A new generation of continuous particulate matter concentration monitoring instruments is presented, covering a measuring range of 0 to 1000 mg/m3. The laser-based scattered light instrument gives an improved stability as well as excellent resolution in the low range. This enables high concentration measurement as well as especially low emission monitoring down to 0.01 mg/m3 with a resolution of 0.001 mg/m3. The high stability of the laser light source, a purge air system avoiding pollution of the optical interfaces and an automatic drift check system allow a maintenance interval of three months. The monitoring is done in an extractive sampling line heating the gas up to 160°C thus allowing the monitoring of wet gases.
Forced by governmental regulations with decreasing emission limits as the new European directives 2000/76/EG for waste incinerators and 2001/80/EG for power plants, the technology for reducing particulate matter emission has been continuously improved over the last years. Latest state-of-the-art waste incinerators and combustion facilities in standard operation emit particle concentrations of 0.1 mg/m3 or even less (TÜV-Bericht: 936/806015, 1997).
To ensure a reliable control of these low emissions and the related cleaning process, particulate matter measuring instruments also had to be improved. These instruments should have at least a resolution of 10% of the lowest emission values given above, i.e. 0.01 mg/m3, to make changes detectable and to demonstrate the instruments functionability. On the other hand, peak emission values in countries with more moderate limits or for older facilities can reach up to 200 mg/m3. So a multi-purpose instrument suited for installations in a variety of industries and applications like waste incinerators, power plants, refineries, cement kilns, etc. has to cover a broad measuring span with a high dynamic resolution.
In addition, the new European Standard EN 14181 for quality assurance of automated measuring systems and the new suitability requirements for continuous emission monitoring systems in EN ISO 14956 brought further requirements for new measuring equipment as automatic drift and reference point check including logging functions.
2 Principle of detection
2.1 Scattered light measurement
The newly developed StackGuard instrument is based on its predecessors CTNR and KTNR, which have been installed in about 350 facilities since more than 20 years and successfully used in a variety of applications and industries. Similar to these instruments, the StackGuard is using the scattered light detection for particulate matter concentration measurement (Fig. 1). A light beam with certain intensity is striking the sample. Due to absorption and scattering at the dust particles, the transmitted light intensity is reduced. This effect is used in the measurement principle of opacimeters especially for high particle concentrations. To detect high and also low particle concentrations, the detection of the scattered light is necessary. In the StackGuard, a certain part of the scattered light is detected under a 20° angle.
The intensity of the scattered light shows in this concentration range a linear relation to the concentration of the particles, as long as the particle properties like size, shape, color, or refractive index do not change. As demonstrated in paragraph 4, a calibration of the instrument can be done using a linear fit between the scattered light intensity and the particulate matter concentration determined by a reference method. Using this calibration, the read-out of the instrument can be displayed in mg/m3. For different particle properties, the scattered light intensity and distribution might also be different (Huber and Frost, 1998). For most applications in dust emission measurement these variations are negligible in the concentration range of interest and the linear fit is still appropriate. In some applications, however, especially with a wide span of the concentration or a variety of different fuels, the particle properties like e.g. the size distribution are changing with changing concentration. In these cases the best fit method for the calibration curve might be different from linear, e.g. quadratic or logarithmic. Due to this fact, for all light scattering CEMS, a calibration has to be done on-site after installation of the instrument.
2.2 Optical setup
The completely revised setup of the StackGuard instrument is given in Figure 2. The measurement is done in two phases:
Phase 1: The laser remains switched off. The zero points of the three detectors are determined. This is especially important to reduce any zero-point offset and thus improve the accuracy at low dust levels. The laser is a temperature controlled semiconductor laser light source with a wavelength of 650 nm.
Phase 2: The temperature-stabilized laser is switched on. The light impinging on the two reference detectors and the 20° scattered light detector is measured. The measurement reading is calculated from these signals.
Using this two-beam method, a continuous determination and compensation of all uncertainties and errors due to intensity changes and fluctuations of the light source and ageing effects of the electronics is achieved. Any additional drifting is well-proven to be below 2% per 3 months and can be easily compensated by re-calibrating the instrument by inserting glass-rods with a given opacity.
Moreover, the improved sensitivity of the light detecting circuits gives a dynamic range of more than 105 for the scattered light intensity thus covering a concentration range from 0.0005 mg/m3 PLA to 100 mg/m3 PLA. The appropriate measuring range is selected by automatic range switching.
The signal ratio of reference detector inlet (3) to reference detector outlet (7) gives an information on the pollution of the optical components and is monitored continuously. When a certain limit is exceeded, a warning is generated.
An additional improvement in the StackGuard is an automatic zero and reference point drift check system (see also Figure 2). Every 24 hours, the system undergoes the following check procedure:
1. Measuring position: Part of the laser light is decoupled with a partially permeable mirror (10). When the checking unit is in the measuring position, the attenuator (11) blocks this checking beam. The shutter (12) is open and lets the scattered light from the flow cell pass through. It is detected by the scattered light detector (8).
2. Reference point check value: The attenuator (11) lets pass the checking beam unhindered. It strikes the scattered light body (13) and produces scattered light that is picked up by detector (8). The shutter (12) is closed and blocks the scattered light coming from the flow cell.
3. Zero point check value: The attenuator (11) reduces the checking beam intensity to 1%. The scattered light produced by the scattered light body is picked up by the scattered light detector (8). The shutter (12) is closed and blocks the scattered light coming from the flow cell.
