providing value-added Feedback. A large part of the quality control can be automated. All data has to be carefully integrated in depth and time as the basis for analysis. A central data manager (broker) should be available with clearly defined data flows and standards. He must take responsibility for the data collection process and quality. The data manager, as a third party, may monitor the collection process and data quality. This auditory function is to monitor data quality and deliver feedback. The objective is to continuously improve the data collection process. As an example the mudlogging service, as one of the data providers on the rig, is discussed. The mudlogger is variously

a consumer of data from service companies and a provider of information to others. Subsequently, the data of the mud logger is directly transmitted to the operators system, rather than printing out a report from the mudlogging system, giving this report to the company man, who then enters this information into the reporting system. The company mans duty should ideally only be to check the entered data for correctness. In a similar way measured data coming from sensors must be quality controlled. This can not be done manually since the amount of data is in the range of several hundred thousands of data points per day. Measured data quality control includes: Outlier removal Checks that the data channels are in the correct range Logical checks, e.g. if the bit is moving, the block must move as well. Or the block velocity itself has technical limits and can not exceed certain values If bad data quality is recognized by the software an alarm is sent to the data manager who can take the appropriate actions to correct the measurements, e.g. let the mud logger do a sensor recalibration. Well defined data flows combined with rigorous quality management will result in a consistent and accurate data set, which can then be accessed throughout the whole company. The technical and economical value of this process is demonstrated by Smith [7]. Data Accessibility and Visualization As soon as the data is quality controlled and stored correctly the next very critical point has to be addressed. Many systems struggle with the amount of data accumulating during a well construction project. Real-time measured data is hardly ever used to its full potential, or even neglected, whereas most systems are very comfortable to hold and access reporting data. The new concept permits the displaying and printing of structured data, which is nothing new since this is the objective of all reporting systems. The key difference is the possibility to literally surf the huge amounts, often several gigabytes, of measurement data. This is possible in the timedomain as well as in the depth-domain. Additionally, one can directly compare the data of different wells. This ability is a novel concept, which helps to extract added value from the data. Other benefits were reported by Holland and Oberwinkler [8, 9]. Further improvements by applying the already mentioned store data only once strategy implies that the same data is presented in different ways depending on the objective: Technical planning or analysis: In this case the drilling engineer or geologist wants to see the data depth related, either MD or TVD Process analysis and performance analysis: In this case the drilling engineer wants to see the data timerelated Another strategy is to apply the 4 rights (the right information to the right people in the right time in the right context) whenever data is displayed. A simple example on

SPE 103211

probably the most misused parameter, ROP, clearly shows the importance of this matter. ROP is used for many different applications, mostly but not exclusively performance related: Bit performance analysis Daily drilling performance analysis Planning a whole hole phase If the objective is to analyze bit performance, the instant ROP is used (data has to come from measurements), to produce meaningful results. This ROP includes only the time where the bit was really drilling on bottom. In the second case, with information typically coming from morning reports, also the wellbore treatment time, the off-bottom time, and the tripping time is typically included to calculate the average ROP. For the last case it is important to know which nonproductive time operations are also included in the time used to determine phase ROP (logging, casing, cementing, and BOP work). All applications are based on a ROP, but the value is different in each case. Clearly, selecting data applicable to the analysis objective is necessary, e.g. serious bit analysis is not possible from morning report data but requires measurement data. Use of Real-Time Measurements As mentioned earlier, the use of time based sensor data is a mandatory step in generating an objective description of the drilling process. By using automated operations recognition it is possible to filter information from gigabytes of data, which are normally only inspected visually by a human and not used further. The approach presented by Thonhauser [3] uses a few measurements commonly gathered by the mud logging unit to extract a highly accurate and standardized description of the well construction process. Figure 3 shows the raw data of one day. This data was processed by the automated operation recognition; the results are visualized in Figure 4. The chart shows one day where each column represents one hour of chronological data from bottom to top (left axis with scale in minutes). The black and gray lines represent the hole MD and the bit MD over time. Just by looking at the raw data of Figure 3 experienced drilling engineers get a clue as to what happened that day, although this is very cumbersome. Whats obvious is that the second half of the day was spent pulling out of the hole (POOH). The first part of that day is far more complicated. It takes a lot of time and effort to read the data correctly. But having Figure 4 available changes this completely. Once accustomed to the color codes this chart is very easy to read. Knowing that the green colored areas represent ream and wash, one can identify where and when these operations happened right away. The visualization strategy taps into a psychological aspect of the human brain: colors and charts can be understood far better than tables with codes. This is the case for the majority of people, especially for engineers who are used to graphical representation of data. The same colors are applied throughout the whole system. So every time green can be identified, this time interval can be associated with ream and wash. Figure 5 shows a chart of the whole BHA Run including the previous day. Again, the same color codes apply,

