In response to the customer’s needs, an IT system has been developed and implemented, which learns on an ongoing basis based on historically recorded failures and calculates the probability of failure of individual sections of heat pipelines.

The basis for statistical inference are the parameters of network sections such as: their diameter, length, age, material from which they were made, technology, location in the field and history of previous failures.

During the heating season, elements of the heating pipelines fail. For this reason, selected heat pipelines are preventively replaced in the summer season.

However, the problem is to define which sections of the heating network have the greatest risk of failure in the coming heating season. Such a task was assigned by specialists from our team to be developed on the basis of available historical data.

In response to the customer’s needs, an IT system has been developed and implemented, which learns on an ongoing basis based on historically recorded failures and calculates the probability of failure of individual sections of heat pipelines.

The basis for statistical inference are the parameters of network sections such as: their diameter, length, age, material from which they were made, technology, location in the field and history of previous failures.

After many weeks of testing, the system prepared by our team precisely indicates the sections with the highest probability of failure.

About 45% of the sections indicated by the system in the summer season as vulnerable to failures actually fail in the heating season, which means that the implementation of our solution has halved the winter failures of the heating network.