Adaptive neural network filtering device for enhanced downhole oilfield measurements

Slickline-deployed, instruments are commonly used to perform measurement and. service operations in oilwells. In order to accurately determine the downhole position of a suspended instrument, a device known as an electro-mechanical casing collar locator (EMCL) is sometimes used. The device is first lowered into a well, and then, slowly withdrawn. When a casing collar is,encountered, the EMCL device causes changes in the line tension. These changes in line tension, which are measured at the surface, are correlated to collar locations, and hence, instrument depth. Like many other applications, noise from different sources during slickline jobs may add contaminating noise or cause destructive interference in the monitored of tension signals. In this paper, a method of using an adaptive neural network to filter the tension signal to remove unwanted noise is described. A theoretical discussion and the review of results of experimental testing are presented.