The periodically detected values are compared with the nominal values. In the event of an excessive deviation, a warning is given.
All these features described above from the highly stable laser light source to the various self control mechanisms allow an extension of the maintenance interval to 3 months, which gives significant relief to the operators.
3 Extractive sampling system
For flue gases which are normally or occasionally saturated with water or acid gases, e.g. after wet scrubbers or flue gas desulfurication units, interference with optical measurement may occur due to fluid droplets also scattering the light. To overcome this problem, the sample for the scattered light measurement in the StackGuard is extracted from the stack and heated in a sampling line up to 160°C. Some national regulations like PS 11 in the U.S. even require extractive arrangements for this type of applications. Figure 3 shows the complete setup of the StackGuard including the loop line for sample preparation.
The sampling line (c) takes the sample from the stack (d) at a flow rate of 60 m3/h using a blower (e). This high flow rate is chosen for two reasons. One, to ensure representative sampling, and second, to prevent deposits by using a high gas velocity in the loop line. From this loop line, the sample for the photometer (b) is taken with a flow rate of 35 l/min. The probe of the CTNR sampling line in the stack uses over-isokinetic sampling with a factor of 1.5 to reduce the relative error for changing gas flow velocities in the stack. The absolute error is considered with the calibration of the instrument. Using temperature controlled heaters, the sample temperature is kept at a level of at least 160°C. The control unit (a) handles the operation, control, display, and signal processing functions.
The photometer is built up in two completely separated compartments. The first compartment contains all the electronics, light source, detector, chopper, and the related optical elements. The sample compartment contains the flow cell for the detection of the particles by scattered light measurement (Figure 4). The sample enters the flow cell from top (P1). A heated and filtered purge air flow (S1) surrounds the sample stream with a protective shroud, and additional purge air injection in front of the optical lenses S2 prevents soiling of the optical surfaces L1 and L2. Sample and purge air leave the cell at the bottom outlet (P2).
This setup makes the cell insensitive to any changes of the signal caused by soiling even for gas streams with high particle load. As given above, the drift of the complete instrument including errors by potential deposits on the windows, is less than 2% per 3 months and can be easily adjusted using optical control glasses.
Figure 4: Sample flow cell of StackGuard Figure 4: Sample flow cell of StackGuard
All instruments are factory calibrated related to PLA (polystyrene latex aerosol) spherical particles with a diameter of 1 micrometer to ensure the linearity of the calibration over the complete measuring span. The on-site calibration relating to the actual particle load usually can be done just by inserting a fix factor between the PLA value and the real value in mg/m3. Usually, this factor can vary between 1.0 and 40.0, depending on the characteristics of the particles as size distribution, material, surface structure, refractive index, etc.
For the calibration of the particulate matter CEMS in these low ranges, exceeding accuracy has to be taken for the reference method. It is obvious, that the accuracy of the calibration (and thus the accuracy of the continuous measurement done later on) cannot be better than the accuracy of the manual method. So either a well-known method like given e.g. in the guideline VDI 2066 should be used, or paired data with two trains should be collected for the manual method. The latter is especially the case, if a well-known method is modified, or if the results are not satisfying. The paired data approach will allow evaluating in a first step the accuracy of the manual method before starting the calculation of the calibration curve. To test the manual method, a plot of train 1 versus train 2 should be done, where the slope should be near to 1 and the correlation coefficient should be 0.95 or better. Otherwise the manual method has to be improved. The method applied for the results given below is the manual method according to guideline DIN EN 13284, part 1, April 2002.
5 Field test
To demonstrate the increased sensitivity and stability of the StackGuard instrument compared to its predecessors, a test installation at a municipal incinerator waste incinerator has been done. The unit was installed specifically in the intermediate gas duct following the scrubber and ahead of the flue gas nitrogen oxide control. It has been investigated within the suitability test of the system (TÜV-Bericht 936/21202165/A, 2005).
The measuring systems were installed in a horizontal flue gas duct. The inlet and outlet sections were at least three times the diameter. The duct’s circular cross-section is 1.60 m in diameter. The intake openings of the measuring installations were arranged at the measuring level, spaced less than 0.3 m apart. The measuring level for the reference measurements was located in the measurement volume of the measuring installations.
The particle emission of this facility is under normal operating conditions around 1 mg/dsm3. Values between 0.4 mg/dsm3 and 2.4 mg/dsm3 have been seen during the campaign of 6 months. An artificial spreading of the values by modification of the operating conditions is no longer allowed according to EN 14181 standard. Two instruments have been installed for 6 months and within this time two measuring campaigns of 17, resp. 16 test runs using manual reference method have been carried out. The measuring range used for this test was 0..1 mg/m3 PLA. The results are given in table 1 and figures 5 and 6, respectively.
Although the low values of the manual reference method have been at the limit of this method, the statistical values demonstrate a good reliability. Also both investigated instruments react very homogeneous. Full scale in the selected measuring range of 1 mg/m3 PLA is corresponding to a particulate matter concentration of approximately 12.5 mg/m3. The lowest detected concentration was 0.4 mg/dsm3. The lowest detectable concentration, which is mainly determined by the stray light background, was determined to 1% of full scale in the lowest possible range of 0 to 0.05 mg/m3 PLA, i.e. 0.006 mg/dsm3.
The tests reported in this paper have been successfully performed by TÜV Rheinland in Germany, leading finally to approval of the StackGuard for use with power stations and waste incinerators according to 13th and 17th BImSchV and TA Luft.