|
| 1 | +""" |
| 2 | +Script to check FFT of SPS quadrupolar circuit currents |
| 3 | +""" |
| 4 | +import fma_ions |
| 5 | +import numpy as np |
| 6 | +import matplotlib.pyplot as plt |
| 7 | +from scipy.fft import fft, fftfreq |
| 8 | + |
| 9 | +# Simulation parameters |
| 10 | +fs = 1000 # Sampling frequency (Hz) |
| 11 | +duration = 2 # Duration in seconds |
| 12 | +t = np.linspace(0, duration, int(fs * duration), endpoint=False) |
| 13 | +N = len(t) |
| 14 | + |
| 15 | +# Signal parameters |
| 16 | +I_dc = 70.0 # DC current (A) |
| 17 | +I_50hz = 0.001 # 50 Hz component amplitude (A) - at the threshold |
| 18 | +I_100hz = 0.0005 # 100 Hz harmonic (A) |
| 19 | +I_150hz = 0.0003 # 150 Hz harmonic (A) |
| 20 | + |
| 21 | +# Add some broadband noise to simulate other fluctuations |
| 22 | +noise_level = 0.003 # A (RMS) |
| 23 | + |
| 24 | +# Create synthetic DCCT signal |
| 25 | +current = (I_dc + |
| 26 | + I_50hz * np.sin(2 * np.pi * 50 * t) + |
| 27 | + I_100hz * np.sin(2 * np.pi * 100 * t + 0.5) + |
| 28 | + I_150hz * np.sin(2 * np.pi * 150 * t + 1.2) + |
| 29 | + noise_level * np.random.randn(N)) |
| 30 | + |
| 31 | +print(f"Signal statistics:") |
| 32 | +print(f"Mean current: {np.mean(current):.4f} A") |
| 33 | +print(f"Current fluctuations (std): {np.std(current):.4f} A") |
| 34 | +print(f"Peak-to-peak variation: {np.ptp(current):.4f} A") |
| 35 | + |
| 36 | +# Remove DC component (subtract mean) |
| 37 | +current_ac = current - np.mean(current) |
| 38 | + |
| 39 | +# Perform FFT |
| 40 | +fft_result = fft(current_ac) |
| 41 | +frequencies = fftfreq(N, 1/fs) |
| 42 | + |
| 43 | +# Calculate normalized FFT amplitude (amplitude/N) |
| 44 | +fft_amplitude_normalized = np.abs(fft_result) / N |
| 45 | + |
| 46 | +# Only consider positive frequencies |
| 47 | +pos_freq_mask = frequencies >= 0 |
| 48 | +frequencies_pos = frequencies[pos_freq_mask] |
| 49 | +fft_amplitude_pos = fft_amplitude_normalized[pos_freq_mask] |
| 50 | + |
| 51 | +# Find the 50 Hz component |
| 52 | +freq_50hz_idx = np.argmin(np.abs(frequencies_pos - 50)) |
| 53 | +amplitude_50hz_normalized = fft_amplitude_pos[freq_50hz_idx] |
| 54 | +amplitude_50hz_actual = amplitude_50hz_normalized * 2 # Convert back to actual amplitude |
| 55 | + |
| 56 | +print(f"\nFFT Analysis:") |
| 57 | +print(f"50 Hz normalized FFT amplitude: {amplitude_50hz_normalized:.6f}") |
| 58 | +print(f"50 Hz actual current amplitude: {amplitude_50hz_actual:.6f} A") |
| 59 | +print(f"Threshold (5e-4): {'PASS' if amplitude_50hz_normalized < 5e-4 else 'FAIL'}") |
| 60 | + |
| 61 | +# Create plots |
| 62 | +fig, (ax1, ax2, ax3) = plt.subplots(3, 1, figsize=(12, 10)) |
| 63 | + |
| 64 | +# Time domain plot |
| 65 | +ax1.plot(t[:500], current[:500], 'b-', linewidth=0.8) |
| 66 | +ax1.set_xlabel('Time (s)') |
| 67 | +ax1.set_ylabel('Current (A)') |
| 68 | +ax1.set_title('DCCT Current Measurement (first 0.5 seconds)') |
| 69 | +ax1.grid(True, alpha=0.3) |
| 70 | +ax1.set_ylim([I_dc - 0.02, I_dc + 0.02]) |
| 71 | + |
| 72 | +# AC component time domain |
| 73 | +ax2.plot(t[:500], current_ac[:500], 'r-', linewidth=0.8) |
| 74 | +ax2.set_xlabel('Time (s)') |
| 75 | +ax2.set_ylabel('AC Current (A)') |
| 76 | +ax2.set_title('AC Component (DC removed)') |
| 77 | +ax2.grid(True, alpha=0.3) |
| 78 | + |
| 79 | +# Frequency domain plot |
| 80 | +ax3.semilogy(frequencies_pos[:N//10], fft_amplitude_pos[:N//10], 'g-', linewidth=1) |
| 81 | +ax3.axhline(y=5e-4, color='r', linestyle='--', label='Threshold (5e-4)') |
| 82 | +ax3.axvline(x=50, color='orange', linestyle=':', label='50 Hz') |
| 83 | +ax3.set_xlabel('Frequency (Hz)') |
| 84 | +ax3.set_ylabel('Normalized FFT Amplitude') |
| 85 | +ax3.set_title('FFT Analysis (Normalized to Number of Samples)') |
| 86 | +ax3.grid(True, alpha=0.3) |
| 87 | +ax3.legend() |
| 88 | +ax3.set_xlim([0, 200]) |
| 89 | +ax3.set_ylim([1e-6, 1e-2]) |
| 90 | + |
| 91 | +plt.tight_layout() |
| 92 | +plt.show() |
| 93 | + |
| 94 | +# Additional analysis: Find all significant peaks |
| 95 | +significant_peaks = [] |
| 96 | +for i, freq in enumerate(frequencies_pos): |
| 97 | + if 1 <= freq <= 500 and fft_amplitude_pos[i] > 1e-5: # Above noise floor |
| 98 | + significant_peaks.append((freq, fft_amplitude_pos[i], fft_amplitude_pos[i] * 2)) |
| 99 | + |
| 100 | +print(f"\nSignificant frequency components:") |
| 101 | +print(f"{'Frequency (Hz)':<15} {'Normalized Amp':<15} {'Actual Amp (A)':<15}") |
| 102 | +print("-" * 45) |
| 103 | +for freq, norm_amp, actual_amp in sorted(significant_peaks, key=lambda x: x[1], reverse=True)[:10]: |
| 104 | + print(f"{freq:<15.1f} {norm_amp:<15.6f} {actual_amp:<15.6f}") |
| 105 | + |
| 106 | +# Compensation quality assessment |
| 107 | +print(f"\nCompensation Quality Assessment:") |
| 108 | +print(f"50 Hz component: {amplitude_50hz_actual*1000:.3f} mA") |
| 109 | +if amplitude_50hz_normalized < 5e-4: |
| 110 | + print("✓ Good compensation achieved (below threshold)") |
| 111 | +else: |
| 112 | + print("✗ Poor compensation (above threshold)") |
| 113 | + |
| 114 | +print(f"\nFor reference:") |
| 115 | +print(f"- Total RMS fluctuation: {np.std(current_ac)*1000:.2f} mA") |
| 116 | +print(f"- 50 Hz component represents {(amplitude_50hz_actual/np.std(current_ac)*100):.1f}% of total fluctuation") |
| 117 | +plt.show() |
0 commit comments