|
Knowing the size and distribution of coal particles can significantly help optimize combustion and boiler performance. The traditional way to measure this important property is to collect samples from the coal pipe, then measure particle size distribution by sieving and analysis in a laboratory environment. That means considerable delay between collecting samples and seeing their results, which makes direct correlations to operational mill and boiler settings impossible.
With the increasing popularity of fuel switching, such as converting to Powder River Basin (PRB) coal to reduce plant emissions, coal-fired plant operators face new problems, such as excessive fouling and slagging. These problems can be solved by measuring particle size spectrum inside coal pipes using laser-based online technology.
Each variety of coal has its distinctive grinding and sampling characteristics, which change with production conditions. Even when standard sampling methods are used, results of sampling differ among analyses taken. These differences may stem from the inherent variability in coal properties and from the fact that measurements of a variable substance are only estimates of average values. Obtaining a true value requires a series of analyses on the same coal to determine the variability in the samples’ inherent properties and in sampling itself.
The direct cost of coal grinding is a function of the wear and tear of the powdered coal mill, which in turn is influenced by the expected particle size spectrum of the coal dust and thus of the mill settings (such as hopper load, mill speed and separator adjustment). Therefore, cost-effective mill operation requires well defined grinding quality and particle size limits. As simple as this task sounds, it is difficult to realize in practice due to the variability described.
Next to obvious mill related issues, the coal particle size spectrum affects the combustion process inside a steam generator in several ways. It influences ignition delay; combustion efficiency and loss of ignition (LOI); NOX and CO emission levels as well as tendencies for slagging and fouling. Ignition and flame properties can be determined by the ability to influence the devolatilization and composition of volatiles at the burner. That, in turn, can affect NOX levels. Also, small, medium and large particles have different flow characteristics and vary in their furnace residence times. Since a coal particle’s heat balance and combustion time depend on its size, some particles may still be heated beyond their ash melting point or even undergo reaction at their surface when entering the convection pass heating surfaces.
In any case, the furnace exit gas temperature (FEGT) plays an important role when it comes to slagging and fouling. The carbon-in-ash level and LOI are other important properties affected by the size spectrum of burning coal particles. Contributing to the problem, the particle-size spectrum is not a constant but may undergo rapid changes. Not only does it vary with the type of coal and composition (including water content) but also with the mill set-up, operating conditions and its wear and tear.
Despite the significant role that this property plays, there has not been a robust and reliable measuring system that works both online and inline. To date, the method used has been to collect samples then measure particle size distribution in the laboratory. As a result, the collection of samples itself may bias the result.
Because the method is cumbersome and costly, sampling and sieving is not often done. Worse, the time between sampling and analysis makes it impossible to use the data for feedback control. Optimization, if possible at all, will always be limited to static settings. Regardless of the plant, measurement conditions inside a coal pipe are demanding. The temperature is typically as high as 392 F and the environment is abrasive, especially if the coal has a high sand content (as with lignite). A high moisture content may further complicate matters by causing the particles to lump together, distorting results. To make things worse, elements such as coal composition, particle or load density, or total mass flow can continuously change. The particle load may reach up to 1,000 g/m³ and the particle stream will be distributed unevenly and fluctuate in time.
A laser system that uses a time-of- transition technique rather than laser light diffraction has been built to solve this problem. The Eucoalsizer system measures particle size distribution, density, velocity and temperature within a measurement volume placed at the tip of an insertable probe. By traversing the lance through the pipe, a spatially resolved distribution along the cross section of the coal pipe can be measured. In contrast to laser diffraction methods, this measurement technique also covers particles up to 5 mm, while the lower range covers particle sizes down to 20 micrometers (for example, 7.8 * 10-4 inches). This is sufficient for most pulverized coal combustion furnaces. Particle velocity distribution proves to be an essential add-on information as it allows determination of mass flow in the pipe.
The end of the measuring probe allows the exact measurement volume of the sample. Coal particles crossing this volume intersect a lattice formed by closely spaced laser beams. The laser beams are emitted from one side of the measurement volume. The intermittency pattern is detected on the opposite side by a correspondingly spaced array of fibre optical detector elements. This arrangement allows for the simultaneous, straightforward measurement of particle velocity and particle size spectrum. While the particle velocity (vp) is determined through methods of frequency analysis, the particle size (xp) is determined by the time of flight (tp) and the particle velocity.