so identifying all ream and wash operations is not difficult any more. Real-Time Tracking The beauty of this system again is that these automatically generated reports can be compared to the planned process simply by comparing two scenarios. One example is shown in Figure 4 where the Planned (in blue) and the Best Performance (in green) are Tracked against the actual performed Trip Out (in grey). The differences between the three pull out of hole processes are obvious. Lets take a closer look and do a sample calculation. In Table 2 the facts of the real POOH process are displayed in the column Actual Scenario. The data is from the Trip Out section of the BHA Run report (see Figure 5 and 6). The reason that the number of connections in Figure 6 deviates from the Planned Scenario in Figure 2 is that connections longer than 15 minutes are not considered in this report. However, the final result is virtually unaffected. The first calculation compares the Actual Scenario to the Planned Scenario (displayed in blue). To calculate the planned trip time from 4210m to 302m the number of connections was multiplied with the average connection duration and added to the necessary running time for the 3908m distance. This calculated time of 09:10 hours compared to the actual duration of 11:05 hours results in a difference of about 2 hours. This difference expressed in cost on a land rig with 50,000$/d this is equivalent to about 4,000$. The second calculation compares the Actual Scenario to the Best Historic Performance (displayed in green) of this rig or crew. The best historic performance shows an average connection time of 2.5 minutes instead of the 2.9 minutes observed in Trip Out operations. The other conditions remain the same. For this case the difference is close to 3 hours, worth about 6,000$. If this example related not to a land but an offshore rig, these numbers would be 40,000$ and 60,000$ assuming a day rate of 500,000$/d. Importantly, this analysis illustrates the improvement possible in a single trip out of hole. This shows the tremendous potential of this concept. More information about automated operation recognition reports was presented by Thonhauser [6]. Automated Reporting Reporting, with a focus on the process description in this case, is done by every operator. As most companies use this process description as the basis to plan future wells every task is classified with operations codes to enable automated analysis. This is supported by more or less all reporting systems in different ways. Some companies already use this to do extensive performance analysis and process optimization as presented by Iyoho [1] and Adeleye [2]. However this process has some substantial drawbacks: Human reporting is very subjective A lot of training is necessary to get trustful and objective data Considerable effort is required to quality control the data A very well defined classification system is needed and must be rigorously applied

SPE 103211

To overcome these disadvantages the industry has for a long time demanded better use of the measured data (as presented by Irrgang [4]). Thus, the results of the automated operation recognition are used to enable automated reporting. However, the results can not be used directly. With a typical number of 100,000 recognized operations per month this would flood existing analysis tools where one percent of this number is typical. The processing time would be increased by a factor of 100, presupposing that the system could handle the amount of data. The automated reporting is an algorithm that lumps together the high detailed information into tasks with lower granularity. The level of detail is configurable, a typical time of 5 minutes turned out to be very efficient. Table 1 shows the results for the same day of the previous examples (Figures 3 and 4). All information in this table is found automatically. Most of it can also be found in most of the common daily drilling reports. But one parameter, the tool MD, is introduced with the new concept. Looking at the operation description of the task reveals no big surprises, a text like Circulation (at 3717m) can be found in any report. However the two columns Start and End tool MD make the big difference. This enables not only the determination of how much time was spent on a particular operation, but also the depth where it happened. Importantly, the information is not only present in the text but also in listed form. This enables automatic analysis. A detailed example follows. Benchmarking The fully automated concept allows the introduction of benchmarks which would not be feasible otherwise. Some parameters do have a tremendous influence to the total performance of a rig or crew: Connection time during tripping Off bottom time during drilling Wellbore treatment time as total time or along hole depth Bit ROP performance depending on operating parameters (flow rate, rpm, WOB) and geology In one specific analysis the connection times of 12 wells drilled by 3 different rigs were compared. More than 200,000 single connections were analyzed with two main results: The average connection time per rig over time was increasing. The reason for that was the replacement of experienced with less experience staff. This tendency was known, however the financial impact was tremendously underestimated The saving potential if it would be possible to reach average performance where above average performance was in the order of $10 million, nearly one million per well. The second example is the wellbore treatment time. The ability to detect duration, time and location of operations opens totally new possibilities to identify hidden lost time. Figure 7 shows the so-called wellbore treatment operations along hole depth. The treatment time consists of circulation,

