For the first time in design optimization of microwave circuits, the aggressive ( ) space mapping SM optimization technique is applied to automatically align electromag-( ) netic EM models based on hybrid mode-matching r r r r r network theory simulations with ( ) models based on finite-element FEM simulations. SM optimization of an H-plane resonator filter with rounded corners illustrates the advantages as well as the challenges of the approach. The parameter extraction phase of SM is given special attention. The impact of selecting responses and error functions on the convergence and uniqueness of parameter extraction is discussed. A statistical approach to parameter extraction involving l l and 1 penalty concepts facilitates a key requirement by SM for uniqueness and consistency. A multipoint parameter extraction approach to sharpening the solution uniqueness and improving the SM convergence is also introduced. Once the mapping is established, the effects of manufacturing tolerances are rapidly estimated with the FEM accuracy. SM has also been successfully applied to optimize waveguide transformers using two hybrid modematching r r r r r network theory models: a coarse model using very few modes and a fine model using many modes to represent discontinuities.
The advent of deep-reading directional electromagnetic logging-while-drilling measurements brings opportunities in imaging complex formations around the wellbore. In this context, basing the inversions on an assumption of a layered medium is not always sufficient to decouple geometrical effects in complex formation scenarios.Rather than attempting to invert all variables of a generic 2D formation model simultaneously, we describe in this paper how the use of 2D inversions can be enabled through a robust workflow specifically designed for a class of formation models, in this case the angular unconformity. We demonstrate how the successive application of 1D inversion to subsets of tool data gradually builds a formation model for use as input to a 2D inversion, which accounts for complex couplings present in the data in such environments.The validity and the performance of the method are established through synthetic and field datasets.
IntroductionSince the introduction in 2005 of deep and directional logging-while-drilling (LWD) electromagnetic (EM) measurements, thousands of wells have been geosteered based on remote boundary imaging, in a variety of reservoirs and geological environments. The standard processing applied to the measurements is a multilayer 1D inversion. Although this type of processing is valid in a large number of cases and has supported many successful operations, this paper introduces a different way of processing measurements to produce a more structured image of the formation around the wellbore and successfully address nonlayered scenarios in which the risks associated with the errors in running the standard processing might increase beyond desirable levels.We present a processing workflow running automatically through multiple inversion steps and minimizing the bias introduced by users. The steps gradually build a richer model of the subsurface, while introducing measurements with sensitivities matching the model features added at each step. The concept was implemented and tested on a specific family of nonlayered scenarios, the angular unconformity.The first sections summarize the LWD tools and the traditional 1D and 2.5D forward modeling and inversion concepts used in the workflow and throughout the industry. Then, the characteristics of the angular unconformity scenarios are defined, before some arguments are given to justify the need for an imaging workflow rather than a single inversion in the first place as is standard. The workflow steps are detailed next, and in the final sections the results and stability on synthetic and field examples are discussed.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.