Downhole mud loss transient simulation and detection with downhole dual measurement points

Abstract Mud loss is usually challenging to discover in time by detecting the mud pit's liquid level, leading to a significant wastage of time and money. Due to the measurement-while-drilling system's low data-transmission speed, downhole measured data is usually ignored in mud loss detection. What is more, surface detection methods comprising pressure and flow-rate sensors require professional knowledge and many input parameters, some of which are required to be assumed. In this study, we proposed a downhole mud loss detection method based on downhole dual measurement points without any use of surface input parameters, which significantly increased the detection method's feasibility. Firstly, we established an end-to-end neural network to recognize the downhole ten working conditions. Once the static and circulation working conditions are recognized, the drilling fluid's density and rheological parameters can be automatically determined with the proposed method. Then, an unscented Kalman filter was applied to perform a backward calculation of mud loss rate dynamically. Thirdly, we proposed a new pressure difference measurement method with high precision, capturing the flow friction pressure consumption in a short distance. To evaluate the mud loss detection method comprehensively, we built the downhole mud loss transient simulation model to generate transient data with the highest possible accuracy. The characteristic lines method was used to solve the partial differential equations of the mud loss simulation model. The accuracy of the simulation model was successfully validated with a small-scale laboratory experiment in the horizontal direction, a full-scale laboratory experiment in the vertical, and a real field experiment. Finally, we generated a set of mud loss well data to describe the proposed mud loss detection method's processes in detail. The results showed that the proposed mud loss detection model successfully detected mud loss and obtained the accurate mud loss rate.