ADIPEC 2013 Technical Conference Manuscript

Transcription

1 ADIPEC 2013 Technical Conference Manuscript Name: Muhammad Safdar Company: Schlumberger Middle East S. A. Job title: DeepLook EM Data Processing Center Manager Address: Abu Dhabi, UAE Phone number: Category: EOR with Gas Injection (CO 2, N2 HC, acid gas) Abstract ID: 705 Title: WAG CO 2 EOR Horizontal Wells Pilot Surveillance at a Giant Abu Dhabi Oilfield: First of a kind 3D monitoring plan using DeepLook Electromagnetics (DLEM) Author(s): Muhammad Safdar (Schlumberger), Nidhal Al Alawi (ADCO), Hani Al Sahn (ADCO), Tawfiq Obeida (ex ADCO), Joseph Khoury (Schlumberger), Michael Wilt (Schlumberger), Mustafa Biterge (Schlumberger) This manuscript was prepared for presentation at the ADIPEC 2013 Technical Conference, Abu Dhabi, UAE, November This manuscript was selected for presentation by the ADIPEC 2013 Technical Committee Review and Voting Panel upon online submission of an abstract by the named author(s). Abstract: DeepLook Electromagnetics (DLEM) crosswell resistivity technique has proven its value as new reservoir scale surveillance method in many EOR projects. However, its application in WAG CO 2 injection and deployment in horizontal wells has not been fully assessed yet. The development of 3D inversion and improved reservoir understanding in WAG CO 2 EOR processes were trigger for first study and survey design for pilot located in ADCO s giant oilfield in transitional zone. Wells drilled there have confirmed that there is considerable oil in place. Hence pattern based, local EOR flooding pilot including program of 6 months long alternating WAG CO 2 injection cycles have been proposed. A horizontal linedrive pilot with one horizontal producer emplaced in middle between two horizontal injectors plus three vertical observation wells are planned for pilot in near future. The main objective of this pilot is assessment of CO 2 EOR potential in this area of field. In this paper we discuss DLEM detailed modeling and field survey design results. The modeling study results indicated that DLEM method is suitable for tracking both CO2 and water flooding; it showed sufficient sensitivity to distinguish between fluid volumes and proved that inverted images are good approximation of modeled saturation volumes. Under watered out conditions and/or when CO 2 injected after water cycle, DLEM technique has good sensitivity to CO 2 phase. 2D and 3D modeling is applied to vertical vertical, horizontal vertical and horizontal horizontal well pairs to ensure modeling the whole volume of reservoir between the wells. Both modeling and survey design for the pilot using recent 3D inversion algorithms, together with improved understanding of WAG CO 2 impact on reservoir resistivity distribution, indicated that time lapse DLEM inter well saturations provide useful monitoring data of the injected fluids; water and CO 2 fronts. Finally this data can be utilized to calibrate the dynamic simulation models for more robust history matching enhancing the overall evaluation of EOR pilot. 1

