Skip to content

Multi Domain Signal Processing

Batuhan edited this page Sep 21, 2026 · 1 revision

Multi-Domain Industrial Signal Processing

Sofia Engine extends far beyond mechanical vibration, providing comprehensive digital signal processing for electrical power grids, acoustic emissions, process telemetry, and multi-axis inertial dynamics.

These algorithms are derived and ported from Rootcastle REI SignalLab.


1. Electrical Power Quality (IEEE 519 / IEC 61000-4-30)

Electrical telemetry from voltage ($v(t)$) and current ($i(t)$) probes is analyzed for power efficiency and power quality disturbances:

1. Power Metrics ($P, Q, S, PF$)

  • Active Power ($P$): $$P = \frac{1}{N} \sum_{n=0}^{N-1} v_n \cdot i_n$$
  • Apparent Power ($S$): $$S = V_{rms} \cdot I_{rms}$$
  • Reactive Power ($Q$): $$Q = \sqrt{S^2 - P^2}$$
  • Power Factor ($PF$): $$PF = \frac{P}{S}, \quad PF \in [-1.0, 1.0]$$

2. Total Harmonic Distortion ($\text{THD}$)

Evaluated up to the 50th harmonic order: $$\text{THD} = \frac{\sqrt{\sum_{h=2}^{50} V_h^2}}{V_1} \times 100%$$

Where individual harmonic magnitudes are verified against IEEE 519 limits ($3.0%$ for odd harmonics, $1.5%$ for even harmonics).

3. Fortescue 3-Phase Symmetrical Components

For 3-phase unbalanced AC power systems, Sofia decomposes line voltages into symmetrical components: $$\begin{bmatrix} V_0 \ V_1 \ V_2 \end{bmatrix} = \frac{1}{3} \begin{bmatrix} 1 & 1 & 1 \ 1 & a & a^2 \ 1 & a^2 & a \end{bmatrix} \begin{bmatrix} V_a \ V_b \ V_c \end{bmatrix}, \quad a = e^{j 120^\circ}$$

  • Voltage Unbalance Factor ($\text{VUF}$): $$\text{VUF} = \frac{|V_2|}{|V_1|} \times 100%$$ A $\text{VUF} > 2.0%$ triggers warnings for three-phase motor de-rating.

4. Power Disturbance Events

  • Voltage Sag: $V_{rms} \in [10%, 90%]$ of nominal.
  • Voltage Swell: $V_{rms} > 110%$ of nominal.
  • Interruption: $V_{rms} < 10%$ of nominal.
  • Inrush Current: Instantaneous peak-to-RMS ratio $> 3.0$.

2. Acoustic Emission & Cavitation Monitoring (ASTM E1316)

High-frequency acoustic emission (AE) and ultrasound telemetry detect stress wave bursts from micro-cracking, partial electrical discharge, and fluid cavitation:

  • AE Energy: $$E_{AE} = \sum_{n=0}^{N-1} v_n^2 \Delta t$$
  • Counts: Number of rising threshold crossings ($> 3\sigma$).
  • Rise Time: Elapsed time from first threshold crossing to peak amplitude.
  • Duration: Total time the burst envelope remains above threshold.
  • Cavitation Index ($C_p$): $$C_p = \frac{\sum_{k: 5\text{kHz} \le f_k \le 20\text{kHz}} |X_k|^2}{\sum_k |X_k|^2}$$ Detects pump and valve cavitation before mechanical pitting occurs.

3. Thermal & Fluid Process Dynamics

  • Temperature Rate of Change: $\frac{dT}{dt}$ (in $^\circ\text{C/s}$ or $^\circ\text{C/min}$) identifies sudden thermal runaways or heat-sink failures.
  • Thermal Inertia & Delta: Dynamic difference between peak and ambient process temperature.
  • Hydraulic Pressure Pulsation: Peak-to-peak pressure swings and pressure crest factor for detecting water hammer and valve chatter.

4. Multi-Axis Inertial Dynamics (IMU)

  • Acceleration Vector Magnitude: $$|\mathbf{a}| = \sqrt{a_x^2 + a_y^2 + a_z^2}$$
  • Dynamic Tilt (Pitch & Roll): $$\text{Pitch} = \text{atan2}(-a_x, \sqrt{a_y^2 + a_z^2}), \quad \text{Roll} = \text{atan2}(a_y, a_z)$$
  • Dynamic Jerk: $$\mathbf{j} = \frac{d\mathbf{a}}{dt}$$ Monitors mechanical shock and high-stress robot arm reversals.

Clone this wiki locally