Gas turbine performance depends on the cleanliness of inlet air, the optimal air being free from solid and liquid impurities.

Air impurities are typically collected with fibrous filters, but these have a number of disadvantages.

The cleaning efficiency of fibrous filters varies markedly with particle size. But the particle sizes to be encountered by a particular gas turbine depend on the environment in which it is operating.

Standard filter classification tests often measure efficiency and pressure drop at only one particle size, which can be misleading and not applicable to real conditions. A typical classification for fine filters would be the efficiency of removing 0.4 µm test particles with a dust load corresponding to a 450 Pa pressure drop. This methodology is problematic.

The standard classification tests do not adequately consider low initial pressure drop and the rise in pressure drop as the fibrous filter becomes loaded with dust, nor the decrease in performance as the filter material becomes aged and deteriorates.

Different size particles cause different problems in gas turbines.

Very small particles are typically not picked up by fibrous filters and cause fouling in the compressor, which decreases the output with time. Compressor fouling can be dealt with by on-line and off-line washing, but the latter can of course contribute to unavailability.

An increasing pressure drop, as fibrous filters become loaded with dust in use, has a similar effect on performance. A two-stage fibrous filter cleaning system typically has a pressure drop of around 7 mbar. The compressor must move the air through the pressure drop and thus a higher pressure drop decreases overall plant efficiency and output.

Because of the increasing pressure drop, fibrous filters must be replaced regularly, incurring spare part costs and unavailability. Losses can be significant if the replacement has to be done in unscheduled outages due to blocked filters.

Because water and dirt cannot be removed from fibrous filters during operation, filters become eventually saturated with water and can release large droplets and dirt downstream in carry-over/wash-through incidents.

Such droplets pose the risk of icing in the inlet air duct in cold conditions, with the possibility of pieces of ice damaging compressor blades – as has been encountered at several plants. If a compressor blade breaks, part of the blade may drift downstream and cause massive destruction.

The risk of icing is usually averted by preheating the inlet air under certain ambient conditions. But, as the density of air decreases when its temperature rises, inlet air preheating decreases the air mass flow through the compressor, leading to reduced gas turbine performance. If the heating air is itself taken from the compressor, the air flow is further decreased.

Installation of the preheating equipment also requires a major investment, of course.

Electrostatic precipitation

As an alternative to fibrous filters, electrostatic precipitation can be used. This entails charging particles in the inlet air and removing them using an electrostatic field created between two electrodes.

The negative discharge electrode is significantly smaller than the positive collection electrode. Thus the electric field is strongest near the discharge electrode and weakens towards the collection electrode. Normally, gas ions are neutral, but the strong electric field near the discharge electrode detaches electrons and thus creates free electrons and positive gas ions. Positive ions migrate towards the negative discharge electrode. The electric field accelerates the free electrons towards the positive collection electrode. These high velocity electrons collide with gas molecules. In the collision, another electron, another gas ion and visible radiation energy are produced. The area of the electric field where the collision produces free electrons is called the corona region. Each additional electron is further accelerated, and more collisions occur. The amount of free electrons increases in this avalanche-like process in the corona region.

Beyond the corona region, the electric field rapidly diminishes. Electron velocities are no longer sufficient to detach more electrons. Instead of producing new free electrons, the free electrons are captured by gas molecules, and negative gas ions are created. Negative gas ions migrate towards the positive collection electrode. Negative gas ions collide with dust particles and are caught in them. The particles become negatively charged and start moving towards the collection electrode.

The negatively charged dust particles are collected on the surface of the collection electrode and are removed by draining the surface with water.

Fortum’s system based on these principles is called FACT – Fortum Air Cleaning Technology – and has been developed in-house, with a number of R&D projects over the 1990s. It basically consists of a combination of different electrostatic precipitators, with the design optimised for each plant’s ambient conditions and surroundings.

The dimensions of a FACT system depend on the quantity and quality of the air to be cleaned. A critical measure is the time a particle spends in the electric field and the speed of the air flow. These depend on the required cleaning and water removal efficiency and the voltage levels to be used.

Operating FACTs

Experience with the FACT system at the Glanford Brigg power station in the UK (formerly owned by Fortum) has been positive. The payback time for this installation is estimated to be less than three years and the lifetime of the investment is expected to be over 10 years, promising a significant net present value.

Filter change eliminated

The dirt collected is removed by on-line washing of the electrostatic precipitators, while every year a more thorough wash is done, using pressurised water. Elimination of the need for filter changes has decreased maintenance work costs, spare part costs and improved plant availability. The only operating costs for FACT come from electricity and water. Electricity consumption is about 100 kW, water consumption is about 10 cubic meters per year and there are no specific water quality requirements. Electricity consumption depends on the amount of air cleaned. Water consumption depends on the quantity of impurities in the ambient air.

Decreased fouling

Unlike a fibrous filter, an electrostatic precipitator efficiently removes even the smallest particles – which cause most of the compressor fouling.

Mesurements show that FACT is much more efficient than conventional filters at particle sizes less than 1 µm.

Cleaning efficiency can also be measured indirectly through compressor fouling, which is reflected in efficiency. Figure 6 shows how, with FACT, compressor efficiency does not degrade, as happens with conventional filters.

Decreased air preheating

Water in the inlet air – mostly arising from fog, rain or snow – and the consequent risk of icing in compressor IGV blades or inlet air ducts, is eliminated by the FACT system, while the lower pressure drop across it, compared with a fibrous filter, decreases the tendency for condensation in the duct when humidity is high. At Brigg, the amount of inlet air heating needed decreased by over 90 per cent. Figure 7 shows the results of a water removal test. During the test the FACT system was switched on and of and different voltage levels were tested. When the FACT unit is on, over 99 per cent of droplets were removed.

Decreased pressure drop

The efficiency and output of a gas turbine plant improve when pressure drop decreases. The pressure drop across a FACT system is about 1 mbar. At Brigg this meant that the pressure drop decreased by about 4 mbar. Pressure drops for four identical gas turbines are shown, as well as the reduction in pressure drop when fibrous filters are replaced by FACT.

Improved availability

In addition to elimination of filter replacements and decreased need for off-line washing, the FACT system has improved availability by avoiding carry-over incidents and the risk of blocked filters. The FACT system has no structures which can become blocked or saturated with water.

Competitive alternative

Overall, the electrostatic gas turbine inlet air cleaning system developed by Fortum has proved to be a competitive alternative to fibrous filters. The efficiency is at least as high and the need for inlet heating is minimised. Additional benefits are a lower pressure drop, removal of water droplets, which lowers the risk of icing, decreased maintenance costs, and less gas turbine degradation. Although the investment costs of an electrostatic precipitator are higher than the costs of a fibrous filter, FACT is an economically competitive solution.