2 Introduction This paper presents surveillance modeling study results on CO 2 EOR pilot planned in one of the giant carbonate fields located onshore Abu Dhabi UAE. This field has been on production for more than 50 years. The reservoir comprises a series of interbedded shallow marine bioclastic carbonate rocks of Lower Cretaceous Thamama Group. The average thickness of the reservoir is about 150 feet at the crest. The porosity varies from 8% to 35%. The permeability ranges between 10 to md in the upper part and 1 to 10 md in the lower part of the reservoir. Each reservoir zone is isolated above and below by dense intervals. Reservoir quality decreases from crest to flanks, good reservoir quality are related to both depositional lithofacies and constructive diagenatic processes. The current production mechanism is influenced by peripheral, crestal and pattern gas and water injections executed in different parts of the field. Two EOR pilots are planned (North Flank Pilot and East Flank Pilot) for trials beginning in near future at northern transitional zone of the studied field (Figure 1). Recently drilled wells in this area have confirmed that there is considerable oil in place. Although oil is present in most of the reservoir layers but the water saturation is relatively high in lower layers. It is likely that much of this water is movable (Figure 2). Hence pattern based, local EOR flooding pilot including program of 6 months long alternating WAG CO 2 injection cycles have been proposed. The horizontal line drive pilot design with one horizontal producer emplaced between two horizontal injectors is selected. Additionally multiple observation wells are planned for data collection and monitoring of the pilot. DeepLook Electromagnetics (DLEM) based crosswell resistivity measurement technique has proven its value as a new reservoir scale surveillance method in many EOR projects. However, its application in WAG CO 2 injection and deployment in horizontal wells has not been fully assessed yet. The recent development of 3D inversion capability and the improved reservoir understanding in WAG CO 2 EOR processes were the trigger for the first modeling study and survey design. This paper describes DLEM detailed modeling and field survey design results of North Flank Pilot study. Pilot Objective, Concept and Design The main objective of WAG CO 2 EOR pilot is the EOR potential assessment of this area of the field. The selected pilot scheme is a horizontal well configuration and works in line drive injection (Figure 3). This pilot design features one horizontal producer emplaced in the middle between two horizontal injectors. Additionally three vertical observation wells, each chrome cased for DLEM survey considerations, are planned to drill for monitoring purposes. The first observation well () is already drilled for injectivity testing and other data gathering purposes. A second vertical observation well will be drilled before the start of the injection in this pilot. Additionally a third OB well will be drilled after CO 2 breakthrough to the producer to recover cores to determine residual and remaining oil saturation. As per plan this third observation well will be drilled in the close proximity of the producer, however, the location and timing of this well to drill could be optimized as per the pilot performance and surveillance considerations. The pilot will feature injection of 6 months long alternating cycles of CO 2 and water during the pilot life. Pilot Monitoring and Surveillance Objectives The field asset team is proposing several technologies to monitor the performance of EOR pilot during its execution. Among these technologies is DLEM, which measures the inter well resistivity. The resistivity data are in turn, sensitive to oil, water and CO 2 volumes and may be used for tracking fluid fronts, finding bypassed hydrocarbons and locating infill wells. The DLEM technology is mainly sensitive to the water phase. The CO 2 and oil phase are both insulating and will have roughly equivalent responses for the DLEM. We feel however, that the measurement of inter well water saturation will provide the required information to monitor the flood because all phases are moving. Under watered out conditions and/or when CO 2 injected after the water cycle, technique has good sensitivity to CO 2 phase. 2

3 Figure 4 shows that in the lower layers along the injection well () as the water saturation increases, the resistivity is decreased from pre injection readings of 1 5 ohm m to less than 0.5 ohm m, at 90% water saturation. Whereas in the case of CO 2 injection cycle as the CO 2 saturation increases, the resistivity is increased from the pre water injection readings of 0.8 ohm m to about 2 ohm m and/or post water injection readings from <0.5 ohm m to as close as 1 ohm m and above, pushing water saturations from 70 80% range to as low as 50%. These changes in saturations (or resistivity) therefore aid in the determination of CO 2 and oil saturation at various stages of pilot s operation. DLEM technology is studied for the WAG CO 2 injection monitoring to achieve following objectives; 1. Monitor flow and saturations of CO 2 and water during WAG CO 2 injection cycles 2. Evaluate DLEM sensitivity to injected fluid volumes at various stages 3. Examine resolution and accuracy of DLEM inversion reconstruction 4. Develop inter well survey design and surveillance plan DeepLook Electromagnetics (DLEM): Background DeepLook Electromagnetics (DLEM) uses electromagnetic induction principles and tomography to provide an image of the resistivity distribution between boreholes. The typical field set up (Figure 5) includes two boreholes, spaced X distance apart with a depth range Z. Inter well measurements are conducted by placing the transmitter in one well to broadcast electromagnetic signals throughout the formation. At the other well, an array of receivers detect these signals. Imaging the inter well space is completed by locating these transmitter and receivers at regularly spaced (about 2 to 5% of the inter well distance, X) intervals below, within, and above the depth range of interest. Ideally, the tomographic imaging aperture (Z/X) should equal one, so the depth range should be greater than or equal to the well separation. However, there have been many successful imaging examples where the depth range was less than that, making the aperture much less than one; in these cases the problem was fairly well constrained. Field data are interpreted by fitting the measurements to calculated data from a numerical model, using an inversion procedure called Gauss Newton. It begins with a resistivity model, usually derived from prior knowledge of the field area including geology, logs and seismic data. The model response is calculated using a forward code and the model parameters are adjusted by inversion until the observed and calculated data fit within a specified tolerance. Although the inversion results in non unique models, using field data as a guide and exercising reasonable model constraints in fitting the data aids in selecting the best model. Recently this workflow which is typically used for 2D sections in case of vertical wells is expanded to newly develop 3D workflow mainly required for the modeling of horizontal wells. Feasibility Modeling Normally with each DLEM project a feasibility phase is undertaken where the project viability is determined and the benefits quantified. With the selection of monitoring and surveillance inter well technology as DLEM the next step is full modeling phase and survey design. Here we worked closely to design a field program in tune with the reservoir simulation model. The plan would be to do DLEM modeling along with the reservoir simulation to; 1) design the data acquisition surveys, 2) determine the optimum times for time lapse measurements, and 3) optimize wells configurations for DLEM survey. The output of this phase will be a prediction model for the pilot, a survey plan for all upcoming field surveys and a starting model for the data acquisition and inversion. The availability of vertical observation wells, in pair with horizontal injectors/producers provide an excellent DLEM data gathering opportunity for multiple well configuration options in this pilot. The observation wells are each chrome cased for DLEM considerations. They provide triangular coverage of the injection movement in different layers between the injector and producer. Horizontal tomography using the open hole injector/producer wells will provide a high resolution 3

