Commingled Well Model CWM
Our Commingled Well Model (CWM) is the result of many years of experience in well analysis. This package will be a powerful tool for anyone attempting to understand commingled well flow behaviour. The Software works as a predictive or as an analytical tool:
In predictive mode the CWM can be used to model the behaviour of a new commingled completion or to look at the effect of changing an existing completion (adding perforations or shutting a zone off, for example). The model predicts the performance of each layer including overall production and crossflows between layers. In this way it acts as a powerful completion management tool.
In analytical mode the model can be used to match well surveillance results such as production logs or distributed temperature (DTS) data. Combining this information with the well’s production history and its initial conditions allows the user to infer how reserves are distributed between layers. This facilitates optimization of production and future field development.
The software is easy to use and is available only from WTKI. It is significantly cheaper and easier to use than other softwares and can be installed easily on any PC.
To purchase a license, click download below and follow the registration process. For product support contact CWM@wtki.com.au.
CWM 2.2 Introduction
The CWM is specifically designed for modelling commingled gas wells, though it is perfectly valid for the simpler single layer case. The results of the CWM in terms of layer and well rate and pressure evolution are broadly consistent with the discussion in the classic SPE paper 01329-G: A Study of the Behavior of Bounded Reservoirs Composed of Stratified Layers (Lefkovits et al , 1959-1960).
The CWM can be used for forward modelling of the expected behaviour of a commingled well. It also incorporates a methodology for estimating the GIIP (gas initially in place) associated with each individual layer, following the workflow outlined in SPE paper 158733: Estimating Zonal Gas-in-Place in a Commingled Well Using Results from Production Logs (Last, 2012) The main features of the CWM are:
- Inflow performance calculations for each layer based on the input layer properties and a pseudo steady-state (PSS) inflow equation. The use of pseudo steady-state inflow is more in line with the method of Tempelaar-Lietz, which is referenced in the Lefkovits paper, and introduces some inaccuracies at early time in each timestep. However, as Lefkovits pointed out in his paper, the differences between a more exact solution (transient inflow transitioning to PSS) and this simplified approach are usually small.
- The CWM is a time-stepping model, with the total-well gas flowrate at each timestep matched to the corresponding entry in an input Production History. At each timestep a material balance calculation is made for each layer, calculating the new layer GIIP and p(bar)1 based on the average layer rate during the timestep and the timestep duration. This GIIP & p(bar) are used as the starting point for the next timestep.
- Material balance is based on a closed-tank (i.e. P/Z) model.
- The inflow of each layer is based on layer properties, PSS inflow and the difference between the current p(bar) and the wellbore flowing pressure (Pwf) at the mid perforation depth of that layer.
- Friction and hydrostatic pressure losses are calculated, in each timestep, both between each layer and back to the wellhead (i.e. FTP or flowing tubing head pressure) 1 P(bar) = average reservoir pressure for the layer under consideration
- The total-well rate at each timestep is matched to the input in the Production History by setting Pwf at the bottom active layer, and marching upwards from that layer calculating layer inflow, pressure losses to the next layer, Pwf and corresponding inflow from that layer, and so on. Pwf is adjusted automatically until the specified total-well rate is matched.
- A logical test is applied to ensure that layer pressure – p(bar) - cannot fall below the Pwf that prevails during a given timestep.
- The CWM can match either to FTP or to the total-well rate, as input in the Production History. It cannot match both at the same time!
Commingled wells can also be modelled using, for example, Saphir’s multi layer model or the Petroleum Experts IPM suite (using GAP, MBAL and Prosper). Having been designed specifically for commingled systems, the CWM provides a more streamlined and cost-effective way of modelling such wells, and especially of analysing suitable data sets in order to derive layer-by-layer GIIP estimates. It also overcomes some of the limitations of these alternatives (for example in Saphir wellbore pressure drops between layers cannot be accounted for). Results from the CWM have been benchmarked against both Saphir and IPM and, where those tools’ limitations can be neglected, yield nearly identical outcomes.
The various plots produced by the CWM – some examples of which are shown below - provide an easy way to picture how the well performs over time, or to forward-model and compare various different completion scenarios and play what-if games by varying the input parameters and quickly generating alternative outputs (plot images can also be saved to the clipboard for comparison purposes). In this way the CWM functions as a quick and easy Initial Completions planning tool. This is what might be termed the predictive function of the tool.
Selective Inflow Performance (SIP) Matching
The SIP matching routine, on the other hand, represents a key analytical function of the CWM. The methodology provides a way to understand how the GIIP associated with a well is distributed among the separate perforated intervals, and also to compare the “measured” PI from a PL survey with the expected PI based on the permeability and skin input into the model. The starting point for the SIP matching is to have an external data set which provides the following information for each layer:
1. Current average reservoir pressure;
2. Flowrate and Pwf during a stable survey.
3. Productivity Index (in MMscf/d/psi or Mscf/d/psi)
Traditionally this type of information would come from a multi rate PL survey, typically with 3 flowing surveys at different rates, and a shut in survey. However any type of survey that provides those 2 pieces of data will be OK – albeit potentially with lower accuracy/reliability. For example:
- A 2-rate (e.g. single flowrate plus shut in) PL survey. The same information is obtainable from a standard 2-rate survey as from a multi rate survey. The results are generally lower confidence: 2 points can always be joined by a straight line, but then we have no information concerning the uncertainty in the SIP results;
- A set of interpreted flowrates from fibre optic (DTS) survey data at multiple flowrates (not separated by significant time intervals), together with a pressure data set either with direct measurements of pressure at each layer (preferred) or with a single point pressure that has been depth-corrected to provide a pressure opposite each layer at each flowrate of the survey. The latter method is likely to be inaccurate in multi-phase flow since it relies on a good understanding of the sources of water production, and a reliable slip model that robustly predicts phase hold ups.
The SIP Matching routine provides an interactive method to define the area (and therefore the Gasin- Place) associated with each layer by matching the SIP & PL results in terms of layer rates, pressures and PI’s. The areas are tuned until a match is achieved.
CWM 2.2 Workflow. CWM 2.2 Guided Example.