Programmatically Extracting g-function/lntts Data

[mitchute]

What's the best way to extract g-function (STS and LTS) and ln(t/ts) values for a particular solution of the bisection search?

For example, I have this:

    # Single U-tube
    # -------------
    # Perform field selection using bisection search
    coordinates_domain = dt.domains.square_and_near_square(1, 64, bh_spacing)

    bisection_search = dt.search_routines.Bisection1D(coordinates_domain, vol_flow_rate, borehole, single_u_tube,
                                                      fluid, pipe_single, grout, soil, sim_params,
                                                      loads, disp=False)

    # Find the near-square design for a single U-tube and size it.
    ghe = bisection_search.ghe
    ghe.compute_g_functions()

I see that the ghe variable (which looks like it is an instance of the GFunction class) has self.g_lts and self.log_time member variables. And the BaseGHE class has combine_sts_lts and grab_g_function methods.

What do you recommend?

[j-c-cook]

Check out how I exported the g-functions in this oak_ridge_export function. I limit the number of short time step g-function values and sample it because either OpenStudio or EnergyPlus didn't like all of them.

ghedt/ghedt/search_routines.py

Lines 385 to 405 in 3a4ad15

	# create a local ghe object
	ghe = bisection_search.ghe
	H = ghe.bhe.b.H
	B_over_H = ghe.B_spacing / H
	g = ghe.grab_g_function(B_over_H)
	total_g_values = g.x.size
	number_lts_g_values = 27
	number_sts_g_values = 30
	sts_step_size = int(np.floor((total_g_values - number_lts_g_values) /
	number_sts_g_values).tolist())
	lntts = []
	g_values = []
	for i in range(1, (total_g_values - number_lts_g_values),
	sts_step_size):
	lntts.append(g.x[i].tolist())
	g_values.append(g.y[i].tolist())
	lntts += g.x[
	total_g_values - number_lts_g_values: total_g_values].tolist()
	g_values += g.y[
	total_g_values - number_lts_g_values: total_g_values].tolist()