4 3D view of the sweep and conformance along the horizontal trajectory. The additional option of vertical horizontal profile provides a full section transect along a horizontal well. These horizontal horizontal and vertical horizontal well pairs provide opportunity to monitor injection sweep along the horizontal trajectories which is not possible with verticalvertical configuration tomography. However, the crosswell logging of horizontal wells requires a different process from the tomography typically applied in vertical wells such as tool deployment, logging time and 3D workflow. Modeling Workflow Sector mechanistic model of North Flank Pilot, already built by field asset team was the starting point for this study. This model contains geological static and reservoir dynamic models built at 10mx10m grid spacing available with porosity property and simulated saturations for initial conditions (before injection) and time lapse alternating 6 months long WAG CO 2 injection cycles over the six years of pilot duration. The injection scheme first starts with CO 2 cycle. DLEM workflow requires resistivity property to do forward modeling and inversion process. Resistivity property is computed from Petrel modeled property; porosity and Eclipse simulated saturations; water saturation utilizing field based parameters. However, during pilot execution phase it is recommended to conduct study for suitable parameters and salinity analysis of injected water (in modeling study salinity for both formation water and injected water is assumed same) and these findings should be used in saturations computation from DLEM resistivity measurement. To validate mechanistic model a comparison of synthetic logs computed from modeled properties and measured logs along the borehole of observation well is shown in Figure 6. The comparison shows that there is good match between the modeled properties and measured logs although small differences on saturations are noted which explains the resistivity mismatch at places while the porosity match is excellent. We believe that this modeling is sufficient for the DLEM modeling study. However, for the field survey campaign we recommend building a new model at finer scale (5mx5m) utilizing North Flank Pilot wells data to capture smaller reservoir variations. This new model will be used as a background model for the DLEM inversion process as well as input model for integrated reservoir modeling. DLEM surveillance requires a well pair, where two different wells are simultaneously modeled for the transmitter and receiver tools response at suitable acquisition parameters. Out of these acquisition parameters logging frequency is of most important which is highly dependent on the well completion and well spacing beside others. During this study following well pairs of different configurations are modeled. 1. Vertical Vertical Profiles (Completion: chrome cased chrome cased) a. (towards, well spacing 300m apart) b. OB3 (between and, well spacing 250m apart) c. OB3 (towards, well spacing 300m apart) 2. Vertical Horizontal Profiles (Completion: chrome cased open hole) a. (closest well spacing 125m apart) b. (closest well spacing 50m apart) 3. Horizontal Horizontal Profiles (Completion: open hole open hole) a. (well spacing 250m apart) b. (well spacing 500m apart) The DLEM modeling workflow is used for two main operations; sensitivity and inversion. Eclipse simulated saturations at initial condition are considered as base model while each injection cycle is used as target in the scenario models. The workflow then simulates DLEM surveys in well pairs where one can place sensors/sources, select the frequency and add noise. The output of the sensitivity step determines if there is sufficient DLEM signal difference between surveys 4

