Saras optimized production using software for Aspen Mtell maintenance

2019: Optimization of production using software for Aspen Mtell maintenance

On July 1, 2019 the AspenTech company reported that at the oil refinery specialists of Saras applied algoritmymashinny training to job analysis of key types of the equipment and selected the software of Aspen Mtell for reliability augmentation of the transactions on oil processing production requiring considerable capital investments and available industrial assets.

The Italian energy company Saras used this approach at the 300 thousand barrels oil refinery in day which is located in the Mediterranean region. The company applied machine learning to the equipment of four types: raw pumps, pumps of flushing oil, compressors of make-up hydrogen and recirculation compressors. The program of implementation of digital technologies was started for only several weeks and allowed to predict precisely failures for each component of the equipment without false operations.

Thanks to these opportunities the company predicts failures with advancing in 24–45 days. Saras also set the object to reduce unplanned idle times by 10 days, to increase income by 1–3 percent and to reduce costs for maintenance and operation of the oil processing equipment for 1–5 percent.

Task:

• Reliable work of the 300 thousand barrels oil refinery in day and the 575 megawatts power plants with the combined cycle of complex gasification.
• Strategic task of reduction of idle times and costs for maintenance.

The initial project which was implemented all for few weeks included creation of software agents of Aspen Mtell for identification of failures of the equipment. Data on conditions and processes were provided for these agents.

For recirculation compressors the created software agent showed the high accuracy (91%) with time of anticipation of 30 days. The template received on the basis of work of this agent allowed to predict two events for valves:

• failures of the valve because of high temperature at the exit, with time of anticipation of 39 days;
• need of replacement of the valve because of a hardware failure, with time of anticipation of 25 days.

Software Aspen Mtell allowed to implement a pilot project for only several weeks and to be highly appreciated by Saras for the deployment speed, high accuracy of early detection of failures of the equipment, lack of false operations and readiness for scaling of the solution to the level of all system. Saras is going to attract related company Sartec which is engaged in industrial automation, to deployment of this software at all oil refinery. In the initial project the equipment of four types was considered:

• raw pumps — give the liquid received from other installations to the line of the heat exchanger;
• pumps of flushing oil — return flushing oil from the separator on the mixer;
• compressors of make-up hydrogen — give make-up hydrogen from a hydrogen source to the line of the main heat exchanger;
• recirculation compressors — perform recirculation of the hydrogen arriving from the line of the heat exchanger.

For a pilot project the following objects were set:

• Exact detection of characteristic models of normal operation of the equipment, failures and anomalies.
• The solution should issue early notifications with considerable time of anticipation from detection before the actual failure.
• It is necessary to receive signatures of failures which will allow to detect failures of the same or similar equipment according to unknown data.

Software agents precisely defined a possibility of specific potential failure – and made it without false operations. For agents of Aspen Mtell data from 52 million sensors, including data on conditions and processes were used. The command studied 163 problems with quality of data (incorrect or missing values) and carried out cross reconciliation with the history of orders on four types of the equipment, including according to 340 orders for preceding periods. The history of maintenance included 17 codes of classification of problems.

On the project all tasks are carried out, and agents of Aspen Mtell successfully predicted failures with essential time of anticipation:

• High temperature of valves: 36 days
• Replacement of oil laying: 45 days
• Replacement of laying of pumps: 33 days
• Replacement of gas laying: 24 days

Based on this research, RON Beck, the director of strategy and the market of AspenTech company, says that it is not necessary to try to embrace the immensity. It is better to focus on a specific problem much quicker to achieve important results for all business. The Saras oil refinery is an example of approach "begin with small". The company well formulated approach to use of digital technologies and saved it. It applied the advanced algorithms of detection of signs of problems and could predict failures with time of anticipation of 24-45 days, reduce unplanned idle times by 10 days, increase income by 1–3 percent and reduce costs for maintenance and operation of the oil processing equipment for 1–5 percent.

Thanks to effective management of industrial assets and also collection of data on the history of production operations and on transactions in real time along with the checked advanced technologies - the organizations can transform process of maintenance of assets, providing optimal reliability, prolonging service life of assets and increasing profitability of the enterprises.

2017: Choice of the software of Aspen Mtell

On February 1, 2018 company Aspen Technology, Inc. announced the choice of the software of Aspen Mtell by Saras S.p.A. company for reliability augmentation of the transactions on oil processing production requiring considerable capital investments and available industrial assets.

Software of Aspen Mtell is a part of the software package aspenONE Asset Performance Management (APM) combining knowledge of production processes with the solutions Big Data and machine learning for increase in efficiency in the field of engineering design, operation and service of industrial assets for all their lifecycle. Aspen Mtell collects data as on the history of production operations and maintenance, and on transactions and maintenance in process in real time precisely to detect signs of failures which precede possible breakdowns and degradation of industrial assets and also to predict future failures and to order detailed recommendations to actions for mitigation or the solution of these problems.

The Saras plant in Sarroch. Photo: lanuovasardegna.it

The choice of Saras company for benefit of software of Aspen Mtell was based on competitive selection process of suppliers by the principle of "justification of the concept" which was concentrated on the crucial oil processing equipment, such as big compressors and pumps. With Aspen Mtell it was succeeded to satisfy all necessary conditions within several weeks. The tender prize from Saras received thanks to the speed of deployment, accuracy of detection of failures of assets at early stages to a possibility of avoiding of false alarms and capability to scale the solution within all system became result. The Saras company is going to use the sisterly engineering enterprise Sartec specializing in industrial automation for large-scale deployment of Aspen Mtell at the oil refineries.

Thanks to effective management of industrial assets and also collection of data on the history of production operations and on transactions in real time along with the checked advanced technologies of the organization can transform process of maintenance of assets, providing optimal reliability, prolonging service life of assets and increasing profitability of the enterprises, noted in AspenTech.

"Reliability augmentation of production operations and assets positively influences a broad spectrum of questions, from decrease in current expenses on maintenance before planning of abnormal conditions and deviations in production and also exceptions of abnormal or unplanned idle times and effective control over unpredictable expenses and requirements. The Saras company expects to reduce costs at the expense of this initiative which is a part of the important project of digitalization of production". Alessandro Zucca, digital production platforms Manager in the "production operations and assets" direction of Saras company