New Model Enhances Flux Management in Sand-Control Completions
This paper discusses a probabilistic flux and erosion model and work flow that extend the ability to estimate inflow through sand screens on a foot-by-foot basis along the wellbore using the well's completion details, production rate, and reservoir and bottomhole flowing pressures.
This paper discusses a probabilistic flux and erosion model and work flow that extend the ability to estimate inflow through sand screens on a foot-by-foot basis along the wellbore using the well’s completion details, production rate, and reservoir and bottomhole flowing pressures. The model is then calibrated and history matched using well data from pressure transient analyses, well tests, and production logs as available. Extensive laboratory testing coupled with computational flow dynamics modeling provided the algorithms for different screen types to relate flux and sand production to the expected service life for any given future production profile. This allows the well’s planned production profile to be optimized by balancing risk, rate, and reserves recovery.
The flux management work flow was field trialed on various failed and successful sand-control completions and has been used on many wells subsequently. It has been applied to wells currently on production by history matching to track current erosion levels and then to predict future erosion or time to failure. Failure here is taken to mean a loss of sand control, typical higher rates of sand with large particles and proppant. The work flow can also be applied to planned wells to help optimize the completion design. The work flow is currently being adapted to be applied to producing wells in real time.
When compared with conventional flux limit models, the work flow drives increased production while the risk of sand production is low and where the completion quality is good. Conversely, where pack quality is low or uncertain and sanding information is of low quality, the work flow can result in a comparative reduction in the allowable rates. The prediction uncertainty ranges reflect the quality of the input information and are driving improvements to the measurement of completion quality and sand production as well as to new, cost-effective, erosion-resistant screen designs.
The new work flow considers pack quality and sand production (particle size, type, and concentration) and allows an operator to maximize well economics by optimizing the production profile for rate, reserves recovery, and service life. Conventional screen erosion models, on the other hand, do not consider a sand-control screen’s erosion history and future service life.
Historical Perspective and Review
Sand-control failures can be caused by installation damage, compaction, corrosion, or improper deployment or inappropriate system selection. The complete paper presents a historical perspective and review of these failure mechanisms and methods of addressing them and includes a discussion of fluid velocities and constraints.
Enhanced Flux Management Work Flow
Conventional erosion models are aimed at calculating fluid-velocity limits and ensuring that the fluid velocity is constrained to below these limits at all times. The erosion models discussed previously also provide as an output the predicted time to erode for a given flow rate. The main inputs to these models can be summarized as follows:
- Superficial fluid velocity approaching the screen surface as derived from the casing tunnel exit velocity
- Particle size
- Particle concentration
- Particle mineralogy (hardness)
- Sand-screen design and geometry
However, several issues are not addressed properly by the models. These include the following:
- Multiphase flow. The models do not fully account for the different fluids expected downhole at the sand face and are based mainly on single-phase water. [Compressed air is not representative of gas under normal downhole conditions (i.e., greater than 1,000 psi). Erosion rates can be expected to differ with gas, water, and heavy oil.]
- Mobile particles. Production constraints applied do not typically account for the presence (or, more importantly, absence) of moving particles.
- Localized fluid velocities. No models have been identified that account for fracture geometry, presence of void sections, or high-permeability heterogeneity at varying production rates, pressures, and skins. Models have been developed that use production log or permeability information to derive an inflow profile, but, typically, this profile then is varied linearly to match the overall production rate.
The probabilistic flux and erosion model and work flow goes some way toward addressing these issues. However, although it has been applied to multiphase wells and evaluates the velocities in such wells, the effect of multiphase fluid properties on erosion rates with specific sand screens has not been fully addressed and so is treated as an uncertainty.
An overview of the work flow is presented in Fig 1. The work flow calculates the time to erode for each selected depth interval (point along the sand screen) and for each time period (e.g., daily or monthly historical production or forecasted production). The output is a risked time to failure for a given forecasted production profile, which then can be used to optimize production and reserves recovery from the well and field. The work flow requires completion equipment, reservoir, and completion-quality information on a depth-by-depth basis and solids and production information over time.
