Waterflooding is one of the most important recovery methods in mature oil fields. By injecting water into the reservoir, operators can maintain pressure and push oil toward producing wells. However, one major question always controls the success of waterflooding:
Which injector is really communicating with which producer?
A recent paper by Mohamed Adel Gabry, Amr Ramadan, and Mohamed Y. Soliman, titled “Inter-Well Connectivity Estimation Using Continuous Wavelet Transform: A Novel Approach,” introduces a new answer to this question using routine injection and production data. The paper presents a Cross-Wavelet Transform Coherence, or CrWTC, workflow to estimate inter-well connectivity in waterflooded reservoirs.
Why This Study Matters
In waterflood management, understanding inter-well connectivity helps engineers answer practical questions such as:
- Which producers are strongly affected by each injector?
- Are there high-permeability streaks or preferential flow paths?
- Are some producers isolated by faults or barriers?
- Is injected water moving toward the desired part of the reservoir?
- Can injection be adjusted to improve sweep efficiency?
- Can water breakthrough be predicted or explained?
Traditional methods such as reservoir simulation and the Capacitance–Resistance Model, or CRM, are useful, but they have limitations. Full reservoir simulation can be slow, expensive, and highly dependent on geological-model quality. CRM is faster, but it uses simplifying assumptions and may struggle when the reservoir response is nonlinear, nonstationary, or strongly affected by heterogeneity.
This paper introduces a faster and more dynamic alternative based on wavelet signal processing.
Main Idea of the Paper
The paper treats the reservoir as a dynamic system:
- Injector rate = input signal
- Producer liquid rate = output signal
- Connectivity = how strongly the producer response follows the injector behavior
Instead of using simple correlation between injection and production rates, the method uses Continuous Wavelet Transform, or CWT, to analyze the signals in both time and frequency.
This is important because reservoir behavior changes with time. A producer may be weakly connected to an injector early in the waterflood but become more connected later as the waterfront advances.
The proposed method can capture this changing behavior.
What Is New in This Work?
The novelty of the paper can be summarized clearly:
- It applies Cross-Wavelet Transform Coherence to injector and producer rate data.
- It uses the complex Morlet wavelet, which captures both amplitude and phase relationships.
- It provides a continuous time–frequency connectivity map, not just one static correlation number.
- It can be updated as new field data become available.
- It helps track connectivity evolution and waterfront movement during waterflooding.
- It was tested on both simple synthetic cases and more realistic reservoir datasets.
- It was benchmarked using the Volve sandstone field dataset and the COSTA carbonate reservoir model.
Why Wavelets Are Useful Here
Production and injection data are not always simple. They may include:
- Shut-ins
- Rate changes
- Delayed pressure support
- Water breakthrough
- Operational noise
- Nonlinear reservoir response
- Time-varying connectivity
A normal time plot may not reveal these relationships clearly.
Wavelet analysis works like a mathematical microscope. It can show when two signals behave similarly and at what time scale this relationship appears.
In this paper, the CrWTC value ranges from 0 to 1:
- 0 means weak or no coherence.
- 1 means strong coherence.
- Higher coherence means stronger relative injector–producer communication.
However, the authors emphasize that CrWTC should mainly be used to rank producers connected to the same injector, not to compare absolute values across different injectors. This is because each coherence value is normalized by the energy content of the specific injector and producer signals.
How the Workflow Works
The workflow can be explained in simple steps:
- Collect injection-rate data from injector wells.
- Collect liquid production-rate data from producer wells.
- Synchronize the time series on a common time grid.
- Remove invalid long gaps and avoid artificial coherence from missing data.
- Apply CWT to each injector and producer signal.
- Calculate CrWTC for every injector–producer pair.
- Rank producer responses for each injector.
- Compare the results with simulation, CRM, geological interpretation, or water saturation maps.
- Update the analysis when new production and injection data become available.
This makes the method practical for reservoir surveillance.
Validation Cases Used in the Paper
The study did not rely on one example only. It tested the method at different levels of reservoir complexity.
1. Simple Synthetic Reservoir Model
The first test used one injector and four producers placed in areas with different permeability values.
The goal was simple:
- Check whether CrWTC can identify the producers that are more strongly connected to the injector.
- Compare the ranking with known permeability distribution.
- Validate the method against reservoir simulation.
The CrWTC method successfully ranked the connectivity between injector and producers in agreement with the expected reservoir behavior and with CRM results.
2. High-Permeability Streak Case
The second test used a synthetic reservoir with high-permeability streaks.
