Challenge

Companies with complex products need to identify early which components are likely to fail in the future, what costs this will entail, and where action should be taken first. Conventional point forecasts often fail to account for uncertainties. The quality of results and their interpretation depend heavily on the approach and experience of individual experts. This makes it difficult to prioritize field issues, make reliable recall decisions, and plan for spare parts based on actual needs.

The IAV Solution

IAV Hirundo is a cloud-based SaaS solution for probabilistic lifetime prediction. The patented method combines statistics with modern machine learning approaches and uses fleet and failure data to predict future failures, including a confidence interval. The workflow is fully automated – from data cleansing and model selection to result visualization – and delivers consistent, decision-ready results in a short amount of time – without requiring Weibull expertise on user side.

Key Benefits

  • Probabilistic Forecasting: Reliable predictions instead of point estimates
  • Patented Weibull mixed model with intelligent model selection: Always the optimal model without the need for statistical expertise
  • Consistent, standardized workflow: From data import to result visualization
  • SaaS solution on AWS: Ready to use immediately without own infrastructure
  • Minimal data requirements: Only timestamps, age index, ID of fleet and failures are needed
  • Data generator: Generate synthetic datasets for reference and testing purposes

Additional Features

  • Upload fleet and breakdown data (ID, timestamp, usage index (e.g., mileage))
  • Create a dataset and gain initial insights
    • Distribution of the datasets
    • Cumulative failure rate and Kaplan-Meier estimator
    • Historical failure trends and fleet usage information
  • Create a forecast and look to the future
    • Future failure trends with confidence interval
    • Visualization of fleet usage and failure behavior
    • Determination of relative and absolute part failures

  • Cloud-based SaaS solution, implemented on AWS (also available via the AWS Marketplace)
  • Fully automated, scalable processing of fleet and failure data of any size
  • Applicable across various product groups and industries with complex, mechatronic products
  • Relevant IAV standards: TISAX, ISO 27001, and ISO 9001

  • Automated data cleansing ensures maximum data quality in forecasting
  • Process-oriented implementation with a strong focus on data protection within the IAV system environment

  • Patent DE 10 2021 115 804: “Method for Estimating the Service Life of a Vehicle Fleet”
  • Validated and recommended by Fraunhofer ITWM
  • Used by BMW, a leading German OEM
  • Scientific basis and publications:
    • ATZ worldwide, April 2024: “Probabilistic Models for Prognostics of the Condition of Vehicle Fleets”
    • PHM Europe 2021: “Probabilistic Lifetime Prediction for Fleet Failure Prognostics”

  • Self-service usage (pay-per-use) as an IAV-hosted SaaS solution on AWS
  • Adaptable to different fleets, product groups, age indices and data quality levels
  • Optional data generator for synthetic test data
  • Demo accounts available (upon request)

Interested?

Turn uncertainty into reliable forecasting: Prioritize future field issues, assess risks, and plan spare parts requirements based on probabilistic forecasts.

Feel free to contact us using the form below to schedule a demo or get started with your fleet and failure data.

Contact

Your Inquiry

Danny Vanis

Product Manager
LinkedIn