reaming and washing time. In this chart only the operations performed during the BHA run in Figure 5 are displayed. In Figure 8 the wellbore treatment times from spud until the end of the mentioned BHA Run is visualized. The dashed red horizontal lines mark the casing shoe position. At the casing shoes wellbore treatment is intended to clean the hole prior to casing and cementing. Not so obvious are the peaks marked by the three red shaded zones. These peaks mark potential trouble formations since increased wellbore treatment time was necessary in these intervals. These peaks are the result of accumulated wellbore treatment every time the bit passed this zone. So even if the driller recognized problems in these formations he might not find it worth reporting the 15 or so minutes of reaming, washing or circulation. But the treatment time accumulates with each trip. This is a very good indicator for trouble formations. With this indication from the operation time statistics a closer look or a technical analysis might reveal problems which can be prevented with appropriate planning. This can save time and increase performance for the next wells being drilled in this area. Figure 9 shows the same analysis for rotary, sliding, total drilling and total wellbore treatment time. The drilling curves are basically the inverse ROP. In the form of time spent per depth interval this can be compared with the treatment time. The important thing to note is that all these analyses and performance evaluations are easily done because they are integrated into a highly automated workflow. Drilling Process Coding Recognizing the fact that the BHA run is a key element of the well construction process the coding system has a special class of tasks for tubular assembly runs. A hole phase is often planned as a series of BHA runs, e.g. 2 drilling runs, one hole conditioning run and a casing run. However, hardly any commercial reporting product is able to specify BHA runs directly in the process description. So it is not possible to compare planned versus actual runs later on. The extended operations classification system includes: Phases Jobs Runs Main Operations Operations The system includes a concept to cater for the reporting of trouble time. In the presented concept this is done as Feedback. Feedback is given, whenever a human observer recognizes differences between expectation and actual. Feedback is reported in a structured way defining: Symptoms Diagnoses Action/Recommendation These observations do not only give negative feedback (problem with bit XY in soft formation YZ) but also positive (exceptional high ROP with bit XY in formation YZ). The word Feedback communicates a more positive meaning than Trouble with the target to let the individual person know

