The Linde Group is a world leading gases and engineering company operating in more than 100 countries. With such large and widespread operations it is no surprise that maintenance activities are well run and that the company is no stranger to advanced monitoring techniques and equipment.
In USA and UK particularly, but also elsewhere in the world, on-line machine monitoring has been used for a number of years at Linde plants. And, at its Shanghai headquarters it has a large and impressive 'remote operations centre' where it monitors and tracks the process operations of all it's major gas plants in China 24 hours a day.
But Henry Aung-Kyi, Linde Reliability Engineering Manager for Greater China also knows that the key to successful maintenance operations is the right balance of techniques, systems and methodologies. An experienced and knowledgeable engineer, Aung-Kyi, is convinced of the value of highly sophisticated on-line monitoring system - for critical machinery- but, for some of the plant equipment at his gas plants, this level of technology and cost is not needed for optimum management of his company's resources. And he has, for a long time, believed in the value of predictive maintenance (PdM), even if there is a price to pay in terms of time to acquire the skills and train the workforce.
Many process and manufacturing plants operate preventive maintenance activities (PM). PM activities are usually time based, with machine adjustments and replacements parts being made on a routine basis without regard of the operational performance or condition of the major piece of equipment or its component parts. However, predictive maintenance (PdM) is a 'condition' based approach that builds a picture of the anticipated condition of a component or piece of equipment based on trend data taken while the equipment is in service. Such data is mostly vibration analysis data, but can include lubricant analysis, temperature measurements etc. Based on this trend data, and applying principles of statistical process control, an engineer can predict when is the best time to carry out maintenance activity. This allows him to plan work when it is most cost effective with a full understanding of the relative performance loss at risk. Since most PdM actions are made while equipment is in service there is minimal disruption to production and can result in substantial cost savings and higher equipment reliability.
When Linde Group acquired the BOC gases company in 2006 it embarked on a programme of integration of it's new assets in China into Linde's 'High Performance Organisation'. With Linde operations and maintenance being process driven, focusing on process optimisation and continuous improvement it was only natural for Aung-Kyi to assess the situation at each plant separately and develop a plan for improvement to the next level of technology. The long-term goal is to bring all maintenance activities at all plants to the same advanced level, and share information and knowledge for optimal benefit. But, that will take time due to differences in the existing experience at each plant and the required level of learning needed by the workforce at each plant.
So the Nanjing plant (joint venture with Sinopec SPC) was the first plant selected for a pilot application of a predicitive maintenance programme (PdM). Before running the pilot application Linde invited specialist companies, including local Chinese suppliers, to do some PdM service work at the Nanjing plant and propose a long term programme along with details of the services provided and the costs and anticipated benefits to Linde. After one year of trials and discussions with the competing vendors Linde selected SKF to start up the PdM activity for their operations in China.
When asked about the reasons for their choice Aung-Kyi said: "A number of key points separated SKF from the other vendors. As well as providing excellent services in the actual monitoring of the plant they showed their experience by identifying what equipment to monitor and went on to produce work recommendations that were far broader than the other vendors.
"Naturally, the local Chinese vendors had a price advantage, even though the SKF people are from the local SKF China organisation, but from SKF we got more than just a 'condition report'. We got constructive feedback of a depth that local suppliers could not match and a feeling that they really knew the machinery under investigation and had as much desire as Linde to manage that equipment efficiently. And that transmitted a huge amount of confidence to the local Linde engineers, which is really needed to get PdM understood and implemented properly. It wasn't enough that I was satisfied with a particular vendor. What was really important was that the plant engineers in Nanjing wanted the vendor, and effectively it was joint decision by myself and the actual Linde people receiving the service, to select SKF.
"In addition I knew we were embarking on a period of growth and that we would need an experienced and stable supplier that could support us as we grew, so I was very happy to select SKF who are also a global vendor with a large and also growing presence in China."
Consolidating and widening PdM
There are now nine Linde plants in China running PdM programmes. The SKF engineers make a three monthly visit to the plants to take measurements and follow up with in-depth discussions and recommendations as necessary. All data is uploaded to a central Linde database where all plants can access, and benefit from, the historical trending data for the particular equipment being monitored, many of which are applied in more than one plant. Simon Chen is Linde's condition based monitoring specialist for Greater China and he has many years experience in condition monitoring applications on many kinds of rotational machinery, eg compressors, motors, gearboxes, fans, pumps, etc. He oversees the internal co-ordination of knowledge transfer between the nine Linde plants to ensure the company extracts the maximum benefits in the shortest time. During the process of delivering the service to these nine plants, Chen's experience assisted greatly to enable SKF reliability engineers to make effective condition-based maintenance monitoring to Linde.
The main equipment being monitored are the high speed compressors and ancillary equipment such as pumps and expansion machines. Readings are taken at positions on the compressor that indicate the 'condition' of each of the compressor stages, and this is important to the overall efficiency of the compressor. Using SKF's portable data collector Microlog CMVA60 or CMXA70, measurements are routinely and easily collected without any disturbance to operations. Relevant data is also compared with that collected by the integrated distributed control system (DCS) in the compressor. The DCS data is needed to control and protect the production process and it measures flows, pressures, shaft vibrations etc, so that very fast automatic decisions can be made depending on what is measured. Any strong deviations between the SKF data and the DCS data are then discussed to the point of deciding what corrective action is to be taken.
Commenting on the situation now Aung-Kyi said: "At the time that the pilot application was started Linde had a lot of projects running, so the success of the pilot was very much in the 'own hands' of the vendor company. They had to really work hard to inform, educate and train the local Linde engineers how to handle and assess their machine data in order to get the most understanding how to go further.'
For more information, visit www.skf.com