Under normal injection or production conditions, eroding a sand-control screen with fluid flow alone is not considered feasible within a normal well lifetime. Under simulated flow conditions during testing, no erosion is observed with fluids alone. Moving solid particles are needed to cause significant erosion risk to the sand screen. The presence of such particles, and when they may be expected to appear, needs to be considered at each point along the sandface when estimating time to failure. The complete paper includes discussions of sand-screen-erosion mechanisms. Additional topics include sand measurement and prediction, solids production across the sandface, fluid-velocity estimation, flow convergence, completion proppant-pack voids, erosion prediction, and model verification.
Observations and Conclusions
- A sand-control system forms one of the more critical elements in a typical production well’s reliability profile. Understanding and managing its associated risks, therefore, is very important.
- Conventional industry guidelines provide simple flux limits but do not account for the cumulative effect of sand production on the screen’s functional requirements.
- Conventional flux models use skin data or drawdown to estimate one average perforation efficiency for the well. The average perforation efficiency then is applied using permeability heterogeneity or a production log to give an estimate of the maximum perforation tunnel fluid velocity. This velocity is often unrealistic because it does not account adequately for nonlinear velocity profile changes at elevated production rates, fracture geometry, convergence, pack quality, or changing local perforation efficiencies. Detailed modeling indicates that the inflow profile for a high-skin well is unlikely to match the permeability profile.
- If sand production is associated with increasing water cut and depletion, applying arbitrary rate limits during the sand-free phase of a well’s production life can unnecessarily reduce the well’s economics or adversely affect the reservoir management plan.
- Voids and poor pack quality may be present in low-skin completions. This combination is a cause for concern and is another reason to verify annular pack quality with a suitable gravel pack or production type log. Better still would be a permanently installed flux or noise-monitoring system.
- Sand concentration, particle-size range, shape, and hardness remain uncertainties. Estimating downhole local sand entry from surface sand measurements is also uncertain. The use of in-situ distributed flow, acoustic, temperature, and pressure monitoring may provide insights into pack quality and the location of solids influx over well life.
- High perforation fluid velocities may lead to pack destabilization under certain circumstances. If this is to be used as a constraint, the tunnel superficial velocity should be calculated while accounting for convergence and other factors. Note that the screen can be eroded with moving proppant only.
- A screen failure may not necessarily mean that the well needs to be repaired or replaced immediately. This also could be considered in any decision analysis when setting well rates.
- Gas and multiphase flow effects on screen-erosion rates require further investigation.
- Setting production rates involves a complex interplay of issues—safety and environment, flow assurance, reservoir management, and facilities and operations management, for example. The effect of any relaxation of constraints should be considered carefully in a detailed risk analysis, and appropriate mitigations should be put into place.
- Localized flow concentrations should be prevented by ensuring annular pack quality. This is critical for the long-term reliability of sand-control completions.
- Measuring annular pack quality is key to managing production from cased-hole frac-packed (and openhole gravel-packed) completions. Pack-quality logs run and retrieved with the pack service tools can be open to interpretation and represent the quality of the annular pack only at the time of installation. It is important to consider subsequent well operations when assessing pack quality.
- Critical wells with poor pack quality are candidates for the enhanced flux management work flow.
- Conversely, wells with good pack quality constrained by conventional limits are also candidates for the enhanced work flow.
This article, written by JPT Technology Editor Judy Feder, contains highlights of paper SPE 191598, “Enhanced Flux Management for Sand-Control Completions,” by John Cameron, SPE, Karim Zaki, SPE, Colin Jones, SPE, and Antonio Lazo, SPE, Chevron, prepared for presentation at the 2018 SPE Annual Technical Conference and Exhibition, Dallas, 24–26 September 2018. The paper has not been peer reviewed.