SIEMENS

Block 2 of a major cogeneration plant in Munich, Germany burns some 800,000 tons of hard coal per year. Every day, three or four trains, each with an average of 22 cars, deliver the fuel, which is then ground into dust and blown into a boiler via 24 burners.

The power plant produces both electricity and heat. Block 2 includes a Siemens turbine, and achieves a thermal output of 550 mega watts and an electrical output of 237 mega watts. The facility is also an economy champion, as its overall efficiency level is an excellent 85 percent.

But it isn’t just the quality of equipment at a power plant that determines output. Efficency also depends on fuel quality. Whereas plant operators in the past usually employed only one type of coal, today they utilize coal from all over the world in order to limit costs. Munich, for example, burns coal from Venezuela, South Africa, Poland, and the Czech Republic. But while this saves money, it also leads to quality fluctuations. That’s because the calorific value of the coal varies, as does its moisture and sulfur content. “Using very poor coal can cause a plant’s output to fall,” says Prof. Maximilian Fleischer, a sensor expert at Siemens Corporate Technology (CT) in Munich. “If you want to always achieve the output needed, you have to adjust the amount of coal you feed into the burner in line with that particular coal’s calorific value.”

What’s needed here is a kind of incoming goods inspection system that allows plant operators to know exactly what type of coal is being burned at any given moment. Combustion processes can then be adjusted depending on the quality of the coal fed to the furnace. Fleischer, a physicist, is working on the development of such a system using infrared (IR) spectroscopy. The idea is to employ infrared light, which is invisible to the human eye, to determine the composition of the coal. This can be done because chemical elements emit electromagnetic waves at specific frequencies after being stimulated by an energy source such as IR rays in the 0.7–2.5 micrometer wavelength range (near infrared - NIR), or the 2.5–50 micrometer range (medium infrared - MIR). Such exposure causes atoms or even entire molecules to vibrate. This energy is then radiated back, creating a “fingerprint” of the molecules present, which can be analyzed in a spectrometer.

“Measurements taken in the MIR reveal clear peaks that can be conclusively assigned to specific chemical structures,” Fleischer explains. “These connections can not be as clearly established in the NIR range because the peaks are wider and therefore overlap.” This is due to the fact that the molecular vibrations are coupled with one another - but here too, the information on molecular composition can be revealed using neural networks.

While this makes the NIR process somewhat more expensive than MIR, NIR spectrometers have a simpler design and are more robust and cheaper, costing between €15,000 and €25,000, as compared with roughly €100,000 for their MIR counterparts, which are therefore mainly used by large laboratories. NIR and slightly less precise micro-spectrometers (€2,000–€5,000) are thus ideal candidates for monitoring and controlling industrial processes such as coal combustion.

In the first phase of their project, Siemens experts demonstrated their ability to distinguish between different types of coal with the help of IR spectroscopy. Their NIR probe consists of a corona of infrared light emitting diodes (LEDs) that illuminate the fuel as it moves through a pipe from the coal grinder to the burner. A sensor absorbs the light reflected by the coal dust and sends it via a glass fiber cable to a spectrometer.

“As a next step, we want to ascertain the quality of the coal on the basis of its calorific value and sulfur and moisture content in order to optimize the combustion process,” says Paul Herrmann from Siemens Energy in Karlsruhe, Germany, who manages the project in the Instrumentation, Controls & Electrical Strategy department. “After the current feasibility study is completed, we plan to implement this second step in the early summer of 2010 - and if everything goes well, we’ll be able to get the first pilot facility up and running in 2011. That facility will be designed for long-term and continuous operation.”

Information gleaned from infrared sensing will allow power plant operators to implement better countermeasures against facility contamination, since if they knew the properties of their fuel in advance, they will be able to adjust the amount of air fed in and thus avoid furnace slag formation. Slag formation occurs because coal ash begins to soften or melt at a certain temperature, which depends on the type of coal used, and settles in the boiler as a pasty mass. Operators normally have to shut down their plants once every two to three months to remove slag in what is a complicated and costly process. Slightly increasing the air supply could keep the heat under the critical value. To achieve this goal, operators need to know the precise composition of their coal. In other words, with the appropriate measures, the furnaces wouldn’t have to be cleaned as often.