Controller design for V-Loop
The examples here will consider designing the RST and ILC controllers for the voltage loop. For simplicity, the examples will consider the desired control parameters taken from the default function (see API Reference):
>>> import pyfresco.obcd as cd >>> device = 'RFNA.866.01.ETH2' >>> user_pars_v = cd.get_default_v(device) >>> user_pars_v <pyfresco.obcd.props.UiParams object at 0x7fba77baf7f0> >>> vars(user_pars_v) {'damp_bw': 75, 'damp_z': 0.8, 'volt_mm': 0.5, 'opt_method': 'Hinf', 'volt_bw': 50, 'volt_z': 0.8, 'ref_delay': 0, 'control_mode': 'V', 'debug': False, 'positive_coeff': False, 'kd_0': False} >>> fgc_v = cd.FgcProperties.from_fgc_v(user_pars_v, device) >>> fgc_v <pyfresco.obcd.props.FgcProperties object at 0x7fba77baf0f0> >>> vars(fgc_v) {'LOAD.OHMS_SER': 9.9999997e-06, 'LOAD.OHMS_MAG': 0.13600001, 'LOAD.OHMS_PAR': 100000000.0, 'LOAD.HENRYS': 0.0063999998, 'LOAD.GAUSS_PER_AMP': 1.0, 'VS.ACTUATION': 'VOLTAGE_REF', 'VS.FIRING.DELAY': 0.00166667, 'VS.ACT_DELAY_ITERS': 1.5, 'FGC.ITER_PERIOD': 9.9999997e-05, 'VS.FILTER.FARADS1': 0.039999999, 'VS.FILTER.FARADS2': 0.0099999998, 'VS.FILTER.OHMS': 0.0070000002, 'VS.FILTER.HENRYS': 0.00060000003, 'REG.I.INTERNAL.PURE_DELAY_PERIODS': 0.0, 'MEAS.I.DELAY_ITERS': 0.51999998, 'MEAS.V.DELAY_ITERS': 0.51999998}
Model-based design (V)
The model-based design uses the FGC properties to design a controller. It is thus imperative that the correct load properties are set in the FGC. This method can be used in cases where a frequency response measurement cannot be obtained from a particular power converter. The model-based design can be completed by simply calling the following functions:
>>> X = cd.OptimizeV(fgc_v, user_pars_v) >>> opt_result, df_sens, margins = X.model_opt() >>> vars(opt_result) {'Kd': [0.2702714426345273, 0.015855706981816254, 1.3441402655593147], 'Kv': [0.14642296449100214, 108.34838485558069, 0.3202947134041462]}
The opt_result object contains both the damping loop and voltage loop parameters, given by Kd
and Kv, respectively. The results from the optimization problem can then be sent to the FGC as follows:
>>> cd.FgcProperties.to_fgc_v(opt_result, user_pars_ib, device)
The output dataframe df_sens contains the closed-loop sensitivity functions, and is discussed in
Sensitivity functions. Finally, margins contains the calculated robustness
margins (in dictionary format) of the closed-loop system: modulus_margin, gain_margin,
phase_margin and delay_margin.
Data-driven design (V)
To perform a data-driven design for the voltage loop, three frequency responses are needed:
F_REF_LIMITED to I_CAPA
F_REF_LIMITED to V_MEAS_REF
F_REF_LIMITED to I_MEAS
In addition to these frequency responses, the DC gain of the load is also required; let us assume that a PRBS experiment was performed with the following log data:
>>> df_vreg t_global V_MEAS_REG_(V) F_REF_LIMITED_(V) I_CAPA_(A) - t_local sample t_local sample t_local sample 0 1.597837e+09 1.597837e+09 0.000000 1.597837e+09 0.0 1.597837e+09 -3.763275e-21 1 1.597837e+09 1.597837e+09 0.000000 1.597837e+09 0.0 1.597837e+09 -3.726354e-21 2 1.597837e+09 1.597837e+09 0.000000 1.597837e+09 0.0 1.597837e+09 -3.687887e-21 3 1.597837e+09 1.597837e+09 0.000000 1.597837e+09 0.0 1.597837e+09 -3.647898e-21 4 1.597837e+09 1.597837e+09 0.000000 1.597837e+09 0.0 1.597837e+09 -3.606406e-21 ... ... ... ... ... ... ... ... >>> df_imeas1 t_global I_MEAS_(A) V_REF_(V) V_MEAS_(V) - t_local sample t_local sample t_local sample 0 1.597837e+09 1.597837e+09 0.000000 1.597837e+09 0.0 1.597837e+09 0.000000 1 1.597837e+09 1.597837e+09 0.000000 1.597837e+09 0.0 1.597837e+09 0.000000 2 1.597837e+09 1.597837e+09 0.000000 1.597837e+09 0.0 1.597837e+09 0.000000 3 1.597837e+09 1.597837e+09 0.000000 1.597837e+09 0.0 1.597837e+09 0.000000 4 1.597837e+09 1.597837e+09 0.000000 1.597837e+09 0.0 1.597837e+09 0.000000 ... ... ... ... ... ... ... ... >>> df_dcgain t_global I_MEAS_(A) V_REF_(V) V_MEAS_(V) - t_local sample t_local sample t_local sample 0 1.597831e+09 1.597831e+09 7.352385 1.597831e+09 1.0 1.597831e+09 0.999998 1 1.597831e+09 1.597831e+09 7.352385 1.597831e+09 1.0 1.597831e+09 0.999998 2 1.597831e+09 1.597831e+09 7.352385 1.597831e+09 1.0 1.597831e+09 0.999998 3 1.597831e+09 1.597831e+09 7.352385 1.597831e+09 1.0 1.597831e+09 0.999998 4 1.597831e+09 1.597831e+09 7.352385 1.597831e+09 1.0 1.597831e+09 0.999998 ... ... ... ... ... ... ... ...
The frequency responses and DC gain can then be obtained as follows:
>>> import pyfresco.frm as frm >>> df_vmeas = frm.prbs(df_vreg, prbs_vmeas, device) >>> df_icapa = frm.prbs(df_vreg, prbs_icapa, device) >>> df_imeas = frm.prbs(df_imeas1, prbs_imeas, device)
The parameters prbs_vmeas, prbs_icapa, and prbs_imeas are simply the objects containing the
UI input parameters (see PRBS Example).
The data-driven design can then be completed by simply calling the following functions:
>>> X = cd.OptimizeV(fgc_v, user_pars_v) >>> opt_result, df_sens, margins = X.data_opt(df_icapa, df_vmeas, df_imeas, df_dcgain) >>> vars(opt_result) {'Kd': [0.30363474386072414, 0.014489922006368857, 1.2091239687250508], 'Kv': [-0.046388735100899886, 47.702158277587756, 0.7295920508127712]}