Kauzen.Diagnose

The AI module for visual fault diagnosis in complex machine environments. Fewer failures, shorter downtimes, more safety – with a system that truly understands your machine structure.

3 Phases to Implementation

From the OPC UA connection to final acceptance – three phases, only 3 days on site.

From the first analysis to active monitoring – three consecutive phases

01

Initial analysis

< 1 month*

Using the Digital Twin Creator, the existing production bus is analyzed automatically and a digital twin of the plant is created. On this basis, compatibility with the Kauzen system is checked and any required adjustments are identified.

Departments involved: IT, production, maintenance
02

Installation

< 1 week*

The Kauzen system is deployed on the plant infrastructure – via installation of suitable hardware on site or as a rollout within your company's own cloud – then connected to the production plant and configured for operation.

Departments involved: IT, production
03

Real-time monitoring

< 1 day*

The system is ready and live operation begins: the current data of the plant is automatically read, processed and analyzed – anomalies and relevant plant states are made continuously visible.

Departments involved: production, maintenance

* Time indications are guidelines and depend on the plant-specific environment.

Pre-requisite

OPC UA – 1-day interface connection

Phase 1 · < 1 month* Step 1 1 day

Digital twin

On site: environment exploration & twin generation

Phase 1 · < 1 month* Step 2 optional, 2 weeks

Configuration

Remote: software & NN architecture design – only if customization is needed, otherwise straight to installation

Phase 2 · < 1 week* Step 3 1 day

Installation

On site: setup & OPC UA connection

Phase 2 · < 1 week* Step 4 overnight

Autonomous training

Automatic training of the neural networks

Phase 3 · < 1 day* Step 5 1 day

Acceptance & training

Final acceptance & team training

OPC UA – natively integrated

Kauzen speaks the universal language of Industry 4.0 – OPC Unified Architecture. This ensures seamless connectivity with equipment from any manufacturer.

  • Manufacturer-independent: Connect to Siemens, Allen-Bradley, Mitsubishi and hundreds more
  • Secure by design: Built-in encryption and authentication
  • Platform-independent: Works across Windows, Linux and industrial OS
  • No code required: Plug-and-play configuration

Security & Data Sovereignty

Your data stays yours – always.

  • On-Site Installation: All Kauzen software runs on your premises. No cloud dependencies, no external servers.
  • Data Never Leaves the Factory: Training data, inference results and all analytics remain within your facility boundaries.
  • Enterprise Security: Role-based access control, audit logging and integration with existing security infrastructure.

How Kauzen.Diagnose works

First the production environment assessment – then four processing layers, from raw OPC UA data to intelligent fault diagnosis.

Production Environment Capture & Assessment

Before data processing begins, the Kauzen team captures your production environment and evaluates the connectivity. In this step the digital twin of your facility is generated – to make sure the software and hardware work flawlessly on your site before the first data stream flows.

At the start of implementationDigital Twin is generatedOn site, verified in the customer environment
↓

Layer 1: OPC UA Data Ingestion

Connects to your existing production infrastructure via the standard OPC UA protocol. No hardware changes required – manufacturer-independent connectivity.

100,000+ signals supportedReal-time data streaming
↓

Layer 2: Signal Processing & Normalization

Cleans, normalizes and transforms raw sensor data into feature vectors. Handles missing values, outliers and temporal alignment across heterogeneous sources.

Edge computing – all processing on-siteNo data leaves the factory
↓

Layer 3: Auto-Trained Neural Networks

Hundreds of neural networks automatically train on your production data during the passive learning phase – under optimal conditions, overnight. They are then available for monitoring your facility.

200+ automatically trained NNsZero manual tuning
↓

Layer 4: Real-Time Fault Diagnosis

Continuous monitoring with instant anomaly detection and root cause analysis — repair with a target, not with guesses.

<100 ms inference latencyInstant diagnosis

What's Included

Hardware & Software

  • Kauzen edge server (on-premises)
  • Kauzen AI platform license
  • OPC UA connectivity modules
  • Digital Twin generation engine

Neural Networks

  • 200+ automatically trained networks
  • Ready to use for monitoring your production
  • Learn directly from your production data
  • No manual tuning required

Services

  • 3 days on-site engineering support
  • Digital Twin generation
  • System configuration & integration
  • Employee training (half-day)

Documentation

  • Digital Twin documentation
  • User manual & quick-start guide
  • API documentation (if applicable)
  • Maintenance guidelines