Much can be done to improve the Well Testing through effective use of minimal electronic instrumentation on the well head and the test separator. The purpose of this paper is to describe Shell tools and experiences using the resulting real time data to enable well test optimization and automation. 1.0 Introduction The purpose of well testing is to periodically determine oil, gas and water flows for accounting, reporting and surveillance purposes. Hydrocarbon allocation provides official reports of well and reservoir production for lease owners, petroleum revenue tax purposes and management reports as well as feeding into hydrocarbon reserve figures and reservoir simulations which are used for major field decisions e.g. where to drill the next out-step well. Surveillance is key to determining well and reservoir behaviour and ensuring optimal well productivity and integrity. Routine well testing is an established procedure. Wells are for the most part manually diverted to a gravity separator or multi-phase meter and oil, water and gas phases are measured discreetly. Tests are periodically conducted, for example once a month, or once a week. The duration of purge and test periods are usually fixed, for example 30 minutes to purge the test separator and eight hours to test the well. The manual well testing process is subject to error and uncertainty - the wrong well may be put on test, the wrong instrumentation may be used, the instrumentation range may be incorrect and instrumentation may be dysfunctional. After the well test is complete the final result may be good, bad or suspect, hence results are usually subject to a manual validation process. This begs a number of questions related to well test optimization and automation: The purpose of this paper is to discuss Shell E&P experiences and real time software applications relating to the above well test optimization and automation issues.

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractMuch can be done to improve the Well Testing through effective use of minimal electronic instrumentation on the well head and the test separator. The purpose of this paper is to describe Shell tools and experiences using the resulting real time data to enable well test optimization and automation.

This paper describes a cost-effective, pragmatic way to continuously monitor and control well performance for most oil well types. Well fluid flow rate estimates are displayed in graphical form as trends on user desk top PCs. Deviations from the norm are flagged as process alarms. In some cases measured fluid flow changes are used to automatically control the well by adjusting choke settings for performance improvement. The main benefits of this approach are safer operations and reduced operating costs - operators visit only wells with problems thus reducing exposure to hazard and eliminating spurious travel and logistics. Other benefits include increased production due to early warning of well problems, better understanding of well behavior, improved allocation accuracy and improved well testing. The purpose of this paper is to describe Shell experience with low cost continuous well performance devices and some of the benefits that can be gained from this approach. Introduction The traditional approach to well surveillance - Operators go to the wells. For effective surveillance, we need to know when a well is off production, producing abnormally low or high, or behaving in an unstable manner. We need to know how a well responds to changes in operating parameters, e.g. gas-lift injection rate, pumping speed, choke setting, etc. Measurement of well productivity in the Oil Industry has traditionally been done by well testing - the oil, water and gas phases are separated by gravity in a tank and then measured individually. A test separator is usually shared amongst a number of wells. Hence well testing is inherently discontinuous (e.g. once per month). Also, because of back-pressure effects the test conditions may be different from actual well production conditions and also well test results may be adversely affected by leaking manifold valves. We need to know when to test a well, and when a test is not needed so we can concentrate testing on wells where information is most needed. Well surveillance is traditionally a manual process. Operators routinely visit the well production sites, put wells on test, take samples and physically inspect the wells to determine if there is a problem. Most of the time the operators find that the wells are healthy. Consequently, most of the time well head visits prove to be unnecessary. Lots of unnecessary lab samples and data result from the visits. Samples and data have to be processed, again resulting in much spurious activity. Sometimes the volume of data and errors can obscure the results, or give rise to false and misleading results e.g. well tests frequently have to be repeated and in some cases well problems are missed due to the shear volume of data. The ideal would be a well head device that could continuously, cost effectively and sufficiently accurately measure well productivity. This information would be transmitted back to the operator, who would then use the data to judge which locations have problems justifying manual intervention. Hence, limited manpower resource could be concentrated on problem areas to correct well problems sooner and more effectively. Even better would be to apply sufficient intelligence to automatically control the well such that unsafe, or unproductive conditions are automatically recognized and corrective action taken by immediately activating control valves. Such monitoring and control devices now exist and the purpose of this paper is to describe these devices and outline Shell's experience to-date with these techniques. Continuously measure well productivity and "bring the wells to the Operators" - FieldWare FlowMonitor Well liquid flow rates can be estimated by measuring the pressure drop (delta-P) across a fixed restriction in the well flow line with a delta pressure transmitter. A pressure drop of .1 - 1 bar is required across the line restriction. The instrumentation signal may be telemetered back to a remote control room, or via the Internet to a PC in the operator's office.

Downstream Refineries and Chemical Plants have benefited from real time optimization systems (RTO) for the last 30 years. Downstream RTO is a well established and permanent fixture in many plants – the "way we do things ‘round here!". Upstream E&P operations have "come to this party" much more recently and are using RTO more sparingly, even though the economic and HSSE benefits can be very significant. There are key differences between downstream and upstream. For example, downstream facilities do not deal with sub-surface uncertainties, multiphase flow and isolated/harsh environments; while upstream operations do not usually have to deal with complex chemical processes. Integrated Oil Companies run upstream and downstream operations and integration of tools/practices across both regimes is often perceived to be of significant value. Hence, the purpose of this paper is to compare and contrast downstream and upstream RTO learnings with a view to identifying and describing: similarities in production unit operations e.g. fluid separation, compression etc.; key differences between production unit operations; cultural differences between operations; RTO activities from a technical perspective; RTO business benefits and how these might be leveraged and sustained in both directions. What will emerge from this analysis will be a comparison, highlighting points of commonality and differences, leading to a better understanding of how RTO can be more effectively exploited in the upstream business – the cheapest oil available! Specifically, it is concluded that RTO in upstream operations is feasible and lucrative, but is relatively rare with sustainability a challenge. Downstream RTO is more common and sustainable, significantly less lucrative, but a "must do" to compete in a highly competitive, margin constrained business.

Downstream Refineries and Chemical Plants have benefited from real time optimization systems (RTO) for the last 30 years. Downstream RTO is a well established and permanent fixture in many plants -the "way we do things 'round here!". Upstream E&P operations have "come to this party" much more recently and are using RTO more sparingly, even though the economic and HSSE benefits can be very significant.There are key differences between downstream and upstream. For example, downstream facilities do not deal with sub-surface uncertainties, multiphase flow and isolated/harsh environments; while upstream operations do not usually have to deal with complex chemical processes.Integrated Oil Companies run upstream and downstream operations and integration of tools/practices across both regimes is often perceived to be of significant value. Hence, the purpose of this paper is to compare and contrast downstream and upstream RTO learnings with a view to identifying and describing: similarities in production unit operations e.g. fluid separation, compression etc.; key differences between production unit operations;  cultural differences between operations; RTO activities from a technical perspective; RTO business benefits and how these might be leveraged and sustained in both directions.What will emerge from this analysis will be a comparison, highlighting points of commonality and differences, leading to a better understanding of how RTO can be more effectively exploited in the upstream business -the cheapest oil available! Specifically, it is concluded that RTO in upstream operations is feasible and lucrative, but is relatively rare with sustainability a challenge. Downstream RTO is more common and sustainable, significantly less 2 SPE 163710 lucrative, but a "must do" to compete in a highly competitive, margin constrained business.