/* This file is part of Lemma, a geophysical modelling and inversion API */ /* This Source Code Form is subject to the terms of the Mozilla Public * License, v. 2.0. If a copy of the MPL was not distributed with this * file, You can obtain one at http://mozilla.org/MPL/2.0/. */ /** @file @author Trevor Irons @date 06/19/2009 09:12:20 AM The Birth of Lemma! @version $Id: lemma.h 203 2015-01-09 21:19:04Z tirons $ **/ // \image html lemma.png /** \mainpage Lemma is an ElectroMagnetics Modelling API \authors Trevor Irons and M. Andrew Kass and others Originally Lemma was intended as a recursive acronym standing for Lemma is an ElectroMagnetics Modelling API. As the breadth of the project has expanded, the name has remained appropriate in a more literal sense. Lemma is a flexible cross-platform library delivering an expressive API that can be used to easily create versatile programs. Lemma is not itself a program, instead it is a collection of building blocks to make geophysical applications. We retain this name because: - In mathematics a Lemma is a proven proposition which is used as a stepping stone to a larger result rather than as a statement in-and-of itself. - In addition to the electromagnetic modelling, some other facilities are provided such as numerical optimization and inversion capabilities. These tools are also considered stepping stones to final products. We feel that this is a particularly approprate name, as Lemma's API can be leveraged create powerful applications such as forward modelling and inverting frequency and time-domain surveys of arbitrary survey design, sNMR surveys, CSAMT and more. \section Motivation Why another Geophysical EM project? For starters, there aren't that many quality open source packages out there. Those that do exist are generally specialized to perform a single task and extending them is a major undertaking. Lemma's approach is much different, by providing a set of general tools users can easily assemble applications that suite their needs. Furthermore, most are written in either Fortran or MATLAB, and can be difficult to integrate into multiphysics applications. \section Capabilities Capabilities In the long term, we have many goals for this software project. Due to its design, Lemma can be built upon and extended easily. The initial aim is to provide flexible 1D and 3D EM modelling in the time and frequency domains. The project is still in beta, but we have made a lot of progress already. We will release our first non-beta release as soon as the following are supported. \subsection FDM Frequency-domain forward modelling Lemma was initially called EMMODFD: Electromagnetic Modelling in the Frequency Domain. As such this is the most mature area of Lemma. \par 1D Frequency domain solutions to electrical and magnetic dipoles can be computed quasi-analytically in 1D. Calculations can be made in or above the layered media, and complex electrical conductivity and magnetic susceptibility are supported according to the Cole-Cole model. Sources may be embdedded in the media or in the resisitive air layer. Lemma can also can compute fields due to arbitrarily shaped ungrounded wire loops, topography of the loops is also supported. Two separate approaches to solving the Hankel transform, one based on Anderson's digitial filtering technique, and another based on Gaussian quadrature. \par 3D A fast 3D solver that can modify the 1D results based on arbitrary electrical conductivity model is nearing completion. \par future work We are also planning on supporting grounded wires in the near future. \subsection TDM Time-domain forward modelling A 1D time-domain solution has been implemented that utilises both a dipole source as well as a wire loop. Currently, only one receiver is modelled at a time, but will be generalised. In addition, utilities to read in data files for modelling have been implemented. We would like to offer 3D time domain support, but this will not be provided before our first stable release. \subsection DataFormats Data Formats The EM community is plagued with myriad data formats. Often each equiptment manufacturer provides their own data format and interoperability is a constant struggle. We are working on a flexible data format based on the XML format that can be adapted to many types of data. The template for this format will be publically released and we hope it catches on in the community. At the least, it will provide a mechanism to compare datasets and datatypes within Lemma. \section Modules Modules Due to Lemma's design, it is easy to extend the platform. In some cases this extension results in adding functionality that is not directly related to ElectroMagnetics. The following modules utilise parts of Lemma to provide their functionality. \section Tutorials - \ref Tutorial - Basic intruduction to Lemma, including aquiring and compiling the code, class structure, and building your own applications. - \ref Extending Tutorial on how to extend Lemma. \section Development Development and design Ths package was initially developed by the Center for Gravity, Electrical, and Magnetic Studies (CGEM) at the Colorado School of Mines (CSM), the United States Geological Survey (USGS), and Broken Spoke Development, LLC. Where it drew on work by many others including Ki Ha Lee, and Walt Anderson. All new work and interfaces are written entirely in C++. Several small external projects are included, which are written in standard C, and FORTRAN 77. We adapt a modern, test driven, object oriented, C++ framework. More recent development has been undertaken at the University of Utah through the Energy and Geoscience Institute. \section Legalities \subsection Copyrights The following copyrights apply to the source. Most of the code was developed either by Trevor Copyright (C) 2008-2010 Trevor Irons or M. Andrew Kass Copyright (C) 2010 . The 1D EM solver was derived (but updated heavily) from a Fortran programme written by Ki Ha Lee in 1984. We have communicated with Ki Ha, and he