Field Performance Analysis often involves analyzing hundreds, sometimes thousands of multifractured horizontal wells with low resolution data. But the demands on engineers remain the same, to analyse these wells and to provide regular forecasts, reserves and performance optimization. With such massive data, there is insufficient time to precisely model every well. Traditional decline curve analysis is often used, however in shale reservoirs these techniques which were originally developed for conventional formations, are not adapted to the transient nature of unconventional reservoirs.

Citrine provides two workflow methods for analysing the performance of a field. The first uses traditional decline curve analyses. All the common empirical decline models are available for fitting a decline curve to the well data and regression tools are used quickly apply the decline curves to all the wells in the field. These methods are simple to use and provide quick forecasts, however in unconventional formations we are fitting decline curves to transient behavior. Consequently, a second workflow method was developed that consists of identifying a ‘type well’, accurately modeling the well with Topaze, and then applying this as a proxy model to a group of wells.

Citrine Workflow

  • Data\nProcessing
  • Diagnostics
  • Analysis
  • Multi-well Processing
  • Results

ADO Load

Citrine allows data load from multi-well spreadsheets, from databases with ODBC connectivity. Citrine loads both static well/reservoir data and dynamic or ‘time-series’ data.

Load from a database

The flexible load maps between the data source and Citrine internal structure. When loading from a database, this mapping can be saved and used for a subsequent update.

Citrine is 98c compliant.

Load from IHS EnerdeqTM

The user can pull data from the IHS web-service subject to their privileges and the defined query criteria. The configuration can be saved so that the project is updated with extra data when available.

2D Map

If well coordinates are present in the incoming data, well locations can be drawn on a 2D map. The map tool can be used to visualize graphically where higher productivity wells are located.

Wells grouping

Wells can be grouped according to their common behavior or properties (location or completion parameters) to identify characteristic underlying flow features, and to compare their performance.

Identify flow regimes

In addition to an extensive list of inbuilt graphs, it is possible to add any derived user-defined data channel.

Citrine can plot ‘anything versus anything’ to arrive at diagnostics to which variables can be normalized and comparisons made. Multiple plots can be shown saved as dashboards.

Normalization tools

Normalization allows the identification of factors affecting production performance by dividing dynamic data of each well by its corresponding parameter. Single or multiple parameters can be used in a normalization.

P10, P50, P90 Statistics

It is useful to display percentile curves to calculate average well behavior in a group of wells and to assess the range of production profiles.

An option generates a statistical curve to be used as a type well for decline analysis on a wider scale.

‘Type well’ selection

Type Wells represent a group of wells identified by having a similar decline behavior in the diagnostics. In the workflow, the user can analyze this ‘type well’ and apply the decline model to the rest of the wells in the group, thus significantly reducing analysis time and workload on the engineer.

Decline Curve Analysis (DCA)

Classical decline curves are available in Citrine. These simple empirical curves are fitted to the data to predict decline rates and used in forecasting. Curves are applied with modifications to the unconventional reservoirs (Exponential, Modified Hyperbolic, Power-law exponential, Stretch exponential, Duong, Logistic Growth).

Regression tools

When a decline model is initialized, a background nonlinear regression on model parameters is performed to obtain an initial match. Model parameters can be adjusted by means of controlled nonlinear regression, with an option of assigning weight to rates or cumulative production.

Model based analysis

Citrine can transfer single well data to Topaze. An analytical or complex numerical model (including advanced fracture models and fractures defined by micro seismic mapping) allows the user to capture transient behavior beyond the capabilities of simple decline curves. The forecast from Topaze is then used as a seed for multiwell analysis and field forecasting.

Multiphase analysis

For multiphase data DCA analyses can be created separately for each fluid phase.

When multiphase data is sent to Topaze, the simulated results can be retrieved for each phase.

From nothing

Multi-well processing creates a chosen decline model for each well in a group simultaneously. The regression is run independently for each of the wells to match the data.

From DCA

Multi-well processing can be initiated from a decline model matched for a particular well, with an option of fixing some of the estimated parameters (for example, decline exponent).

From Topaze Type Curve

A single step Type Curve brought in from Topaze can be used as a seed for automatic matching other wells data by applying X/Y shift on the rate-time log-log plot.

Bubble map

DCA, modeling results and volumetric quantities can be visualised on a map. Once EUR is estimated for all of the wells, the map of bubbles with proportional sizes will give an idea of reservoir depletion and remaining sweet spots.

Custom plots

Forecasts can be plotted against well parameters to detect factors governing well productivity. For example, the user could plot forecasts vs the number of fracture stages to understand the correlation between these parameters.

Results export

Model forecasts can be exported in Excel, CSV and Aries format. Multiple models for a given well can be compared in tabular form.