Incident detection and isolation in drilling using analytical redundancy relations

Abstract

Early diagnosis of incidents that could delay or endanger a drilling operation for oil or gas is essential to limit field development costs. Warnings about downhole incidents should come early enough to allow intervention before it develops to a threat, but this is difficult, since false alarms must be avoided. This paper employs model-based diagnosis using analytical redundancy relations to obtain residuals which are affected differently by the different incidents. Residuals are found to be non-Gaussian -they follow a multivariate t-distribution -hence, a dedicated generalized likelihood ratio test is applied for change detection. Data from a 1400 meter horizontal flow loop test facility is used to assess the diagnosis method. Diagnosis properties of the method are investigated assuming either with available downhole pressure sensors through wired drill pipe or with only topside measurements available. In the latter case, isolation capability is shown to be reduced to group-wise isolation, but the method would still detect all serious events with the prescribed false alarm probability.

Adaptive Change Detection for Long-Term Machinery Monitoring Using Incremental Sliding-Window

“…In addition, it is also observed that on the basis of considering the projection coefficient K from RKHS to Figure 3 Experimental setup for collection of testing data: (a) Front view; (b) Schematics martingale space and the global fixed confidence level a, the threshold value can be computed/adjusted adaptively according to the standard deviation computed from the past data at each step of change decision making as mentioned in Eq. (11), in other words, the threshold value is not fixed in the whole process, that is different from many existing works [7][8][9][10][11]25]. All of them are consistent with the analysis previously made in Section 3.2.…”

“…To address this problem, unlike existing works that directly used the original monitored variables for change detection (e.g., Refs. [6][7][8][9][10][11][12]25]), we utilize the Hilbert space embedding of distribution (HED, also called the kernel mean or mean map) to map the original data fx i g, i 2 f1; 2; . .…”

“…Potential weakness of the majority of exiting algorithms, e.g., Refs. [5][6][7][8][9][10][11], is that they need either human-made instruction/ intervention or an off-line cross-validation to confirm the value of threshold before operation, and thus make them largely limited in real applications. Another contribution of this paper is to develop a new adaptive threshold when performing change decision making.…”

“…Change decision can be made by performing a null hypothesis testing with a threshold. Many effective tools for this goal such as the cumulative sum metric (CUSUM) [3][4][5][6], geometric moving average (GMA) [7,8] and the generalized likelihood ratio test (GLRT) [9][10][11][12] have been widely used. For example, Willersrud et al [12] developed the GLRT to make efficient downhole drilling washout detection with the multivariate t-distribution; Reñones et al [6] used the CUSUM analysis for multi-tooth machine tool fault detection.…”

A Statistical Fitness Test of Newell's 3-detector Simplification Method for Unexpected Incident Detection in the Expressway Traffic Flow

“…And they developed failure time survival model to capture heterogeneous incident variables with fixed and random specifications. Willersrud et al(2015) made a diagnosis using analytical redundancy relations to obtain residuals from the different incidents effects. Analysing data was extracted from a horizontal flow loop facilities.…”