A German startup develops a unique solution for prediction of machine failures. It detects and evaluates anomalies without predetermined limit values, taking a big step towards fully automated surveillance. To test application in the maritime sector, the company is looking for contacts to shipping companies. In a technical cooperation agreement, the partner should test the pilot system under real conditions onboard vessels and contribute with his feedback to further development in a partnership.
Machine or engine failure causes significant damages for shipping companies: Loss of time and loss of a lot of money. Using advanced technologies for predictive maintenance, deteriorations in machinery and equipment can be detected at the onset and preventive measures may be taken before ships enter open water, where repairs are more difficult and expensive. A Northern German startup develops a unique algorithm on the basis of probabilistic models for predictive maintenance of nearly any kind of machine or engine with following advantages compared to other solutions: - Current technologies on the market can detect anomalies in the condition data of machinery, but the worker or engineer still must decide what this anomaly means. The technology offered does not only detect, but evaluates the anomalies and thus offers active "decision support". - Current methods (e.g., autoencoder) code failures as a fixed factor. These methods are extremely error-prone and inflexible. The solution developed learns the probability distribution of the data and takes a sample in latent space. Through the learner distribution, the software can evaluate how confident a prediction is. It makes the setting of limit values easier, is more robust against disturbances or coincidences and small changes do not lead to false intepretations. - Creating a digital twin the solution learns the system behavior of the machines and can even narrow down the origin of the anomaly over time. By learning the system behavior, troubleshooting can be narrowed down to certain areas. This recommendation becomes more and more precise over the course of time and makes it possible to actually provide a real forecast via extrapolations. The solution is a software, but the startup offers combined support with hardware if necessary to ensure a full-service solution for customers. Once reaching market readiness the predictive maintenance solution can be booked in a subscription. The startup is currently further shaping the solution in pilot projects with partners from different industries, showing very promising results. Application on board of ships could bring great benefits for maritime industry, but is challenging with regard to collection and transmission of data. To further develop feasible solutions for maritime applications, the company is looking for contacts to shipping companies to cooperate in a technical cooperation agreement for test phase aboard vessels. The partner should test the pilot system under real conditions and contribute with his feedback to further development in a partnership. The ideal partner should already monitor performance of the vessel and be able to gather a significant amount of data that can be further analyzed by the algorithm.
Type (e.g. company, R&D institution…), field of industry and Role of Partner Sought:
Type: Shipping Company Role: Pilot partner for application in maritime industries. The partner should have access to relevant data onboard of vessels, test the pilot system under real conditions and contribute with his feedback to further development in a partnership.
Stage of Development:
Available for demonstration
Comments Regarding Stage of Development:
Product is currently being tested with application partners from different industries in pilot projects.