Intelligent decisions to stop or mitigate lost circulation based on machine learning

Abbas, Ahmed K.; Bashikh, Ali A.; Abbas, Hayder; Mohammed, Haider Q.

doi:10.1016/j.energy.2019.07.020

Cited by 63 publications

(15 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Its purpose is to use a machine learning model to predict lost circulation and further put forward prevention suggestions and remedial decisions for drilling engineers according to the predicted lost circulation information, such as lost circulation type and the estimated amount of lost circulation. − Jiang et al has combined the wellbore temperature transient pressure coupling model established with unscented Kalman filter to predict the location and the amount of lost circulation . Abbas et al developed a new model using an artificial neural network (ANN) and a support vector machine (SVM) to predict the lost circulation of vertical and deviated wells . Sabah has used the model of decision tree, hybrid artificial neural network, and the adaptive neuro fuzzy inference system to quantitatively predict the lost circulation.…”

Section: Introductionmentioning

confidence: 99%

“…6 Abbas et al developed a new model using an artificial neural network (ANN) 11 and a support vector machine (SVM) 12 to predict the lost circulation of vertical and deviated wells. 13 Sabah has used the model of decision tree, 14 hybrid artificial neural network, 15 and the adaptive neuro fuzzy inference system 16 to quantitatively predict the lost circulation. The main purpose is to estimate the well loss and the available prevention and remedy methods.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Research on an Lost Circulation Zone Location Method Based on Transient Pressure Wave

et al. 2021

View full text Add to dashboard Cite

Accurately identifying the location of loss zone after lost circulation is the key to subsequent plugging operation. In view of the difficulty of identifying the location of lost circulation zone, a method of identifying the location of loss zone by transient pressure wave signal is proposed. When lost circulation occurs, transient back pressure is applied to the wellhead at the surface choke manifold to produce transient pressure wave. The transient pressure wave propagates downward from the wellhead. The propagation process of transient pressure wave in an annulus system is analyzed, and the position of loss zone is determined according to the change of pressure signal at the choke manifold. Based on the simulation of this method, relevant experiments are also carried out. Aiming at the problem of excessive noise of the pressure wave signal collected in the experiment, variational modal decomposition (VMD) is used to decompose the signal into multiple band-limited intrinsic mode function (BIMF) components. Combined with a Hilbert spectrum, the time–frequency characteristics and energy distribution of each BIMF component are analyzed in turn. The main frequency component is selected to reconstruct the signal to achieve the denoising effect. On this basis, a wavelet modulus maxima method is used to decompose the denoised signal, extract the characteristic points of the signal, identify the loss circulation information in the signal, and then identify the thief zone position by a time–domain method. Through experimental verification, the existence of loss zone will affect the change trend of pressure wave; a VMD–wavelet modulus maxima algorithm can effectively remove the noise of the pressure wave signal and locate the pressure change point. The experimental recognition error range of this method is 0.10–9.22%, which has certain guiding significance for field application.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Research on an Lost Circulation Zone Location Method Based on Transient Pressure Wave

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Although physics-based models can be used to avoid LCIs, data-driven models offer another layer of information that can be used to estimate or predict unseen events based on historical data from offset drilling operations as well as real-time data collected during drilling operations. In this direction, different machine learning (ML) and deep learning (DL) models have been developed to predict mud loss of circulation from surface parameters [21][22][23][24]. Moazeeni et al [25] reported one of the earliest studies utilizing ML to develop a model capable of predicting LCIs in different areas of a specific oilfield, as well as to estimate the quantity and quality of LCIs.…”

Section: B Literature Reviewmentioning

confidence: 99%

Deep Learning and Time-Series Analysis for the Early Detection of Lost Circulation Incidents During Drilling Operations

et al. 2021

View full text Add to dashboard Cite

Drilling operations consist of breaking the rock to deepen a wellbore for oil or gas extraction. A drilling fluid, circulating from the surface through the drill pipe and from the annulus to the surface, is used to remove rock cuttings and maintain hydrostatic pressure. Drilling fluid lost circulation incidents (LCIs) are major sources of non-productive time (NPT) in drilling operations. These incidents occur due to preexisting natural fractures (vugs, caverns, etc.) and/or drilling-induced hydraulic fractures. The initiation of an LCI could lead to other hazardous drilling phenomena, such as formation influx or kick/blowout, stuck pipe incidents, among others. LCIs are typically monitored at the rig site by observing drilling fluid levels in the fluid tanks. This manual process incurs missing the occurrence or late detection of LCIs. Machine learning (ML) and deep learning (DL) classification algorithms are powerful in processing time-series data and achieving early detection of such temporal phenomena. In this study, we performed a large-scale analysis of the surface drilling and rheology data obtained from historical wells with LCIs. This analysis includes primary and secondary preprocessing steps including, aggressive sampling, feature engineering, and window normalization to derive generalizable DL models for real-time operations. Focal loss was utilized to account for data class imbalance and train robust and generalizable models. The results obtained from different ML/DL algorithms showed that one-dimensional convolutional neural network models resulted in the best performance with state-of-the-art precision, recall, and F1 scores of 87.34%, 73.40%, and 79.77%, respectively, on unseen test drilling data.

show abstract

“…Machine learning methods seem to be able to replace the methods recommended by the experience of field engineers. These computationally intelligent methods can learn from the experience of plugging projects in this area and give more reasonable recommended methods [27].…”

Section: Introductionmentioning

confidence: 99%

Prediction of the Control Effect of Fractured Leakage in Unconventional Reservoirs Using Machine Learning Method

Song

et al. 2022

Geofluids

View full text Add to dashboard Cite

Bridging plugging is the most used method of plugging in unconventional oil reservoirs, and many factors affect the effect of bridging and plugging. Since the laboratory cannot simulate the actual leakage size of the lost formation and the corresponding leakage plugging process at the drilling site, the laboratory experiment results cannot reflect the actual leakage plugging construction effect. Aiming at the problem of frequent fracture leakage during drilling in Chepaizi block, Xinjiang, China, this paper proposes a set of machine learning methods based on a neural network. Three types of factors and 14 parameters with a strong correlation with the leakage control effect were screened out. Three categories of factors include construction parameters, choice of plugging material, and fluid properties of the carrier fluid. The training was carried out based on the collected field data, the appropriate activation function was set, and the deep well network structure was optimized. By improving the field plugging measures in the later period, the model was verified by these actual cases, and the results showed that the established model produced the highest R 2 of 0.974, has a good fit, and predicts well.

show abstract

Intelligent decisions to stop or mitigate lost circulation based on machine learning

Cited by 63 publications

References 31 publications

Research on an Lost Circulation Zone Location Method Based on Transient Pressure Wave

Research on an Lost Circulation Zone Location Method Based on Transient Pressure Wave

Deep Learning and Time-Series Analysis for the Early Detection of Lost Circulation Incidents During Drilling Operations

Prediction of the Control Effect of Fractured Leakage in Unconventional Reservoirs Using Machine Learning Method

Contact Info

Product

Resources

About