Pipelines - NVH-One

[ Foundry Overview ] Data Aquisition: Bring in all your data The beginning of any workflow is data – wether live, online datastreams or pre-recorded datasets. NVH-One Foundry can utilize all common ways of data input:

Sensor read-out

Utilize IIoT protocols to directly acquire data in NVH-One with no intermediate steps: robust, reliable, easily set-up.

Work with any data format

ASAM ODS ATFX, PAK data and MDF4/MF4 is natively supported, CSV, SDF, TXT, UFF, WAV and many(most) more formats are easily made accessible via the sidecar-principle of adding atfx-headers.

Use external software

ASAM ODS ATFX and MDF4/MF4 as open and standardized data-formats are utilized by various software like B&K Connect/Pulse, Dewesoft, Head Acoustics Artemis, ETAS Inca, Vector CANape and Siemens Test.lab.

A2B-Bus

Coming soon: A2B-Bus datastreams can directly be accessed and measured by NVH-One to create an integrated workflow for your data evaluations.

[ Foundry Overview ] Data Quality & Integrity: Ensure your data quality Ensure data reliability through automated validation, structural consistency checks, and full traceability across the entire data lifecycle.

Signal validation

Automated detection of missing channels, clipping, drift and sampling inconsistencies grant the integrity of your data

Metadata consistency control

Schema support, naming rules and structural validation help to flexibly access and sustainably utilize your data

Data lineage

Consistent datasets and reproducible processing empower full tracability and control over your data.

[ Foundry Overview ] Classify: Give meaning to your data Structure and contextualize data by applying rule-based tagging, semantic models, and similarity-driven grouping.

Rule-based tagging

Automated tagging can be based on defined logic and/or metadata rules.

Pattern grouping

Clustering and similarity detection across the measurements, e.g. variants or operating states help to easily average or compare data.

Semantic data model

Flexible but standardizeable hierarchies and taxonomies can be applied across projects and domains to easily meet the requirements of different stakeholders

[ Foundry Overview ] Enhance: Features & Transformation Transform raw signals into meaningful features through signal conditioning, scalable processing pipelines, and domain-specific extraction.

Signal conditioning

Filtering, resampling, domain transformation, windowing, normalization can be applied to condition the data for the respective usecases

Domain-specific feature extraction

Orders, spectra, psychoacoustic metrics, statistical KPIs and many others can be extracted to be fed to any following algorithm or AI

Scalable processing pipelines

Structured processing automations that work via the ETL principle to flexibly entract the data of interest from any source, making its relevant features accessible and pushing these to the appropriate destinations

[ Foundry Overview ] Analyze: Automatic evaluation Evaluate data automatically using custom logic, statistical models, and cross-dataset benchmarking.

Rule-based evaluation

Limit curves, Threshold checks, pass/fail logic, standards compliance checks can be created and are evaluating your data accordingly

Model-based analytics

Breakout points to statistical models, anomaly detection, predictive patterns or machine learning approaches can be utilized to integrate model-based analytics

Cross-dataset benchmarking

Fleet, variant or time-series comparisons are applied and evaluated to create meaningful and consistent benchmarks

[ Foundry Overview ] Human Review: Experts in the loop Integrate expert knowledge through interactive analysis, structured validation workflows, and feedback-driven refinement.

Interactive signal exploration

Synchronized time/frequency views, Playback and MarkUp functionalities can be supplied

Structured validation workflows

Ready to connect your reports toexisting or custom approval and comment logic systems

Feedback integration

Incorporation of expert decisions whenever necessary

[ Foundry Overview ] Report & Results: Use insights Deliver actionable insights via automated reporting, KPI visualization, and seamless integration into enterprise systems.

Standardized reporting

Automated report generation with templates for instant and consistent reporting and report distribution

KPI dashboards

Feed and incorporate Business Intelligence workflows for management views and rapid decision making

System integration

Interface with BI, ERP, PLM or other enterprise platforms to empower interdisciplinary insights

From recorded test data to standard reports Turn turbine data into results faster

[ ESSENTIAL USE CASE ]

Turbine tests generate large volumes of synchronously recorded dynamic and condition data. During the measurement, the test personnel focus on safely monitoring the test run. The subsequent standard analysis and report generation, however, should be performed as quickly and consistently as possible, without additional manual effort.

The Essential Pipeline automates this step: Relevant measurement data is read, defined parameters are classified, standardized analyses are performed, and results reports are made directly available to the responsible analysis or development departments.

Key message

Mass data from turbine tests is automatically converted into standardized, distributable results – reliably, reproducibly, and without manual report generation.

Human-in-the-Loop data validation Validate, evaluate, correct NVH data

[ ADVANCED USE CASE ]

In NVH test campaigns, data quality is essential for reliable technical conclusions. Typical errors such as swapped channels, incorrect signs, inconsistent direction definitions, dropouts, or signal disturbances must be detected early and evaluated in a traceable manner.

The Advanced Pipeline consolidates measurement data from NVH systems, bus data, and status data, automatically checks it for defined error patterns, and visually presents anomalies for expert evaluation. After approval by experts, affected channels are marked and – where applicable – automatically corrected.

Key message

NVH test data is automatically checked, visually evaluated, and corrected following expert approval – for higher data quality, shorter turnaround times, and traceable results.

Database-driven e-drive analysis Automate e-drive analyses and scale them across vehicle programs

[ SCALABLE USE CASE ]

When analyzing electric drives – especially in third-party or benchmark vehicles – reliable engine speed information is not always directly available via the bus. For precise order analysis, vehicle configuration, drive concept, wheel speeds, and gear ratios must therefore be consistently integrated.

The Scalable Pipeline links raw test bench data with a vehicle database, identifies the vehicle and drive context, calculates the required speed information from this data, and automatically performs defined order analyses. Results are stored in a structured format, and selected results are fed back into the vehicle database.

Key message

Raw test bench data, the vehicle database, and analysis specifications are combined into a scalable evaluation chain – for precise e-drive analyses without manual intermediate steps.