BESS HIL Simulator

Overview

This repository contains a Hardware-in-the-Loop (HIL) Simulator designed for testing and validating Energy Management Systems (EMS) software. The simulator acts as a digital twin of a Power Conversion System (PCS) and Battery Energy Storage System (BESS), modeling the physical dynamics, grid coupling, and communication latencies.

It allows an external EMS to read telemetry and write setpoints via Modbus TCP, while the simulator runs a real-time physics engine internally.

Simulation Results

The following plots demonstrate the simulator's behavior using a Sungrow ST5015UX PCS-ESU model with the configuration:

• Sampling Time: 100ms
• Measurement Latency: 500ms
• P Time Constant: 0.1s
• Q Time Constant: 0.2s
• Max Apparent Power: 0.21 MVA

Power Response Characteristics

Active Power (P) tracking setpoint with first-order lag dynamics

Reactive Power (Q) response showing similar lag characteristics

Grid Parameters & Derived Measurements

Grid voltage profile during simulation

System frequency measurements

Resulting power factor calculated from P and Q measurements

Calculated current injected to/absorpted by the grid

Features

• Real-Time Physics Engine: Models Active ($P$) and Reactive ($Q$) power evolution using First-Order Lag dynamics (Low Pass Filter).
• Transport Delay Simulation: Simulates realistic SCADA/metering delays ($N$ steps) to test EMS feedback loops.
• Modbus TCP Server: Acts as a slave device (Port 502), exposing measurements on Input Registers and accepting commands via Holding Registers.
• P-Q Capability Curve: Implements complex saturation logic where reactive power limits ($Q_{max}/Q_{min}$) float dynamically based on grid voltage and active power levels.
• Grid Scenarios: Includes a scenario generator that injects voltage sags ($0.95$ pu) and swells ($1.05$ pu) to trigger capability curve changes.
• Web-Based Data Viewer: A standalone viewer.html tool to visualize BessData.csv logs with interactive charts and zoom capabilities.

System Architecture

The interaction between the Device Under Test (EMS) and the Simulator is cyclic:

1. EMS sends control references ($u$) via Modbus.
2. Simulator evolves internal physics state ($x$) based on time constants.
3. Simulator returns delayed feedback measurements ($y$) to the EMS.

Mathematical Model

The system uses a discrete-time state-space representation.

1. State Vector ($x$)

The state vector represents the physical state of the plant at the inverter terminals:

$$ x(k) = \begin{bmatrix} P_{phys}(k) \ Q_{phys}(k) \ V_{grid}(k) \ f_{grid}(k) \end{bmatrix}^T $$

2. Input Vector ($u$)

The input vector consists of the setpoints received from the EMS:

$$ u(k) = \begin{bmatrix} P_{setpoint}(k) \ Q_{setpoint}(k) \end{bmatrix}^T $$

3. State Evolution Equation

The discrete-time dynamics are modeled as a 1st-order lag for Power, while Voltage and Frequency are scenario-driven inputs:

$$ x(k+1) = A_d x(k) + B_d u(k) + E_{scenario} w_{scenario}(k) $$

Where the system matrices are defined as:

$$ A_d = \begin{bmatrix} e^{-T_s/\tau_P} & 0 & 0 & 0 \\ 0 & e^{-T_s/\tau_Q} & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \end{bmatrix} $$

$$ B_d = \begin{bmatrix} 1 - e^{-T_s/\tau_P} & 0 \\ 0 & 1 - e^{-T_s/\tau_Q} \\ 0 & 0 \\ 0 & 0 \end{bmatrix} $$

$$ E_{scenario} = \begin{bmatrix} 0 & 0 \\ 0 & 0 \\ 1 & 0 \\ 0 & 1 \end{bmatrix} \quad w_{scenario}(k) = \begin{bmatrix} V_{grid}^{_(k)} \\ f_{grid}^{_(k)} \end{bmatrix} $$

4. Output Vector ($y$)

To simulate realistic SCADA feedback, a transport delay $N = T_{delay} / T_s$ is applied to the physical states. The output seen by the EMS is:

$$y(k) = \begin{bmatrix} P_{meas}(k) \\ Q_{meas}(k) \\ PF_{meas}(k) \\ V_{meas}(k) \\ f_{meas}(k) \\ I_{meas}(k) \end{bmatrix} = \mathcal{H}(x(k-N))$$

Where current $I_{meas}$ is derived non-linearly as $I_{meas} = \frac{\sqrt{P_{meas}^2 + Q_{meas}^2}}{V_{meas}}$.

Project Structure

• Program.cs: The main entry point. Initializes the model, starts the Modbus server, and runs the real-time simulation loop.
• BessPhysicsModel.cs: Encapsulates the physics engine, IIR filters, and delay buffers (Queues).
• ModbusServerWrapper.cs: Wraps the NModbus logic to handle register mapping and TCP connections.
• PqCapabilityCurve.cs: Handles the complex saturation logic and interpolation between voltage levels (0.9pu to 1.1pu).
• viewer.html: A frontend tool for analyzing simulation CSV logs.
• BessData.csv: The output log file generated during simulation.

Getting Started

Prerequisites

• .NET 6.0 SDK or later.
• A Modbus TCP Client (e.g., QModMaster) or an actual EMS to act as the Master.
• Modern Web Browser (for viewer.html).

Installation

1. Clone the repository.
2. Open the solution in Visual Studio or VS Code.
3. Restore NuGet packages (specifically NModbus).

Running the Simulator

Run the application with administrative privileges (required to bind to Port 502).

dotnet run

Note: If Port 502 is blocked or in use, modify ModbusServerWrapper.Start(502) in Program.cs to use port 5020.

Usage

1. Console Control

You can manually inject setpoints directly via the console window to test step responses without an EMS.

Format: P[MW] Q[MVAR]

Enter command (P Q) or 'exit': 1.0 0.5
>>> Command queued: P=1.00 MW, Q=0.50 MVAR

2. Modbus Interface (EMS Connection)

The simulator listens on Port 502 (Unit ID 1). Data is stored as 32-bit Floating Point values (spanning 2 registers each).

Signal	Type	Register Type	Address (Offset)	Description
Setpoint P	Float (32-bit)	Holding (RW)	40001 (0)	Active Power Command
Setpoint Q	Float (32-bit)	Holding (RW)	40003 (2)	Reactive Power Command
Meas P	Float (32-bit)	Input (RO)	30001 (0)	Active Power Feedback
Meas Q	Float (32-bit)	Input (RO)	30003 (2)	Reactive Power Feedback
Meas V	Float (32-bit)	Input (RO)	30005 (4)	Grid Voltage
Meas F	Float (32-bit)	Input (RO)	30007 (6)	Grid Frequency
Meas I	Float (32-bit)	Input (RO)	30009 (8)	Current (Derived)

3. Data Visualization

1. Let the simulation run; it logs data to BessData.csv in real-time.
2. Open viewer.html in your browser.
3. Click Load CSV File and select the generated BessData.csv.
4. Use the tabs to view Charts, Summary Stats, and Data Tables.

Configuration

Key simulation parameters are defined in Program.cs and derived from the system design:

Parameter	Value	Description
Sampling Time ($T_s$)	0.1 s	Simulation Step Size
Total Delay ($T_{delay}$)	0.5 s	Physics to Feedback latency
P Time Constant ($\tau_P$)	0.1 s	Active Power Response Lag
Q Time Constant ($\tau_Q$)	0.2 s	Reactive Power Response Lag
Max Apparent Power ($S_{max}$)	0.21 MVA	Inverter Capacity Limit