On Segezha PPM the system of predictive analytics is implemented

2019: System implementation of predictive analytics

On October 14, 2019 Jet Infosystems reported that Segezha Group implements the system predicting emergence of defects and a stop of operation of the paper-making equipment. Digital technologies of predictive analytics using methods of machine learning (ML) for the Segezha PPM.

One of the main calls for pulp-and-paper production is the break of the cloth moving on paper machines. Break of a cloth – a big problem of both old, and new paper-mills. Even the minimum stop of a production line once a day on the scale of one year means long equipment downtime, an additional consumption of raw materials and considerable financial losses.

For the solution of a task the Center of machine learning Jet Infosystems constructed ML models which analyze data from sensors of the equipment and the indicators of an APCS concerning service of the machine and material substitution. A system determines different levels of alarm and issues an indicator of probability of break of a cloth or a stop of the machine, predicting date, time and the possible reason of damage. Analysis results are brought to monitors of the operator of the machine and the duty technologist.

Algorithms of machine learning reveal: indicators of sensors which can become the equipment stop reason, and difficult dependences, for example, when indications of some sensors begin to influence only provided that others are out of the set range.

For the last few years plant passed several cycles of upgrade, as a result increased the production capacities to 360 thousand tons of a kraft paper annually, using only own lesoresursa. And at the beginning of 2019 the investment project on reconstruction of JSC Segezha PPM was included Ministry of Industry and Trade RUSSIAN FEDERATION in the list of the woods, priority in the field of mastering. At the next stage of updating of capacities we plan performance improvement about 800 thousand tons of products, including added products. We hope that the implemented system of predictive analysts will help us to achieve effective objectives,

The key value of the solution for the technologist — an opportunity to prevent a stop of the machine or breakdown. And for the owner of business is a guarantee of continuity of production and reduction of financial losses. During tests the system of predictive analytics showed the efficiency, having truly predicted several breaks so now we together with the customer prepare for replication of experience on other tasks, comments Vladimir Molodykh, the director of development and deployment of software on Jet Infosystems