5 conducted at different injection cycles to warrant a field measurement. It is meant to de risk the planned field activity. Inversion reconstructs the resistivity distribution at acceptable details and resolution at inter well space. Modeling Results The forward modeling (sensitivity analysis) results are summarized as below; To model a well pair the choice of electromagnetic frequency is dependent on well spacing, completion type and injected volume. Each sensitivity profile is analyzed for the optimum frequency (Hz) where a scenario model meets the minimum acceptable threshold of DLEM field strength (40uV) and field difference (5%). In some cases, where we have large field strength, 4% field difference may be suitable. Although lower frequencies affect images resolution and increase survey time; during field survey, more suitable frequencies can be tested to ensure that better quality of DLEM data is acquired and acceptable SNR (signal to noise ratio) is achieved. For vertical vertical configuration well pairs (chrome cased, m apart) the optimum frequency varies between Hz. For vertical horizontal configuration well pairs (vertical; chrome cased, horizontal; open hole and m closest apart) the optimum frequency varies between Hz. For horizontal horizontal configuration well pairs (open hole and m apart) the optimum frequency varies between Hz. After 1st CO 2 injection cycle most of the well pairs show small electromagnetic field differences although still above the acceptable threshold except for well pair; OB3, it came below the required threshold. This well pair being away from the injection point will have higher sensitivity once the injection is passed the well (Figure 3). Since rest of the well pairs either cross the injector well or have injector well part of their well pair show generally high field differences. Although higher field differences are observed in most cases of water injection cycles, however, certain CO 2 injection cycles also show similar response. Besides doing 2D forward modeling for all well pairs, 3D forward modeling workflow is applied on vertical horizontal and horizontal horizontal configuration well pairs. These 3D simulations show enough sensitivities to be detected with DLEM technology both for CO 2 and water cycles, although after 1st CO 2 cycle, it shows small data differences between base and scenario models but still in the range of 10 20% which is above the required threshold. Here we are looking not only for the size of the difference but also the number of measurements that are affected due to injection. Figures 7 9 show the forward modeling (sensitivity analysis) results for one of the verticalvertical configuration well pair between and. These results show sensitivity analysis over different time steps during WAG CO 2 injection cycles. Each of these results is compared between baseline (time 0, before injection) and subsequent time lapse surveys (time 1, after 1st CO 2 injection cycle, time 2, after 1st water cycle and time 3, after 2nd CO 2 cycle). The sensitivity study is quite useful in planning the time lapse surveys to find out when and how often these surveys to be repeated. For this we not only look at the absolute field difference between base model and next scenario model but also at the relative difference between different scenario models to find out the incremental change. We have observed that most of the times there is enough change in saturations for both water (increase in water saturations) and CO 2 (decrease in water saturations) during alternating WAG CO 2 injection cycles at different well pairs. Accordingly we recommend for time lapse surveys after each 6 months long injection cycle except for few cases where a small field difference (below the required threshold) is observed, these time lapse surveys can be skipped. The inversion results are summarized as below; Based on the sensitivity study results, the inversion process is proceeded for each favorable scenario model using base model as a starting point. The time lapse inverted resistivity images are able to show the areas of change affected by WAG CO 2 flooding. Resistivity changes are imaged at 1% field measurement noise. The inverted images from three 5