SPE 103211

that feedback is a positive thing, and that providing feedback is something especially well received by the team. When used constructively positive or negative Feedback is generally better received by people and more conducive to a learning experience. Knowledge management is a big issue in the oil industry since the top-heavy age pyramid is a very well known problem (further reading: Hansen [10], Seely Brown [11], and Tealdi [12]). Conclusions The following conclusion can be drawn from the concepts presented in this paper: Using measurements as the basis to generate drilling operations reports leads to a standardized and objective description of the drilling process. Measurement derived data can be integrated into conventional reporting to generate an integrated data set in both time and depth domains Automated reporting provides improved monitoring capabilities with reduced effort at the same time. All users can rely on the quality of the data and access their useful view of the data (operational staff, engineers and managers) The financial impact lies in improved performance (reduce time to execute a project) as well as in reduced time to administrate the project (search for data, quality control, etc.) Staff can spend more time on their primary objectives (like giving feedback when the process deviates from the plan or simply spend less time to access the necessary information for their work). This also applies for to all organizational levels, from the company man on the rig all the way up to management Taking advantage of the automated operation recognition and reporting does not increase costs since the analysis relies on a few commonly recorded measurements of the mud logger. Available measurement data can be loaded, processed and instantly utilized in performance analyses and benchmarking. References
1. Iyoho, A.W., Meize, R.A., and K.K. Millheim, Lessons From Integrated Analysis of GOM Drilling Performance, SPE 97464 presented at the 2004 Offshore Technology Conference, Houston (USA), May 2004 2. Adeleye, A.R., Virgnillo, B.K., Iyoho, A.W., Parenteau, K., and Licis, H., Improving Drilling Performance Through Analysis of Historical Data: Case Study of a Canadian Field, SPE/IADC 87177 presented at the 2004 SPE/IADC Drilling Conference, Dallas (USA), March 2004 3. Thonhauser, G., Using Real-Time Data for Automated Drilling Performance Analysis, OIL GAS European Magazine, Edition 4, 2004, pages 170 173 4. Irrgang, R., Kravis, S., Nakagawa, E., Drilling Knowledge Management, What is Missing and Can We Fix It?, IADC/SPE 77249 presented at the 2002 Asia Pacific Drilling Technology Conference, Jakarta (Indonesia), September 2002 5. Thonhauser, G., Wallnoefer, G., Mathis, W., Well Construction Project Tracking to Improve Economics of E&P Activities,

presented at the DGMK/GEW Spring Conference for Exploration and Production, Celle (Germany), April 2005 6. Thonhauser, G., Wallnoefer, G., Mathis, W., Use of Real-Time Rig Sensor Data to Improve Daily Drilling Reporting, Benchmarking and Planning - A Case Study, SPE 99880 presented at the 2006 SPE Intelligent Energy Conference, Amsterdam (The Netherlands), April 2006 7. Smith, A.H., The Economic Advantages of Managing Data, ONCE!, SPE 78337 presented at the 13th European Petroleum Conference in Aberdeen (United Kingdom), October 2002 8. Oberwinkler C., Holland, J., Mayfield, D., Dixon, D., Real-Time Information to the Engineers Desktop How Murphy E&P Improved Production Management, SPE 96645 presented at the 2005 Offshore Europe Conference, Aberdeen (United Kingdom), September 2005 9. Holland, J., Oberwinkler, C., Huber, M., Zangl, G., Utilizing the Value of Continuously Measured Data, SPE 90404 presented at the 2004 Annual Technical Conference, Houston (USA), September 2004 10. Hansen, M.T.; Nitrin, N.; Tierney, T., Whats Your Strategy for Managing Knowledge, Harvard Business Review, Vol. 77, March-April 1999, pages 106 116 11. Seely Brown, J. : Duguid, P., Balancing Act, How to Capture Knowledge without Killing it!, Harvard Business Review, Vol. 78, May-June 2000, pages 73 80 12. Tealdi, L., Kreft, E., Donachie, J.M., Developing Young Exploration and Production Professionals To Solve the Big Crew Change, SPE 99924 presented at the 2006 SPE Intelligent Energy Conference, Amsterdam (The Netherlands), April 2006

SPE 103211

Figure 1: Project Management Concept Data Model

Figure 2: TxDxC (Time versus Depth and Costs) of Planned and Actual Scenario

SPE 103211

Figure 3: Raw Data

Nr 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Start MD [m] 4209.23 4209.23 4209.23 4209.23 4209.23 4210.43 4210.43 4211.06 4211.06 4211.06 4211.08 4211.63 4211.64 4215.97 4215.97 4216.51 4216.56 4216.99 4216.99 4217.00 4217.00 4217.00 4217.00 4217.00 4217.00 4217.00 4217.00

End MD [m] 4209.23 4209.23 4209.23 4209.23 4210.43 4210.43 4211.06 4211.06 4211.06 4211.08 4211.63 4211.64 4215.97 4215.97 4216.51 4216.56 4216.99 4216.99 4217.00 4217.00 4217.00 4217.00 4217.00 4217.00 4217.00 4217.00 4217.00

Start Tool MD [m] 3716.39 3716.78 4181.85 4205.93 4208.68 4208.67 4210.39 4210.00 4208.15 4209.95 4211.08 4211.63 4211.51 4215.15 4215.94 4216.51 4216.56 4215.20 4216.92 4216.66 4210.74 4206.96 4186.99 4182.63 4124.98 1699.32 1688.63

