PopChainNetwork/fiber-laser-cnc-logs

Fiber Laser CNC Machine Logs — Real Industrial Data

The first industrial machine dataset with on-chain cryptographic provenance.

This dataset contains real operational logs from a Bodor BodorThinker 3.0 fiber laser CNC cutting machine, collected over 5 months of continuous industrial production (November 2021 – April 2022).

Unlike synthetic datasets commonly used in predictive maintenance research, this data comes from a real machine in a real factory, with every record verified on the PopChain blockchain using SHA-256 hashing and WOTS+ post-quantum signatures.

Dataset Summary

Metric	Value
Machine	Bodor BodorThinker 3.0 Fiber Laser
Period	Nov 2021 – Apr 2022 (5 months)
Original size	52 MB (SQLite)
Total rows	426,374
Errors (labeled)	12,781
Warnings (labeled)	10,332
Machine states	33 distinct
POP Score	711/1000 (Grade B)
Blockchain proof	PopChain Block #47243
SHA-256	c1fe88b22e5ca19e03172cba318f1fc208c09883095b9805a2bc2f1bec3e450c

Why This Dataset Matters

60% of predictive maintenance ML projects fail due to poor training data. Most available datasets are either:

• Synthetic (AI4I 2020, NASA C-MAPSS) — not representative of real failures
• Proprietary — locked inside manufacturers like Trumpf, ESAB, Bystronic
• Unlabeled — raw sensor dumps without error annotations

This dataset solves all three problems:

1. Real — from a production fiber laser, not a simulation
2. Labeled — 12,781 errors and 10,332 warnings with natural language descriptions
3. Verified — SHA-256 hash and WOTS+ signature on PopChain blockchain prove data hasn't been tampered with

Data Fields

Events (errors & warnings)

• type: "error" or "warning"
• text: Natural language description of the event (mix of English and Chinese)
• time: ISO timestamp of the event

Features

• feature: Feature name (e.g., servo parameters, coolant status, axis positions)
• distribution: Statistical distribution of the feature

Machine States

• 33 distinct states including: initialization, cutting, idle, alarm, emergency stop, coolant failure, servo alarm (X/Y/Z/W axes), safety door, etc.

Error Categories

Category	Count	Description
Servo alarms	4,200+	X, Y, Z, W axis servo faults
Coolant failures	1,800+	Cooling system malfunctions
Safety door	900+	Safety interlock events
Communication	600+	Controller communication errors
Power	400+	Power supply anomalies
Other	4,800+	Various operational faults

Formats Available

Format	File	Use Case
JSONL	c1fe88b22e5ca19e.jsonl	ML pipelines, streaming
CSV	c1fe88b22e5ca19e.csv	Pandas, Excel, SQL
HuggingFace JSONL	c1fe88b22e5ca19e_hf.jsonl	LLM fine-tuning (instruction format)

Usage

from datasets import load_dataset

dataset = load_dataset("popchain/fiber-laser-cnc-logs")

Or directly:

import pandas as pd

df = pd.read_csv("c1fe88b22e5ca19e.csv")
print(f"Events: {len(df)}")
print(df.head())

For LLM fine-tuning:

from datasets import load_dataset

ds = load_dataset("json", data_files="c1fe88b22e5ca19e_hf.jsonl")
# Each row has: instruction, input, output
# Ready for Llama, Mistral, GPT fine-tuning

On-Chain Provenance

This dataset has been verified on the PopChain blockchain — a Layer 1 blockchain built from scratch in Rust that uses Proof-of-Process consensus.

• Chain: PopChain Alpha Testnet
• Block: #47243
• SHA-256: c1fe88b22e5ca19e03172cba318f1fc208c09883095b9805a2bc2f1bec3e450c
• Signature: WOTS+ (post-quantum, RFC 8391)
• Tamper-proof: ✅ Any modification to the data would invalidate the on-chain proof

Verify on the explorer: popchain.tech/explorer

Full Dataset & More

This is a preview. The full dataset (52MB SQLite with 426K rows) and additional machine datasets (Lincoln, Senfeng lasers) are available on the PopChain Data Marketplace:

🔗 bomwt.io/market — Buy with BIN on PopChain

More about PopChain:

• 🌐 popchain.tech — Chain portal
• 💼 bomwt.io — DEX & Wallet
• 📊 PopChain Explorer — Live block explorer
• 📄 Whitepaper — Technical documentation
• 💬 Discord — Community

Citation

@dataset{popchain_fiber_laser_2026,
  title={Fiber Laser CNC Machine Logs — Real Industrial Data with Blockchain Provenance},
  author={Piekut, Paweł},
  organization={PPlastechniek},
  year={2026},
  url={https://popchain.tech},
  note={Verified on PopChain blockchain, block #47243}
}

License

Proof-of-Process Data License v1 — see popchain.tech for details.

Data produced by PPlastechniek, Almelo, Netherlands.