Truvolt.ai Energy Management System (EMS)

ACS Proposes: 150 MW data center utilizing a Truvolt.ai Energy Management System (EMS), the PID (Proportional-Integral-Derivative) controller ingestion form is the primary interface where engineers define how the BESS responds to volatile power demands.

PID loop typically manages Frequency Regulation or Peak Shaving In a data center environment. Because AI workloads can cause 30–50 MW swings in seconds, the ingestion form must be granular enough to prevent "hunting" (oscillation) while remaining fast enough to protect the grid.

1. Control Loop Configuration ("What")

This section defines the objective of the specific PID loop you are tuning.

• Control Objective: (e.g., Frequency Response, Voltage Stabilization, or Demand Charge Management).

• Set-Point ($SP$): The target value (e.g., $60.00$ Hz for frequency or $140$ MW for peak shaving).

• Process Variable ($PV$): The real-time sensor feed (e.g., "Main Incomer Meter 01 Power").

• Control Variable ($CV$): The output command (e.g., "BESS Discharge/Charge Power in kW").

---

2. The PID Tuning Parameters ("How")

These are the core coefficients that dictate the battery's behavior.

Parameter	Field Name	Description
$K_p$	Proportional Gain	Determines the immediate response to current error. High $K_p$ makes the BESS aggressive but can cause 150 MW overshoots.
$K_i$	Integral Gain	Eliminates steady-state error by looking at past errors. Essential for ensuring the data center doesn't drift away from its power target over minutes.
$K_d$	Derivative Gain	Predicts future error by looking at the rate of change. Vital for AI data centers to "catch" a rapid ramp-up in server load before frequency drops.
$T_s$	Sample Time	How often the loop runs. For a 150 MW site, this is typically 20ms to 100ms.

---

3. Constraints and Safety (The "Guardrails")

Since you are managing a massive 150 MW system, the ingestion form will include "Clamp" or "Saturation" limits to prevent damaging the battery cells.

• Output Limits (Saturation): Max Discharge (e.g., $150,000$ kW) and Max Charge (e.g., $-100,000$ kW).

• Slew Rate Limit: The maximum rate of change for power output (e.g., $50$ MW per second) to prevent thermal stress.

• Deadband: A "quiet zone" around the set-point (e.g., $\pm 0.05$ MW) where the PID does nothing. This prevents the battery from cycling unnecessarily on tiny noise.

• Anti-Windup: A toggle or value to stop the Integral ($K_i$) from building up when the battery is already at 100% or 0% State of Charge (SoC).

---

4. Logic & Mode Selection

• Operation Mode: Auto / Manual / Cascade.

• Feed-Forward Input: In advanced EMS like Truvolt.ai, you might ingest a "Workload Forecast" signal from the data center's scheduling software. This allows the PID to react before the servers even draw the power.

Typical Form Layout Example:

PID Loop ID: BESS_FREQ_REG_01

Status: Active

Input Source: Grid_Meter_A_Freq

Target: 60.00 Hz

$P$: 2.50 | $I$: 0.05 | $D$: 0.10

Deadband: 0.015 Hz

Max Ramp: 25 MW/sec

Are you looking to tune this for a specific use case, such as "Peak Shaving" to avoid utility penalties, or is this for "Fast Frequency Response" (FFR) to sell services back to the grid?