Start Tool MD [m] 3716.78 4181.85 4205.93 4208.68 4208.67 4210.39 4210.00 4208.15 4209.95 4211.08 4211.63 4211.51 4215.15 4215.94 4216.51 4216.56 4215.20 4216.92 4216.66 4210.74 4206.96 4186.99 4182.63 4124.98 1699.32 1688.63 135.56

Start Time [hh:mm] 00:00 02:35 04:25 04:30 05:00 06:05 06:10 06:35 06:45 07:00 07:15 07:25 07:30 09:10 09:15 09:45 09:50 10:05 10:10 10:15 11:55 12:00 12:05 12:10 12:25 19:45 19:50

End Time [hh:mm] 02:35 04:25 04:30 05:00 06:05 06:10 06:35 06:45 07:00 07:15 07:25 07:30 09:10 09:15 09:45 09:50 10:05 10:10 10:15 11:55 12:00 12:05 12:10 12:25 19:45 19:50 00:00

Duration [hh:mm] 02:35 01:50 00:05 00:30 01:05 00:05 00:25 00:10 00:15 00:15 00:10 00:05 01:40 00:05 00:30 00:05 00:15 00:05 00:05 01:40 00:05 00:05 00:05 00:15 07:20 00:05 04:10

Operations Long Circulation (at 3717m) Run in hole Wash Ream&Wash Drilling rotary Circulation (at 4210m) Drilling rotary Ream&Wash Run in hole Circulation (at 4211m) Drilling rotary Ream&Wash Drilling rotary Ream&Wash Drilling rotary Circulation (at 4217m) Drilling rotary Ream&Wash Drilling rotary Ream&Wash Run out of hole Wash Run out of hole Wash Run out of hole Wash Run out of hole

Table 1: Automated Reporting Results for one Day (same Day as Presented in Figure 2 and 4)

SPE 103211

00 55

500 50 1000 45 40 35 1500

2000 30 2500 25 20 15 3500 10 4000 05 00 4500 3000

00:00

01:00

02:00

03:00

04:00

05:00

06:00

07:00

08:00

09:00

10:00

11:00

12:00

13:00

14:00

15:00

16:00

17:00

18:00

19:00

20:00

21:00

22:00

23:00

Time [hh:mm]
Drilling rotary Drilling sliding Ream and Wash upwards (backreaming) Ream and Wash downwards (into hole) Wash upwards (out of hole) Wash downwards (into hole) Run in hole Run out of hole Circulation on Make connection Other Operations Bit Position [m] Hole Depth [m]

Figure 4: Automated Operation Recognition Results for one Day (the blue and green Line are Planned and Best Performance Trip Out)

Table 2: Performance Calculation of Trip Out

24:00

Measured Depth [m]

Time [min]

10

SPE 103211

Demowell 7 - Drilling Run 43

2005-03-26 01:38 to 2005-03-31 00:00 Duration 117:21 [hh:mm]

1000

MD [m]

2000

3000

4000

5000

Make Up BHA

Trip In

Drilling Formation

Short Trip Drilling Formation Ream/Wash

Trip Out

Break BHA

2005-03-26 01:38 2005-03-26 06:05 2005-03-26 14:27 2005-03-29 01:55 2005-03-30 05:02 2005-03-30 10:04 2005-03-30 11:52 2005-03-30 22:58 2005-03-26 06:05 2005-03-26 14:27 2005-03-29 01:55 2005-03-30 05:02 2005-03-30 10:04 2005-03-30 11:52 2005-03-30 22:58 2005-03-31 00:00 Duration 04:27 [hh:mm] Duration 08:22 [hh:mm] Duration 58:27 [hh:mm] Duration 27:08 [hh:mm] Duration 05:01 [hh:mm] Duration 01:49 [hh:mm] Duration 11:05 [hh:mm] Duration 01:02 [hh:mm]

