Continuous rate and water cut data reveal reasons for allocation errors and provide means for their correction.

Field trials evaluating the performance of multiphase metering systems proved them as a viable testing and measurement solution, overcoming the limitations associated with conventional gravity-based test separators. The trials offered a broader understanding of the causes of poor production allocation factors and how they can be improved significantly through the use of multiphase metering technology.

Well testing and field allocation problems

Production allocation is necessary to reconcile oil, gas and water measurements at all entry and exit points in a production network. Back allocation procedures using data generated by periodic well testing and production measurements provide the basic requirements for data reporting and volume proration and serve as the primary basis for assessing the performance of producing assets. However, poor well testing and production measurements can result in improper allocation practices, which negatively impact material balance accuracy and reservoir understanding. They also can have serious economic consequences because of potential tax and custody transfer implications. While these issues pose major challenges in on- and offshore field operations, proper measurement and allocation become critical in marginal deep water subsea developments where well testing and production measurements may become the enabling technology.

The oil back allocation factor (defined as the ratio of the sum of the upstream oil volume measurements for all inflows into a network divided by the sum of oil outflows) ideally should have a value of unity (1.0), meaning that all fluid measurements coming into a system should match those going out, after correcting for pressure and temperature variations and accounting for phase mass transfer. However, various measurement errors cause allocation factor values to vary from 0.6 to 1.2, with the worldwide average estimated at about 0.85. This "less-than-one" back allocation occurring in the majority of operations translates to an over-reporting of oil rates at the wellhead.

Mobile or permanent application test separators have exhibited limited performance because of their difficulty in achieving good separation under challenging flow conditions. With their use, performing accurate water cut measurements at wellheads was especially difficult in poorly conditioned well streams or in cases of stable emulsions.

A measurement solution

Unlike conventional test separators, multiphase meters can perform well in cases of high flow instability, stable emulsions and foaming. Also, their online measurements provide representative descriptions of the actual flow dynamics, enabling operators to better understand a well's normal flow behavior versus when it is flowing into a test separator.

Multiphase metering field trials

When deploying multiphase flowmeters, operators must secure agreement not only from partners, but also from other operators involved through custody transfers as well as the proper regulatory authorities. The acceptance of the meters relies in part on metrological performance tests performed in controlled environments, such as flow loops, as well as through field qualifications in the proper test environments.
As part of a new technology introduction to various operators, an in-depth qualification of multiphase meters was undertaken to assess their performance and accuracy for testing wells and measuring production. In applications ranging from remote well testing to gas lift optimization, field trials compared multiphase meter measurements with those of conventional gravity-based test separators.

The use of multiphase meters in the field trials promoted a better understanding of the causes of poor allocation factors. Elements known to jeopardize the field allocation process did in fact surface during the trials. The main contributing elements were:

• unstable well production;
• undersized test separators; and
• well flow dynamics.

While it was possible, in some cases, to rectify the problems with the existing systems, many inherent limitations prohibited viable solutions without the alternative test practices made possible with multiphase metering technology.

Unstable well production

A PhaseTester Vx dual energy gamma multiphase flow meter, jointly developed by Framo Engineering and Schlumberger, was installed on a North Sea well where excessive gas lift injection resulted in large liquid flow rate variations ranging from 600 to 6,000 b/d of fluid. Phase segregation and cyclic water coning caused water liquid ratios (WLRs) to vary from 20% to 60% (Figure 1). In this environment, performing well tests with a test separator would result in erroneous flow rate estimations. A three-phase test separator would have trouble handling the large instantaneous liquid rates present with a dynamic WLR. To correctly estimate true WLR, the separator's oil and water lines would have to be instrumented with inline water cut meters, which often are not available. Consequently, liquid samples from the test separator must be collected at regular intervals to determine the WLR. Getting a representative picture of the actual water production can be difficult when water cuts vary significantly. Often, the atypically high values collected during sampling are eliminated under the assumption that the sampling method is at fault. Invariably, the result is an underestimated water cut, and consequently, an overestimated net oil rate. Many field operators are surprised at the water ratio variations measured by multiphase flowmeters.
Undersized test separators


