Building a Ballistic Fire Control Simulator with Python, C# and Redis

#csharp #microservices #python #science

description: How I built a real-time ballistic computer with Euler simulation, Coriolis effect, blast zones and a satellite map interface using microservices architecture.
tags: python, csharp, redis, opensource

Building a Ballistic Fire Control Simulator — BALISTIC V5

A few weeks ago I started wondering: how hard would it be to simulate real external ballistics? Not just a simple parabola — but actual physics with air resistance, wind, air density and even the Coriolis effect.

The result is BALISTIC V5 — a full microservices ballistic fire control simulator.

👉 GitHub: https://github.com/InsaneInfinity/Balistic

🏗️ Architecture

The system is split into two independent services communicating via Redis Streams:

┌─────────────────────┐     Redis Stream      ┌──────────────────────┐
│   Python / Flask    │ ──────────────────►   │    C# Processor      │
│                     │   ballistics:stream   │                      │
│  - Leaflet Map      │                       │  - Euler Simulation  │
│  - Weather API      │ ◄──────────────────   │  - Coriolis Effect   │
│  - Results Panel    │   ballistics:result   │  - Blast Zones       │
│  - PDF Export       │                       │                      │
└─────────────────────┘                       └──────────────────────┘

Why Redis Streams instead of a simple key-value? Because every shot is an independent entry — no data gets overwritten during sequential multi-target fire.

🧮 Physics Model

Trajectory Simulation

For kinetic energy rounds (APFSDS, HEAT) I use an Euler method simulation with dt=0.01s:

double v = Math.Sqrt(vx*vx + vy*vy);
double drag = 0.5 * cd * rho * area * v * v;
vx += -(drag/mass)*(vx/v)*dt;
vy += (-9.81 - (drag/mass)*(vy/v))*dt;

For artillery rounds (HE, EXCALIBUR) I use an empirical angle formula based on the dist/max_range ratio — artillery always prefers a high-arc trajectory (45°–65°).

Coriolis Effect

This was the most interesting part. The real deflection formula:

d_cor = Ω · sin(φ) · v_avg · t² / 2

Where:

Ω = 7.2921×10⁻⁵ rad/s — Earth's angular velocity
φ — shooter's latitude (dynamic, updates with every position change)
v_avg = dist / tof — average horizontal velocity during flight

Results for M107 HE at ~5km range, 54°N latitude: ~3m deflection to the right — matches NATO FM 6-40 tables.

Air Density

Calculated live from OpenWeatherMap data:

dens = (pressure * 100) / (287.058 * (temp + 273.15))

💥 Blast Zones

Based on NATO / FM 6-40 publicly available data:

Round	Total destruction	Heavy damage	Light damage	Hazard zone
M107 HE / EXCALIBUR	30 m	100 m	300 m	800 m
HE STD (RAK 120mm)	8 m	30 m	80 m	200 m
APFSDS	kinetic only	—	—	—
HEAT	—	—	—	5 m

These render as 4 concentric circles on the satellite map with tooltips.

🔫 Supported Systems

System	Round	v₀	Trajectory
AHS KRAB (155mm)	M107 HE	560 m/s	Artillery arc
AHS KRAB (155mm)	EXCALIBUR	560 m/s	Artillery arc
M120 RAK (120mm)	HE STD	280 m/s	Artillery arc
M120 RAK (120mm)	SMOKE	250 m/s	Artillery arc
LEOPARD 2 (120mm)	APFSDS	1650 m/s	Flat
LEOPARD 2 (120mm)	HEAT	1140 m/s	Flat

✨ Features

🗺️ Satellite map with shooter/target positioning
🎯 Multi-target selection — mark multiple targets, fire sequentially
📊 Live results panel — azimuth, elevation angle, TOF, wind drift + Coriolis
💥 4-zone blast radius visualization
📄 PDF session report export
🔄 Dynamic system/ammo switching without restart
🔐 Session token authorization on all fire endpoints

🚀 Key Lessons Learned

1. Euler simulation has limits
For high-velocity artillery (v0=827 m/s — my initial wrong value!), the vacuum range formula gives unrealistic results. Real M107 HE has v0=560 m/s, not 827 m/s.

2. Bisection method for elevation angle
Finding the correct elevation angle for a given range required a bisection search. The tricky part: artillery prefers the high-arc solution (>45°), not the flat one.

3. Coriolis is smaller than you think
At 5km range with 20s flight time at 54°N: only ~3m deflection. Significant for precision rounds like EXCALIBUR (CEP 10m), negligible for M107 HE (CEP 150m).

4. Redis Streams > Redis String for queues
Using XADD/XREAD with consumer groups means no data loss even if the C# processor is temporarily down — messages wait in the stream.

⚠️ Disclaimer

Educational and simulation purposes only. Real fire control systems use 6-DOF models, balloon sonde meteorological data, barrel wear corrections and dozens of other factors.