Lurgi GmbH optimises its Piping Systems using Simcenter FLOMASTER V7
“FLOMASTER is the right tool to solve problems which have a high impact on the safety and reliability of a plant and cannot be solved by most other solutions”
Chemical plants are continually growing and increasing in complexity. To survive in an extremely competitive market, failures are not allowed. To obtain reliable statements on the behaviour of a plant requires as comprehensive and accurate information as possible. To achieve this level of understanding, the usage of computer simulation is continuing to increase. For more than ten years Lurgi GmbH has used FloMASTER for the simulation and analysis of their complex piping systems.
The challenges for plant design can be very difficult, sometimes with total order values of hundreds of millions of euros. This combined with strict deadlines, for example if the date of mechanical completion is not met, penalties of over a hundred thousand euros per day are not uncommon.
The term ‘large-scale plant’ has changed its meaning over the years. Traditionally methanol production would have a daily production rate of 2,000 tons per day which would have been state-of the-art at the time. Today, the production rate is typically 5,000 tons per day and concepts exist for plants with the capacity of 10,000.
The challenge is not only the rapidly increasing production rate for engineering departments, but the knowledge of how to get there. More and more often multistranded plants are substituted by single-stranded production lines. In such cases, faults can impact on the whole plant, therefore high reliability of processes is absolutely necessary.
More Than 100 Years in Plant Design
One of the enterprises that has a central position in the market is the Frankfurt based Lurgi GmbH. The company which employs 1,450 people worldwide, 750 of them in Germany, is engaged in the construction of plants for the production of petro-chemical intermediates and final products as well as synthetic fuels and oleo chemistry. One of the main fields of activity for plants is gas production and treatment. Lurgi itself doesn’t have a production line, but orders are given to contractors.
Lurgi, originally “Metallurgische Gesellschaft” was founded by a company named “Metallgesellschaft GmbH” in 1897. In 1919 the company changed its name to Lurgi, which previously was the ticker address for “MetalLURGIsche Gesellschaft”. More recently, in 2007 Lurgi GmbH became part of the French Air Liquide Group.
For Plant Design - Simulation is Essential
Process technology requires highly accurate planning and outstanding engineering. The increasing complexities, even for experienced engineers, often mean experience alone is not enough for an efficient and stable solution.
To simulate and analyse models which are unable to be calculated manually, computer based simulation programs are often being used. Computer Aided Engineering (CAE) programs allow reliability planning to be enhanced significantly. Failures can be avoided or corrected before they cost additional time and money.
The “Material, Flow and Stress Analysis“ group is an internal service provider consisting of team leader Manfred Bohle and seven other specialists who provide development engineers with meaningful simulation results. The list of CFD applications used is comprehensive, including FEM analysis (Ansys) and process simulation (Aspen) as well as 3D CFD programs (Ansys Fluent). For the system simulation of the mile long and branched piping systems, Lurgi GmbH has successfully used FloMASTER for more than ten years.
Fast Model Generation and Short Computational Times
Unlike 3D CFD programs which require CAD-geometry to model systems, FloMASTER only requires schematic models which can be generated using standard components and drag and drop functionality. The behaviour of these components (pipes, valves, pumps etc.) is described by physical equations and/or empirical data which are stored in thematically oriented libraries. Alternatively, own calculations and tests can be used to understand the behaviour of components using characteristic curves or arrays obtained from the originator.
For clarity and to minimize the potential number of error sources, the models are kept as simple as possible. For example, the typical behaviour of a whole sub-system can be described by its pressure loss; nevertheless, the analysed models often obtain up to 5,000 components.
“In principle we try to generate our models as simply as possible”, explains Jürgen Bohle, “but to calculate for example pressure losses and distribution, besides the literal system you need the character and position of in- and outlet – that’s just physics.”