RW Dwn00:06

1.2 %

RIH POOH Circ Conn Other Sum

00:05 00:03 00:09 00:30 03:41

2.0 1.0 3.2 11.1 82.7

% % % % %

RIH POOH Circ Conn Other Sum

01:34 00:13 00:04 04:04 02:20

18.7 2.7 0.9 48.6 27.9

% % % % %

Drill R 30:20 Drill S 16:27 RW Up 00:45 RW Dwn01:11 W Up 01:02 W Dwn 01:43 RIH 00:07 POOH 00:07 Circ 01:09 Conn 00:46 Other 04:52 Sum

51.9 28.1 1.3 2.0 1.8 2.9 0.2 0.2 2.0 1.3 8.3

% % % % % % % % % % %

Drill R 00:01 RW Up 00:13 RW Dwn00:15 W Up 01:37 W Dwn 01:24 RIH 02:05 POOH 02:15 Circ 09:33 Conn 03:08 Other 06:38 Sum

0.0 % 0.8 0.9 6.0 5.2 7.7 8.3 35.2 11.6 24.4 % % % % % % % % %

Drill R 02:39 RW Up 00:08 RW Dwn00:13

52.8 % 2.7 % 4.2 %

Drill R 00:01 RW Up 00:57 RW Dwn00:41 W Up 00:01 W Dwn 00:01 RIH 00:00

0.6 % 52.3 37.3 1.0 0.7 0.3 % % % % %

RIH POOH Circ Conn Other Sum

00:03 00:02 00:23 00:05 01:29

1.0 0.7 7.5 1.7 29.4

% % % % %

W Up W Dwn RIH POOH Conn Other Sum

00:27 00:02 00:28 02:16 06:41 01:11

4.1 0.4 4.2 20.5

% % % %

RIH POOH Conn Other Sum

00:01 00:08 00:20 00:32

2.3 % 13.0 % 32.8 % 51.9 %

Other Sum

00:08

7.6 %

60.2 % 10.6 %

04:27 100.0 %

08:22 100.0 %

58:27 100.0 %

27:08 100.0 %

05:01 100.0 %

01:49 100.0 %

11:05 100.0 %

01:02 100.0 %

Figure 5: BHA Run Report Chart

Trip Out
Start Time Duration [hh:mm]

2005-03-30 11:52:51
End Time

2005-03-30 22:58:17 4210.33m

11:05
Delta Bit MD

Start Bit MD End Bit MD

302.55m
Start Hole MD End Hole MD

3907.77m
Delta Hole MD

4217.00m 4217.00m

0.00m

Drill R Drill S RW Up RW Dn W Up W Dn RIH POOH Circ Conn Other SUM

00:00 00:00 00:00 00:00 00:27 00:02 00:28 02:16 00:00 06:41 01:11 11:05

0.0% 0.0% 0.0% 0.0% 4.1% 0.4% 4.2% 20.5% 0.0% 60.2% 10.6% 100.0%

Special Analysis for Trip Out

average gross tripping speed ............................. 352.35m/h # of recognized connections ......................................... 138 average connection time....................................... 2.90min
Figure 6: Details for Trip Out Interval of BHA Run Report Chart in Figure 5

SPE 103211

11

3700

3800

3900 MD [m] 4000 4100

4200

Circulation Ream&Wash Washing 0 1 2 3 Time [h/10m] 4 5 6 7

Figure 7: Wellbore Treatment Time along MD

Wellbore Treatment Time - Demowell 7

0

Wellbore Drilling Time - Demowell 7

0

500

500

1000

1000

1500

1500

2000

2000

MD [m]

2500

MD [m]
1

2500

3000

3000 1 3500
2

3500

2 4000

4000 3 4500

3 4500

5000

0 5 10 Treatment Time [h/10m]

0 5 10 Circulation Time [h/10m]

0 5 10 Washing Time [h/10m]

0 5 10 Reaming Time [h/10m]

5000

0 5 10 Rotating Time [h/10m]

0 5 10 Sliding Time [h/10m]

0 5 10 Total Drilling Time [h/10m]

0 5 10 Treatment Time [h/10m]

Figure 8: Wellbore Treatment Time - Conditioning Time

Figure 9: Wellbore Treatment Time - Drilling