showCase: Smart Facility Management
Energy and Infrastructure
Services
- Enterprise (France, Germany, Europe)
- >40 Tsd. Employees
- >7 Bil.€ Revenue
- Exists since year 1900
- Ventilators with motor vibration sensors
- SCADA system used only for monitoring
- >3 years of unprocessed vibrational telemetry data & maintenance events data from incident management
- High maintenance costs
- Intensive maintenance schedule
- Outdated warning and alarming system
- No transparency about system state
- No feedback about maintenance effects
- Data integration & quality issues with telemetry data (inconsistencies, sampling rates)
- Created elastic data lake in AWS
- Preprocessed and cleaned data using plug&play pipelines
- Built intelligent forecasting and classification model of incidents
- Generated partially synthetic data to overcome the rare occurrence of some incidents
- New predictive maintenance algorithm improves incident detection by 70% as compared to SCADA system
- Incidents are detected early (up to 4 days ahead of time), allowing time for maintenance
- Introduced intelligent monitoring based on dynamic alarm thresholds
- Optimized the maintenance schedule
HOW IT WORKS
Big-data native
Use state of the art clustering methods to integrate different data sources and derive business critical insights.
Deep Learning to learn (almost) anything
Utilize our cloud advantage to build and train deep learning networks for predictive maintenance.
Methods & Algorithms
Deep Convolutional Networks, LSTMs, Decision trees, Boosted Trees, Random Forests, Kernel based methods
Forecast and predict
Combine parametric and non-parametric (tree- and kernel based) forecasting algorithm to accurately predict system failures.
Plug&Play Transfer Learning
Utilize your data and save on model training time by reusing pre-trained models on different machinery.
Request presentation
of our predictive maintenance solution.
Please feel free to contact us. We will hold a presentation and show the product at a convenient time for you.
Send us a letter to the
info@adeal-systems.com
Give us a phone call
+49 (0)6102 574 386 0
RaaS
Recommendations as a Service
Our project "AI Services for German SMEs" was funded under the Distr@l program of the state of Hesse: Validierungs-/Wachstumsförderung
Many small and medium-sized enterprises (SMEs) lack the resources and know-how for an automated analysis of their customers' behavior. However, the results of this analysis form the basis for good digital marketing and sales.
Our AI service RaaS solves this issue for SMEs. This pre-built AI service is intended for business users and is thus very different from common technical AI platforms and frameworks. The design of this service is based on an integration of multiple machine learning approaches such as customer segmentation, transaction clustering, statistical consumption analysis and recommendation engines.
The use of our recommendation service provides significant benefits for SMEs such as increasing customer loyalty, increasing marketing team productivity and improving customer service quality. The primary application group is German SMEs from the retail and e-commerce sectors.