Diffusion

enum tug::APPROACH

Enum defining the implemented solution approaches.

Values:

enumerator FTCS_APPROACH

Forward Time-Centered Space

enumerator BTCS_APPROACH

Backward Time-Centered Space

enumerator CRANK_NICOLSON_APPROACH

Crank-Nicolson method

enum tug::SOLVER

Enum defining the Linear Equation solvers.

Values:

enumerator EIGEN_LU_SOLVER

EigenLU solver

enumerator THOMAS_ALGORITHM_SOLVER

Thomas Algorithm solver; more efficient for tridiagonal matrices

enum class tug::CSV_OUTPUT

Enum holding different options for .csv output.

Values:

enumerator OFF

do not produce csv output

enumerator ON

produce csv output with last concentration matrix

enumerator VERBOSE

produce csv output with all concentration matrices

enumerator XTREME

csv output like VERBOSE but additional boundary conditions at beginning

enum class tug::CONSOLE_OUTPUT

Enum holding different options for console output.

Values:

enumerator OFF

do not print any output to console

enumerator ON

print before and after concentrations to console

enumerator VERBOSE

print all concentration matrices to console

enum class tug::TIME_MEASURE

Enum holding different options for time measurement.

Values:

enumerator OFF

do not print any time measures

enumerator ON

print time measure after last iteration

template<class T, APPROACH approach = BTCS_APPROACH, SOLVER solver = THOMAS_ALGORITHM_SOLVER>
class Diffusion : public tug::BaseSimulationGrid<T>

The class forms the interface for performing the diffusion simulations and contains all the methods for controlling the desired parameters, such as time step, number of simulations, etc.

Template Parameters:

• T – the type of the internal data structures for grid, boundary condition and timestep

• approach – Set the SLE scheme to be used

• solver – Set the solver to be used

Public Functions

inline Diffusion(RowMajMat<T> &origin)

Construct a new Diffusion object from a given Eigen matrix.

inline Diffusion(T *data, int rows, int cols)

Construct a new 2D Diffusion object from a given data pointer and the dimensions.

inline Diffusion(T *data, std::size_t length)

Construct a new 1D Diffusion object from a given data pointer and the length.

inline RowMajMat<T> &getAlphaX()

Get the alphaX matrix.

Returns:

RowMajMat<T>& Reference to the alphaX matrix.

inline RowMajMat<T> &getAlphaY()

Get the alphaY matrix.

Returns:

RowMajMat<T>& Reference to the alphaY matrix.

inline void setAlphaX(const RowMajMat<T> &alphaX)

Set the alphaX matrix.

Parameters:

alphaX – The new alphaX matrix.

inline void setAlphaY(const RowMajMat<T> &alphaY)

Set the alphaY matrix.

Parameters:

alphaY – The new alphaY matrix.

inline virtual void setTimestep(T timestep) override

Setting the time step for each iteration step. Time step must be greater than zero. Setting the timestep is required.

Parameters:

timestep – Valid timestep greater than zero.

inline T getTimestep() const

Currently set time step is returned.

Returns:

double timestep

inline void setNumberThreads(int numThreads)

Set the number of desired openMP Threads.

Parameters:

num_threads – Number of desired threads. Must have a value between 1 and the maximum available number of processors. The maximum number of processors is set as the default case during Simulation construction.

inline void printConcentrationsConsole() const

Outputs the current concentrations of the grid on the console.

inline std::string createCSVfile() const

Creates a CSV file with a name containing the current simulation parameters. If the data name already exists, an additional counter is appended to the name. The name of the file is built up as follows: <approach> + <number rows> + <number columns> + <number of

iterations>+<counter>.csv.

Returns:

string Filename with configured simulation parameters.

inline void printConcentrationsCSV(const std::string &filename) const

Writes the currently calculated concentration values of the grid into the CSV file with the passed filename.

Parameters:

filename – Name of the file to which the concentration values are to be written.

inline virtual void run() override

Method starts the simulation process with the previously set parameters.