Failure Prediction for Rod Pump Artificial Lift Systems

SPE Western Regional Meeting

Jaiswal²,

Yao

et al. 2015

Self Cite

This paper presents the results of using autoencoder-derived features, rather than hand-crafted features, for predicting rod pump well failures using Support Vector Machines (SVMs). Features derived from dynamometer card shapes are used as inputs to the SVM algorithm. Hand-crafted features can lose important information whereas autoencoder-derived abstract features are designed to minimize information loss. Autoencoders are a type of neural network with layers organized in an hourglass shape of contraction and subsequent expansion; such a network eventually learns how to compactly represent a data set as a set of new abstract features with minimal information loss. When applied to card shape data, we demonstrate that these automatically derived abstract features capture high-level card shape characteristics that are orthogonal to the hand-crafted features. In addition, we provide experimental results showing improved well failure prediction accuracy by replacing the hand-crafted features with more informative abstract features.

Section: Introductionmentioning

confidence: 99%

Autoencoder-derived Features as Inputs to Classification Algorithms for Predicting Well Failures

SPE Western Regional Meeting

Jaiswal²,

Yao

et al. 2015

Self Cite

“…Generally, the rod pump failures can be categorized into two types, namely, mechanical and chemical failures [1]. The mechanical failures are mainly caused by improper design, improper manufacturing, normal wear and tear during operations, and excessive wear and tear due to sand intrusions, gas pounding, and asphalting.…”

Section: Introductionmentioning

confidence: 99%

“…These workovers typically shut down the systems, leading to substantial downtime, production loss and operation expense, in addition to the regular maintenance cost. Quick and correct identification of well failure and scheduling appropriate maintenance will reduce wrongly planned repairs and minimize downtime, and subsequently improve the efficiency [1].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Well Failure Detection for Rod Pump Artificial Lift System Through Pattern Recognition

Patel

2013

All Days

Rod pump is by far the most widely used artificial lift method within the oil industry to improve the oil production for reservoirs with energy levels too low to lift fluids to the surface. It is roughly responsible for about two-thirds of the artificial lift producing oil wells. Well failures, including surface, reservoir and down-hole failures, in rod pump artificial lift systems commonly occur, leading to substantial downtime, production loss and operation expense. Correct identification of well failures and scheduling appropriate maintenance will reduce inappropriate repairs, minimize downtime and subsequently improve operation efficiency.This paper presents a systematic Pattern Recognition approach to detect the well failures. In this method, we start with the pattern recognition workflow, explain how we choose the right data and define good features based on the domain knowledge and how we extract features utilizing signal processing techniques. Finally, we will describe how we establish the classifiers via heuristics instead of training. Through this practical application of pattern recognition, we describe some fundamental concepts and steps to build a successful application using pattern recognition technologies.This method was firstly tested on one year history data of 100 rod pump production wells from an active oilfield and the results were compared with the state-of-the-art method from Liu et al (2010); and then on six months of history data of 100 rod pump production wells. The pilot test results showed that our method could correctly identify the well failure with success rate of 82-86% and only 11-15% of false alarm. Comparing to Liu's the state-of-the-art method (2010), our method has much lower false alarm rate while with similar success rate of detecting failure, thus confirming that our systematic method is more capable of detecting well failures successfully.

“…However, the POCs by themselves are not sufficient. In the fields studied in this paper the sheer number of these units compared to the number of personnel still demands a great deal of time and effort in terms of monitoring each and every unit operating [8].…”

Section: Introductionmentioning

confidence: 99%

Automatic Early Fault Detection for Rod Pump Systems

Liu¹,

Raghavendra

SPE Annual Technical Conference and Exhibition

et al. 2011

Typically, rod pump system failures are determined using the dynamometer card which may miss some early warnings. This paper presents a novel approach for early failure detection in rod pump wells using more than 14 parameters that indicate the daily functions of rod pump wells and employs advanced machine learning techniques. Our system recognizes failing, failed as well as normal situations by learning their patterns/signature from historical pump data, that include card area, peak-surface load, minimum-surface load, daily run-time, and production data. These data are automatically pre-processed using expert domain knowledge to reduce noise and to fill-in missing data. Our approach is novel in two ways. First, our machine learning algorithm AdaBNet uses boosting to learn several Bayesian Network models and then combines these models with different weights to form a stronger boosted model. Second, our approach generates this single boosted model that is applicable across all the wells in a field, as opposed to well-specific approaches that generate one model per well. This model detects anomalies, prefailure and failure signals and generates corresponding alerts. Early fault detection in rod pump wells is useful for automatic monitoring of large number of assets remotely, and could be extended to other artificial lift systems. We used a training data set of 12 wells to construct the learning model for the AdaBNet algorithm and tested the algorithm on 426 wells from the same field. The results show that our algorithm detects failures with accuracy higher than 90%. This framework can help field operators not only to remotely recognize and predict failures in advance, but also to help prioritize the available manpower, save significant time, reduce operating expense (OPEX), downtime and lost production. Early fault detection in rod pump systems can allow for proactive maintenance that can delay and even prevent future well failures. The proposed algorithm can enable production engineers remotely detect failures and anomalies before they occur, and assess the situation at control centers before taking any remedial or corrective actions. This approach to using a single model for an entire field is superior to other approaches with individual model for each well.