Maintenance Planning

On the topic "Maintenance of a continuously operated production plant of basic industry," we analysed ten years of maintenance data for a client. The following conclusion was obtained:

An ad-hoc maintainance measure, made necessary by a standstill of the plant, costs approx. 40% more than a planable measure. This refers only to maintenance costs and not the financial loss due to a production stop; this depends upon the value of the plant and materials.

Making maintenance planable is therefore the vision of the entire producing industry. The possibilities available through organisation, management and experience of workers are already exhausted in most plants.

A completely new methode promisses relief: Sensors are installed on all critical equipments and send their measurements to a central computer; Plant Information Management System (PIMS). A novel program analyses the data with the goal of making predictions as to when which Equipment will fail. The prediction is accurate to 90%; this corresponds to an accuracy of one production shift. With this knowledge maintenance can be planed and the extra costs saved.

In a study, algorithmica technologies was able to deduce a potential of 4% per year (relative to the total maintenance costs). At a yearly volume of 15 Mio. Euro, this would come to over 4 Mio. Euro in five years. On this scale, the investment into a system of this kind returns in less than one year.

Neuronator prediction for real data from a chemical reactor

Using real PIMS data of a chemical reactor, algorithmica technologies was able demonstrate the practicability of this approach. The technologies Neuronator and Datanator were used for this.


Deutsch
English
Français