Intelligent predictive methods have the power to reliably estimate water saturation (Sw) compared to conventional experimental methods commonly performed by petrphysicists. However, due to nonlinearity and uncertainty in the data set, the prediction might not be accurate. There exist new machine learning (ML) algorithms such as gradient boosting techniques that have shown significant success in other disciplines yet have not been examined for Sw prediction or other reservoir or rock properties in the petroleum industry. To bridge the literature gap, in this study, for the first time, a total of five ML code programs that belong to the family of Super Learner along with boosting algorithms: XGBoost, LightGBM, CatBoost, AdaBoost, are developed to predict water saturation without relying on the resistivity log data. This is important since conventional methods of water saturation prediction that rely on resistivity log can become problematic in particular formations such as shale or tight carbonates. Thus, to do so, two datasets were constructed by collecting several types of well logs (Gamma, density, neutron, sonic, PEF, and without PEF) to evaluate the robustness and accuracy of the models by comparing the results with laboratory-measured data. It was found that Super Learner and XGBoost produced the highest accurate output (R2: 0.999 and 0.993, respectively), and with considerable distance, Catboost and LightGBM were ranked third and fourth, respectively. Ultimately, both XGBoost and Super Learner produced negligible errors but the latest is considered as the best amongst all.

No abstract

No abstract

No abstract

No abstract

No abstract

Matching pursuit belongs to the category of spectral decomposition approaches that use a pre‐defined discrete wavelet dictionary in order to decompose a signal adaptively. Although disengaged from windowing issues, matching point demands high computational costs as extraction of all local structure of signal requires a large size dictionary. Thus in order to find the best match wavelet, it is required to search the whole space. To reduce the computational cost of greedy matching pursuit, two artificial intelligence methods, (1) quantum inspired evolutionary algorithm and (2) particle swarm optimization, are introduced for two successive steps: (a) initial estimation and (b) optimization of wavelet parameters. We call this algorithm quantum swarm evolutionary matching pursuit. Quantum swarm evolutionary matching pursuit starts with a small colony of population at which each individual, is potentially a transformed form of a time‐frequency atom. To attain maximum pursuit of the potential candidate wavelets with the residual, the colony members are adjusted in an evolutionary way. In addition, the quantum computing concepts such as quantum bit, quantum gate, and superposition of states are introduced into the method. The algorithm parameters such as social and cognitive learning factors, population size and global migration period are optimized using seismic signals. In applying matching pursuit to geophysical data, typically complex trace attributes are used for initial estimation of wavelet parameters, however, in this study it was shown that using complex trace attributes are sensitive to noisy data and would have lower rate of convergence. The algorithm performance over noisy signals, using non‐orthogonal dictionaries are investigated and compared with other methods such as orthogonal matching pursuit. The results illustrate that quantum swarm evolutionary matching pursuit has the least sensitivity to noise and higher rate of convergence. Finally, the algorithm is applied to both modelled seismograms and real data for detection of low frequency anomalies to validate the findings.

No abstract