assured us that this code is in the public domain. A Gaussian quadrature hankel transform originally written by Alan Chave was ported to C++. This code is in the public domain, and the source code was published in Geophysics. A digital filtering approach to the Hankel transform written by Walt Anderson was also rewritten for Lemma. The original Fortran code is also in the public domain. Please note that Ki Ha Lee and Walt Anderson had no part in this work, and the above should not be interpreted as any sort of endorsement by those parties. \subsection License This Source Code Form is subject to the terms of the Mozilla Public License, v. 2.0. If a copy of the MPL was not distributed with this file, You can obtain one at http://mozilla.org/MPL/2.0/. \section Contributing Suggestions and contributions We welcome contributions and suggestions. Feel free to email the development team at info@lemmasoftware.org. Under the terms of the MPL, if you modify a Lemma file, you are obligated to share those contributions back with the community. \section Useful Useful links - Home page https://lemmasoftware.org - Git repository https://git.lemmasoftware.org - Broken Spoke Develpment http://numericalgeo.com - CGEM at the Coloroado School of Mines http://geophysics.mines.edu/cgem/ - EGI at the Eniversity of Utah http://egi.utah.edu/ **/ #pragma once #ifndef __LEMMA_H #define __LEMMA_H #include // Include some basic stuff that will always be needed #include #include #include #include #include #include #include #include #include #include #include #include #include //#include #include /** \brief The only namespace used by Lemma * * \details The rational behind this namespace is that built-in * types should be used wherever possible, but not * not built-in names. This allows for code that is better * enacsulated and easier to modify. The typedefs and constants * specified here are defined so that * precision/inplimentation can easily be changed. * All floating precision types should be typedefed in this file * and should not be used natively within any code. * Lemma uses * the Eigen Matrix/Vector/Linear Algebra library. * and a lot of the namespece typedefs * are specifying Eigen types. */ namespace Lemma { /// Real defines precision for the whole API, default is double #ifdef LEMMA_SINGLE_PRECISION typedef float Real; #else // ----- LEMMA_SINGLE_PRECISION ----- typedef double Real; #endif // ----- not LEMMA_SINGLE_PRECISION ----- /// Complex version of Real. typedef std::complex Complex; /// A 3 component Eigen vector of Reals typedef Eigen::Matrix Vector3r; /// A 3 X Dynamic Component Eigen matrix of Reals typedef Eigen::Matrix Vector3Xr; /// Variable length Eigen vector of Reals typedef Eigen::Matrix VectorXr; /// Variable length Eigen vector of integers (int) typedef Eigen::Matrix VectorXi; /// Variable length Eigen vector of Complexes typedef Eigen::Matrix VectorXcr; /// A 3 Component Eigen vector of Complexes typedef Eigen::Matrix Vector3cr; /// A 3 X Dynamic Component Eigen matrix of Complexes typedef Eigen::Matrix Vector3Xcr; /// Variable length Eigen Matrix of Reals typedef Eigen::Matrix MatrixXr; /// Variable length Eigen Matrix of ints typedef Eigen::Matrix MatrixXi; /// Variable length Eigen vector of Complexes typedef Eigen::Matrix MatrixXcr; //////////////////////////////////////// // Constants used across the programmes /// Restating the obvious, this is pi const Real PI = 4.0*atan(1.0); /// Permitivity of Free Space //const Real EPSILON0 = 8.854187817e-12; const Real EPSILON0 = 8.854187817e-12; /// Permeability of free space const Real MU0 = 4.*PI*1e-7; /// 1/4 of \f$ \pi\f$ const Real QPI = .25/PI; /// Some functions will convert units from SI (standard) to Gauss /// This is because NMR calculations are much more natural in Gauss enum MAGUNITS {TESLA, NANOTESLA, GAUSS}; /// Unit of temperature entered enum TEMPUNITS {CELCIUS, KELVIN}; /// Unit of time entered enum TIMEUNITS {SEC, MILLISEC, MICROSEC, NANOSEC, PICOSEC}; /// Unit of time entered enum FREQUENCYUNITS {HZ, KHZ, MHZ, GHZ}; /// FEM coil relative orientations enum FEMCOILORIENTATION {COAXIAL, COPLANAR}; /// General orientation relative to coordinate system enum ORIENTATION {X, Y, Z, NX, NY, NZ}; /// Type of field enum FIELDTYPE {HFIELDREAL, HFIELDIMAG, EFIELDREAL, EFIELDIMAG}; /// Compenent of vector field enum FIELDCOMPONENT {XCOMPONENT=0, YCOMPONENT=1, ZCOMPONENT=2}; /// Spatial component of vector enum SPATIALCOORDINANT {XCOORD=0, YCOORD=1, ZCOORD=2}; /** Evaluation method for Hankel integrals. * ANDERSON801 Walt Anderson's 801 point filter * CHAVE Alan Chave's gaussian quadrature integration method * FHTKEY201 Key's 201 point filter * FHTKEY201 Key's 101 point filter * FHTKEY51 Key's 51 point filter * QWEKEY Key's Gaussian quadrature integration method */ enum HANKELTRANSFORMTYPE { ANDERSON801, CHAVE, FHTKEY201, FHTKEY101, FHTKEY51, QWEKEY, FHTKONG61, FHTKONG121, FHTKONG241, IRONS }; /** Enum is OK because these are the only physically possible sources. @param NOSOURCETYPE is default. @param ELECTRICDIPOLE is an electric dipole @param MAGNETICDIPOLE is a magnetic dipole */ enum DipoleSourceType {NOSOURCETYPE, GROUNDEDELECTRICDIPOLE, UNGROUNDEDELECTRICDIPOLE, MAGNETICDIPOLE}; /// Only three polarizations are supported. They may be summed to /// approximate others /// @param NOPOLARISATION is uninitialized, default value /// @param XPOLARISATION is a dipole oriented in the x direction /// @param YPOLARISATION is a dipole oriented in the y direction /// @param ZPOLARISATION is a dipole oriented in the z direction enum DipoleSourcePolarisation {NOPOLARISATION, XPOLARISATION, YPOLARISATION, ZPOLARISATION}; /// The polarity may be either negative or positinve enum DipoleSourcePolarity {NEGATIVE, POSITIVE}; /** The fields to make calculations on */ enum FIELDCALCULATIONS {E, H, BOTH}; /** Windowing function type */ enum WINDOWTYPE { HAMMING, /*!< A hamming window */ HANNING, /*!< A hanning window */ RECTANGULAR /*!< Rectangular window */ }; } #endif // __Lemma_H