6 vertical vertical well pairs are imported into Petrel to merge with the existing reservoir model. In addition to the resistivity comparison, we have converted these inverted resistivity data into water saturations using the same work flow explained earlier in the section Modeling Workflow during resistivity modeling (saturation to resistivity) but this time in reverse order (resistivity to saturation). Figures show a 3D view of the saturations against two merged panels ( and OB3), exhibiting the evolution of the water saturation distribution between the baseline (before injection, time 0) and various time lapse alternating 6 months long WAG CO 2 injection cycles (selected times 1 3 for illustration) in the vicinity of the injector well. On the same figures for comparison we have displayed side by side panels of Eclipse water saturation and DLEM saturations computed from DLEM inverted resistivity results as explained earlier. A close comparison of DLEM inverted sections with Eclipse simulated results reveals following observations; 1. There are certain background model differences due to the fact that initial model built at 10mx10m grid spacing is re gridded at even finer scale of 5mx5m to image minor inter well variations. 2. The simulated vs. inverted comparison at time 1 (after 1st CO 2 cycle) shows only minor changes during the inversion process due to the relatively lower signal noticed in the sensitivity analysis. 3. As more injection volume is swept through the reservoir in the later injection cycles, the increase in DLEM sensitivity helps to achieve close to similar images for both Eclipse and DLEM modeling cases. During water cycles since water saturation increases, the inverted images show lower resistivity. Whereas a reduction in water saturations or increase in resistivity is noticed in case of CO 2 injection cycles. The upward push of existing injection volume is observed as the new CO 2 injection cycle starts, more likely replacing water in addition to reservoir oil de saturation. 4. A smoothing effect is generally seen on the DLEM inverted results as compared to blocky Eclipse simulated results due to reasons likely, grid size, boundary effects and resolution mismatch between the two methods, however there is reasonable agreement in the volume distributions and flood fronts shapes as evolved during the pilot operation. Besides 2D inversion for all well pairs, 3D inversion workflow is applied on vertical horizontal and horizontal horizontal configuration well pairs. In contrast to 2D inversion where a section is imaged, 3D inversion images the whole volume of reservoir between the wells. Figures show scenario and inverted models for vertical horizontal and horizontalhorizontal configuration well pairs for time lapse injection surveys. During 1st CO 2 cycle, in spite of lower signal seen on scenario model, the inverted images are able to pick some changes along the injectors with few artifacts or missing parts. The comparison of vertical horizontal configuration well pairs (Figure 13) shows relatively better images in the close vicinity of along the but lower image resolution towards the toe of the due its larger distance from. Whereas shows roughly equal quality of images on both sides of the. The gaps in the inverted images against vertical wells are a blind spot where there is no coupling between the sensors. Horizontal horizontal configuration well pair; (Figure 14) has better quality images than other horizontal horizontal configuration well pair; because of smaller well spacing and high frequency used. However, the images from are also useful to map the other side of the pilot where no observation well is available. These vertical horizontal and horizontal horizontal configuration well pairs provide us opportunity to monitor injection sweep along the horizontal trajectories which is not possible with vertical vertical configuration tomography. Monitoring Plan In the pilot design there are three vertical observation wells (OB) each cased with chrome for DLEM surveillance considerations. The first OB well () is already drilled and used for injectivity testing in addition to other data gathering. The second OB well () will be drilled before the injection starts (to be located in close proximity to 6

7 injector). This will be used to measure sweep and conformance along the injector and from the injector to the. Additionally OB3 will be drilled adjacent to producer following CO 2 breakthrough to the producer. This well in pair with other OB wells will be used for DLEM imaging close to the producer. Vertical DLEM tomography of these three chromecased observation wells will provide excellent triangular coverage of the injection movement in different layers between the wells; and. As the WAG CO 2 injection moves away from the injector the DLEM data collected from these OB wells will help to measure sweep from the injector towards the producer in a pseudo 3D area. Additionally open hole horizontal DLEM tomography in case of horizontal horizontal and vertical horizontal well pair configurations provides a high resolution 3D view of the injection movement as it measures the sweep and conformance along the horizontal trajectories to capture pilot performance variations due to reservoir pressure and heterogeneities (Figure 15, Panel A). In the tentative DLEM monitoring plan a baseline survey of multiple well pair profiles is recommended to capture interwell reservoir variations at initial conditions. We recommend time lapse surveys after each injection cycle except for cases where a small DLEM response is anticipated as indicated by modeling results. After the onset of injection we propose two levels of surveillance; one prior to CO 2 breakthrough and one after CO 2 breakthrough at the producer. Since 3rd OB well will be drilled after the injection breakthrough at producer; the focus of the monitoring shifts as the injection moves from the injector towards producer. The initial time lapse surveys monitor the area close to the injector but then surveillance increases more towards the producer as the injection passes. In the later stages of pilot life (after injection breakthrough at producer) multiple well pair surveys are recommended to cover all the area between injectors and producer to image injection sweep, conformance and oil de saturation (Figure 15, Panel B and C). Conclusions and Project Way Forward The modeling study results indicate that the DLEM method is suitable for tracking CO 2 and water flooding; it shows sufficient sensitivity to distinguish fluid volumes in separate layers and inverted images are a fairly accurate representation of the true saturation volumes. Under watered out conditions and/or when CO 2 injected after the water cycle, the DLEM technique has good sensitivity to CO 2 phase. This bodes well for time lapse saturations monitoring at North Flank Pilot. Time lapse DLEM inter well saturations provide time dependant injected fluid movement which can be utilized in the history matching cycle of reservoir characterization work assuring more robust and reliable dynamic model generation which will help in the reservoir management and pilot performance evaluation. Acknowledgements The authors would like to thank Abu Dhabi National Oil Company (ADNOC), Abdu Dhabi Company for Onshore Oil Operations (ADCO) and Schlumberger for their support and permission to publish this work. References 1. Wilt, M. J., Alumbaugh, D. L., Morrison, H. F., Becker, A., Lee, K. H., and Deszcz Pan, M., 1995, Crosswell Electromagnetic tomography: system design considerations and field results. Geophysics, 60: Abubakar, A., Habashy, T., Druskin, V., Alumbaugh, D., Zhang, P., Wilt, M., Denclara, H. Nichols, E., and Knizhnerman, L., 2005, A fast and rigorous 2.5D inversion algorithm for crosswell EM data, SEG Extended Abstracts, 24, Li, M., Abubakar, A., Li, J., Pan, G., and Habashy, T., 2010, Three dimensional regularized Gauss Newton inversion algorithm using a compressed implicit Jacobian calculation for EM applications, SEG Extended Abstracts 29, 4p. 4. DePavia, L., Zhang, P., Alumbaugh, D., Levesque, C., Zhang, H., and Rosthal, R., 2008, Next Generation Crosswell EM Imaging Tool, SPE Obeida, T., Gibson, A., Al Hashemi, H., Baruah, B., 2010, Detailed Compositional Modeling of Gas Injection Pilot in Giant Carbonate Reservoir in the Middle East, SPE