Undersized test separators can generate problematic and unstable liquid rate measurements (Figure 2). In one field example, when a well flowed at high rates, large flow rate variations were reported by the separator instrumentation even though the well condition (i.e. choke setting) did not change. The undersized separator could not achieve good separation because retention time dropped below the recommended one-minute residence time. Furthermore, in this instance, emulsion and foaming problems compounded the separation process. Instabilities in liquid and gas rates also occurred. In contrast, the multiphase meter performed stable measurements throughout the test period, and its accuracy was not affected by any of these issues, including separation efficiency.

Well flow dynamics and test evaluations

Even without terrain-induced slugging and production instabilities, wells can experience large flow behavior changes caused by natural depletion. Analyzing well behavior over a long time period helps quantify production statistics. For example, a dual energy gamma multiphase meter was installed on a subsea well to continuously monitor its performance for more than 1 year. During this time, conventional sampling would likely have underestimated the increase in water production, for the reasons noted earlier.

Simulations were run using data from this well to estimate the effect of well test frequency and duration on the overall uncertainty of the well's projected production. The relative difference between the rate measured and used for allocation and the actual oil rate shows that large differences were experienced for short duration tests (Figure 3). This could be expected from the well's rapidly changing rates. Conversely, continuous monitoring, 720 hours in this case, led to zero error. However, little improvement in uncertainty occurred for test durations between 100 and 500 hours.

With allocation factors, there is a tradeoff between continuous or quasi-continuous measurements obtained with multiphase meters and non-continuous test separator measurements. In this particular case, test measurement durations of less than 72 hours led to uncertainties greater than ±5% on oil rates. In unstable flow environments, it is clear that the benefits of continuous measurement outweigh the need for certainty in the measurement itself. This translates directly to back allocation discrepancies.

Simulations using a 12-hour duration well testing base case helped determine the optimum testing frequency for this well. For such a short test, the well would have to be tested every few days to significantly reduce the standard deviation of the oil flow rate (Figure 4). For this particular well, the main gain would be obtained from increasing the duration of the tests rather than increasing their frequency. Such analysis, as shown here, is specific to the production conditions and requires input when evaluating the true dynamics of the production system. For continuous flow changes, the accuracy of the measurement system may be overshadowed by the need to make more frequent production measurements. Multiphase meters can address these situations since they are capable of conducting continuous online measurements.

Well testing strategies

Optimizing well testing methodologies to ensure the best back allocation factor requires a proper representation of individual wells in a pad or manifold. The wells flowing into a production manifold contribute differently to the total error of the commingled flow. Therefore, it is important to allocate testing duration and means differently among wells to minimize the uncertainty associated with the existing testing equipment and capabilities. The ideal strategy in allocating testing duration for various wells is to perform it on a basis proportional to the fluid amount to be back allocated. Equal duration testing, performed in many operating areas, is often not the best strategy. Optimal well testing practices dictate giving more coverage to the best producers (Mehdizadeh and Perry, 2002).

Some operators have proposed a hybrid solution in which a small multiphase meter is used on the test manifold to determine the flow rate of individual wells in conjunction with a global metering system that determines the total commingled flow on a continuous basis. While a large meter would provide good quality total flow information, it probably would have difficulties in performing accurate individual well measurements. The hybrid solution provides for the various metering requirements of both individual wells as well as the total commingled production. A clear understanding of the actual metrological performance of a multiphase flowmeter is essential for utilizing these measurements for back allocation purposes.

Editor's note: This article is based on SPE paper 76766 presented at the SPE Western Regional/AAPG Pacific Section Joint Meeting held in Anchorage, Alaska, May 20-22, 2002.