This is important because high-permeability streaks can create preferential flow paths. In waterfloods, these paths may cause injected water to move quickly toward certain producers, resulting in early water breakthrough and poor sweep efficiency.
The method was tested to see whether it could detect these preferential connections from rate data.
3. PyWaterflood CRM Benchmark
The paper also compared the CrWTC results with examples from the PyWaterflood CRM framework.
This comparison was useful because CRM is a common reduced-order method for estimating injector–producer connectivity. By comparing CrWTC with CRM, the authors showed how the new wavelet-based method performs relative to an established connectivity tool.
4. Volve Sandstone Field Dataset
The Volve field case is one of the strongest parts of the paper because it uses real field data.
The Volve field includes:
- Two injector wells
- Five producer wells
- Around nine years of waterflooding history
- A history-matched reservoir simulation model
- Complex operational behavior, including shutdowns
The paper showed that CrWTC results agreed with the field behavior and simulation interpretation.
Important observations included:
- Strong connectivity between I-F-5 and P-F-14
- Strong connectivity between I-F-4 and P-F-12
- Lower connectivity for P-F-15D and P-F-11
- Reduced communication toward P-F-1C
- Connectivity trends consistent with water saturation movement in the simulation model
The history-matched simulation showed the waterfront moving mainly toward the northwest, and the CrWTC results reflected this movement.
5. COSTA Carbonate Reservoir Model
The COSTA model was used to test the method in a more complex carbonate reservoir system.
This is important because carbonate reservoirs often have:
- Strong heterogeneity
- Layered flow units
- Complex depositional architecture
- Large permeability contrasts
- Complicated sweep behavior
Using both Volve and COSTA helped demonstrate that the method can be applied to different reservoir types, not only simple sandstone systems.
Main Technical Contribution
The main contribution of the paper is not simply applying wavelets to production data. The real contribution is using wavelet coherence to map dynamic injector–producer communication.
This is different from older wavelet-based connectivity methods because:
- Older methods often used Discrete Wavelet Transform, or DWT.
- DWT analyzes signals only at fixed decomposition levels.
- Older methods often used cross-correlation, which assumes more linear behavior.
- The new method uses continuous wavelet coherence, which gives a smoother and more complete time–frequency view.
- The result is normalized between 0 and 1, making it easier to interpret as relative synchrony.
This allows the method to detect both amplitude and phase relationships between injector and producer signals.
Practical Benefits for Reservoir Engineers
This method can help reservoir engineers in several ways:
- Quickly screen injector–producer relationships.
- Identify dominant flow paths.
- Detect weakly connected producers.
- Support injection reallocation decisions.
- Track waterfront movement over time.
- Compare dynamic connectivity with geological interpretation.
- Complement CRM and reservoir simulation.
- Reduce dependence on expensive full-field simulation for early diagnostic work.
- Update connectivity maps as new data become available.
The method is especially attractive because it uses data that operators usually already have: injection and production rates.
Important Interpretation Note
CrWTC values should be interpreted carefully.
The paper makes an important point:
- CrWTC is best used to compare producers around the same injector.
- It should not be used as a universal absolute connectivity number across all injectors.
- For example, a CrWTC value of 0.8 for one injector–producer pair does not necessarily mean the same physical connectivity as 0.8 for another injector with a different injection history.
So the best practical use is:
For each injector, rank the producers from strongest to weakest response.
This ranking can then guide reservoir-management decisions.
Why This Paper Is Important ?
This work fits strongly within the research direction of using advanced signal processing and data analytics for petroleum engineering.
It shows that ordinary field data can contain hidden reservoir information when analyzed with the right mathematical tools. Instead of relying only on static maps, simple correlations, or full simulation, the CrWTC method provides a dynamic way to see how wells communicate through time.
The paper also connects several important research areas:
- Waterflood optimization
- Reservoir surveillance
- Inter-well connectivity
- Wavelet transform
- Time–frequency signal analysis
- Data-driven reservoir management
- Field-scale validation
Key Takeaway
This paper presents a clear and practical message:
Cross-Wavelet Transform Coherence can turn routine injection and production data into dynamic inter-well connectivity maps.
By using CWT with a complex Morlet wavelet, the method can identify how injector–producer communication changes with time, detect dominant flow paths, and support better waterflood management. The validation using synthetic models, CRM comparison, Volve field data, and the COSTA carbonate model demonstrates that the approach is promising for both research and field applications.
