Investigations of Non-Ideal Hook Load Behavior during Tripping Operations for Well G-45
MetadataVis full innførsel
This thesis has analyzed hook load data from well G-45 in Calispera C describing a tripping operation. The purpose of the analysis has been to identify and analyze non-ideal hook load behavior representing non-ideal borehole conditions. Certain borehole problems will affect the hook load signal and this makes the hook load signal a diagnostic tool in the evaluation of borehole problems. The ambition has been to develop borehole diagnostic tools applicable to hook load signals. To introduce a source of verification for the real data a drag model based on well G-45 was built. This model was built upon information provided describing specifications of well G-45, such as the borehole trajectory, drilling fluid density and bottom hole assembly. The model was designed so that it could generate hook loads for different bit depths and different operations, such as for example RIH or POOH. Comparisons of the real hook load data to the predictions generated by the model indicated a relation between the two. This correspondence provided a degree of verification of the model and enabled the model to be used in the identification process of non-ideal hook load behavior. Four distinct non-ideal hook load behaviors in the data were identified and evaluated; startup peaks, fluctuations, kinetic increases and kinetic peaks. Non-ideal features of the startup peaks were detected when several hook load sections containing startup peaks were compared. It became evident that the startup magnitudes varied in a manner ideal borehole conditions would not dictate, the reason for this was attributed to variations in the coefficient of static friction. The fluctuation of numerous kinetic hook load sections with similar fluctuation ranges was found to partially correlate to the acceleration of the block. The origin of the acceleration of the block originates both from non-ideal heterogeneities in the coefficients of kinetic friction along the borehole walls and limitations in the ability of the hoisting system to maintain a constant velocity of the block. The reason why only partial correspondence between the hook load signal and block acceleration was detected was also investigated. The origins of non-ideal hook load increases and peaks were not identified, however the ability of these phenomena to introduce threats to the overall tripping operation initiated the introduction of type curves. Type curves identify the characteristics of specific non-ideal hook load phenomena by describing the development of the hook load with respect to time and the development of the pseudo coefficient of friction with respect to measured bit depth. Type curves were initially introduced by Cardoso (1994), and this can be seen as a continuation of this work. Type curves were not introduced to the non-ideal hook load fluctuations since they pose no immediate threat to the overall tripping operation. Type curves for start up peaks already exist. The process of establishing type curves for specific non-ideal borehole phenomenon enables the characteristics of these non-ideal hook load behaviors to be universally accessible. The type curves established in this thesis can be made applicable as borehole diagnostic tools for real-time and/or historical hook load signals from drilling operations. Data processing methods were also investigated and evaluated based on their application to drilling data. Wavelet filter banks based on the correct wavelet domain have the ability to increase the transmission efficiency of drilling data through optimizing the representation of transients and steady states intervals in the data streams. Wavelet filter banks can also be tuned to detect and remove outliers and noise in the data. Further research is required to investigate the implementation of wavelet filter banks to drilling data.