EV-QA-Framework

EV Battery QA Framework — detect thermal runaway, validate BMS telemetry, comply with UN 38.3 / IEC 62660 / GB 38031, and ship with 967 passing tests and a Docker-ready pipeline.

22 modules. MIT licensed. Python 3.10+.

---

Installation

# Core (anomaly detection, validation, scoring)
pip install ev-qa-framework

# With web dashboard
pip install ev-qa-framework[web]

# With CAN bus support
pip install ev-qa-framework[can]

# With ML and monitoring
pip install ev-qa-framework[ml,monitoring]

# Everything
pip install ev-qa-framework[all]

---

30-second value

git clone https://github.com/remontsuri/EV-QA-Framework.git
cd EV-QA-Framework
docker compose up -d
open http://localhost:8081

Done. You have a running battery QA workstation:

• telemetry validation
• ML anomaly detection
• thermal runaway early warning
• cell imbalance analysis
• SOH prediction
• compliance testing against 6 international standards
• live dashboard with Prometheus metrics

No cloud account required. No external services needed. Just a CSV and a terminal.

---

What you get

Input safety layer. Pydantic schemas for voltage, current, temperature, SOC, SOH. Bad VINs, out-of-range values, and malformed rows are rejected before they reach your models.

Anomaly detection. Isolation Forest on voltage/current/temperature streams. Configurable contamination, severity thresholds, estimator count.

Thermal runaway prediction. Rule-based heuristics (temperature, delta-temp, anomaly score, chemistry runaway point). CRITICAL trigger defaults at >=130 C, rapid-rise trigger at >10 C/min. Catches overheating before cascade onset. Confidence score clamped to [0, 1].

SOH prediction. LSTM-based State of Health forecasting from historical telemetry. Transformer-based prediction via soh_transformer for longer sequences.

Cell imbalance analysis. Statistical analysis of cell group voltages with configurable thresholds, outlier detection, linear regression trend.

Battery scoring. Composite health score (0-100) with letter grades (A+ through F). Combines SOH, internal resistance, cell balance, and thermal history.

CAN bus and DBC. CAN 2.0B and J1939 simulation and reception. DBC parser supports Vector CANdb, SavvyCAN exports, Intel/Motorola byte order.

Fleet analytics. Aggregate analysis across vehicle fleets: degradation curves, anomaly distribution, SOH histograms.

Digital twin. Real-time battery simulation mirroring physical pack behavior. Charge/discharge what-if scenarios and aging projections.

V2G scenarios. Vehicle-to-Grid simulation: bidirectional energy flow, grid demand response, cycling impact on battery health, revenue estimation.

AutoML. Automated model selection and hyperparameter optimization for SOH prediction and anomaly detection (RandomForest + GradientBoosting).

Streaming anomaly detection. Real-time sliding-window detector for live telemetry feeds with configurable retrain frequency.

Uncertainty quantification. Bootstrap confidence intervals on thermal runaway risk scores for safety-critical decision making.

HIL integration. Hardware-in-the-Loop interface for physical BMS hardware and test stands via TCP/Serial.

Compliance testing. UN 38.3, IEC 62660, SAE J2464, ISO 12405, GB/T 31484, GB/T 31486, GB 38031.

Observability. Prometheus /metrics endpoint, Grafana dashboard, HTML coverage reports, JUnit XML.

---

Quick start

# Python CLI (direct)
uv run pytest -v
uv run python run_factory_inspection.py

# Docker Compose (recommended for fresh environments)
docker compose up --build

• Tests + HTML coverage: http://localhost:8081/coverage/
• Prometheus metrics: http://localhost:8081/metrics

---

One-liners

Analyze a CSV:

uv run python -m ev_qa_framework.cli analyze -i examples/tesla_model_s_defective.csv -o report.json

Emulate CAN traffic:

uv run python -m ev_qa_framework.cli emulate --dbc my_battery.dbc --duration 60

Train SOH model:

uv run python -m ev_qa_framework.cli train-soh -i examples/tesla_model_s_defective.csv -m /tmp/soh_model

---

Project structure

ev_qa_framework/ framework.py # core QA engine models.py # Pydantic models + telemetry validation config.py # thresholds and ML config analysis.py # Isolation Forest, EVBatteryAnalyzer soh_predictor.py # LSTM for SOH (TensorFlow optional) soh_transformer.py # Transformer SOH predictor can_bus.py # CAN 2.0B + J1939 simulation dbc_parser.py # .dbc file parser (Vector CANdb + SavvyCAN) cell_balance.py # cell voltage imbalance analysis thermal_runaway.py # thermal runaway prediction (rule + ML) battery_scoring.py # composite battery health scoring physics_features.py # electrochemical/thermal feature extraction fleet_analytics.py # fleet-wide analytics and benchmarking digital_twin.py # real-time battery digital twin v2g_scenarios.py # Vehicle-to-Grid simulation automl.py # automated model selection and HPO hil.py # Hardware-in-the-Loop interface metrics.py # Prometheus metrics cli.py # CLI entry point chemistries.py # battery chemistry definitions (LFP, NMC, NCA) tests/ # 980 tests examples/ # sample telemetry and demos run_factory_inspection.py # end-to-end factory QA demo

---

Status

Artifact	Value
Tests	967 passing
Coverage	93%
CI	Lint + Test + Coverage
License	MIT
Python	3.10+
PyPI	ev-qa-framework 2.5.0

Regression risk is tracked in tests/. Coverage artifacts (coverage/, junit.xml) are present in the release pipeline.

---

Roadmap

• GitHub Actions CI badge + nightly coverage job
• Grafana dashboard import JSON + provisioning
• public PyPI release
• real BMS telemetry adapters (Tesla, BYD, Nio)
• V2S + charging-station scenarios
• integration with Vector CANoe / CANalyzer