8 Figures North Flank Pilot X, [m] km East Flank Pilot :62500 Figure 1: Map showing proposed WAG CO2 EOR pilots locations in the northern transitional zone of the studied oilfield. Figure 2: Open hole logs from the recently drilled observation well () in North Flank Pilot. 8

9 Figure 3: North Flank Pilot design and setup with proposed observations wells. Figure 4: The water saturation panels at different WAG CO2 injection cycles, displayed along the cross section of. The resistivity panel (extreme right) shows synthetic resistivity computed from different saturations along 1. 9

10 Figure 5: Schematic setup of DLEM transmitter and receiver tools to carry out inter well electromagnetic tomography. Figure 6: Cross section showing modeled/simulated properties before injection (left panels). Also shown comparison of logs synthetically computed from modeled properties and logs measured along the observation well (right panels). 10

11 Figure 7: Sensitivity analysis between base model and scenario model after 1st CO2 cycle for well pair. Figure 8: Sensitivity analysis between base model and scenario model after 1st water cycle for well pair. 11

12 Figure 9: Sensitivity analysis between base model and scenario model after 2nd CO2 cycle for well pair. Figure 10: Saturation comparison at time 1 between Eclipse simulation & DLEM inversion for vertical vertical well pairs. 12

13 Figure 11: Saturation comparison at time 2 between Eclipse simulation & DLEM inversion for vertical vertical well pairs. Figure 12: Saturation comparison at time 3 between Eclipse simulation & DLEM inversion for vertical vertical well pairs. 13

14 Scenario 1 st CO2 cycle Inverted 1 st CO2 cycle ( ) Inverted 1 st CO2 cycle ( ) Rt Scenario 1 st water cycle Inverted 1 st water cycle ( ) Inverted 1 st water cycle ( ) Scenario 2 nd CO2 cycle Inverted 2 nd CO2 cycle ( ) Inverted 2 nd CO2 cycle ( ) Figure 13: 3D map view for vertical horizontal configuration well pairs; and. Scenario 1 st CO2 cycle Inverted 1 st CO2 cycle ( ) Inverted 1 st CO2 cycle ( ) Rt Scenario 1 st water cycle Inverted 1 st water cycle ( ) Inverted 1 st water cycle ( ) Scenario 2 nd CO2 cycle Inverted 2 nd CO2 cycle ( ) Inverted 2 nd CO2 cycle ( ) Figure 14: 3D map view for horizontal horizontal configuration well pairs; and. 14

15 Figure 15: Tentative monitoring plan where Panel A shows pilot wells, spacing and well pair combinations. Panel B shows surveys time line and Panel C shows surveys plan to track injection movement as progressed during the pilot operation. 15