The procedure illustrated in Figure 1 makes use of a high rate of single particle measurements. Tuneable statistical filtering methods look after double counts, particle coverage effects and other elements. The system can achieve high measurement rates and is highly reproducible. Unlike other methods, including laser diffraction, the system does not require calibration.
The system can be directly applied via an adapter to the coal pipe and there is no mechanical interaction with the particles that can bias the measurement. The system works continuously, delivers results online and can be easily integrated into an existing monitoring and control environment. Handling is through an intuitive graphical user interface. The measuring system consists of the measuring probe inserted into the coal pipe; the pneumatic unit that controls the supply of purging and cooling air to the measuring probe; and the switchgear and control cabinet with the process PC.
The measuring probe only needs to be connected to the pneumatic unit and to the EUcoalsizer control console to have the system ready to operate. The front part of the stainless steel probe contains the measuring optics. The system is designed in such a way that delicate opto-electronical components inside the probe are protected from thermal and mechanical impact. Small bursts of purging air keep the optics clean, while a temperature control of the tip of the probe helps avoid condensation and accumulation of wet coal. The complete system operation is automatically monitored and if problems arise, the operator is immediately notified. To facilitate system handling and data allocation, the lance position is automatically recorded. Data analysis, evaluation and data storage are handled by the process PC. Automatic reporting is also included that can be calibrated according to user-defined requirements.
A systematic series of test measurements was done in different power plants and with different coal types to validate the reliability of the system by simultaneously taking the traditional route of sampling and sieving. The installation is quick and easy and usually there is no calibration needed. The results of the new system show a very good match with the sieving results while having far better reproducibility.
Typically, a single point measurement requires one minute. The analysis is done online and the results are shown in terms of: - Discrete classes of particle distribution and cumulative particle distribution
- Particle distribution of any self-defined sieving class
- Velocity distribution
- Particle load density and
- System parameters
In a case study aimed at optimizing mill function on a 600 MW lignite-fired unit in central Europe, the task was to homogenize the grinding quality of the boiler’s eight mills and optimize the particle size distribution behind the separators. The purpose was to extend the range of speed control for the mills because they were constantly running at maximum speed with no room for adjustments.
It was not feasible to perform an offline optimization based on traditional sampling methods because boiler operating conditions and coal quality varied significantly. So the EUcoalsizer’s online capability was used. The system was applied in the coal pipe directly behind the separators. While the EUcoalsizer took measurements, the mill parameters (speed and load) and the adjustments of the separators were systematically varied and the effects of these variations analyzed online. This allowed for an efficient real-time optimization of the separator adjustments so that an acceptable speed control range could be established. The result was an increase of separator efficiency and particle size homogeneity (reduction of standard deviation in particle size); a reduction of mass circulation rate through separator; and a suitable control range for each mill.
Figure 1 shows the online result of particle size distribution during parameter optimization after separator. While the distribution with parameter setting No. 2 shows a large fraction of large particles, the change parameter setting in this case study achieves higher fractions of smaller particles.
A second case study focused on coal flow balancing and mill maintenance at a 300 MW European plant. The project’s objective was to equalize and balance the mass flow rate of coal dust to the burners. It is well established that meandering streaks of coal dust can cause significant distortions in the distribution of coal to the burners. Since the coal dust’s flow characteristics are also influenced by the particle size distribution, mill settings were modified to adjust to this variability. The cross section in the pipes was scanned and the particle density distribution and particle velocities were measured. That allowed the mass flow rates in the pipes to be estimated as a function of load, mill speed and separator position.
Apart from the well balanced coal mass distribution, the wear of the mills could also be evaluated by looking at the particle size distribution. With increasing wear, the fraction of large size particles typically increases markedly. To some extent mill parameters can be adjusted to compensate for changes in particle size spectrum. While the mill with the most severe attrition had to be overhauled after operating 300 hours, another mill was able to run for more than 6,000 hours before grinding quality reached a predefined limit. Thus, the study indicated that the EUcoalsizer can be used to help establish a condition-based maintenance management program.
Furthermore, EUcoalsizer enables entirely new fields of application. Its online capability makes it possible to include certain mill adjustment parameters - such as the siever setting or mill speed - into closed loop control or be included into a neural network-based optimization tool as a manipulated variable. In another application the boiler suffered from slagging problems due to NOX reduction measures (staged combustion) and a change of coal quality. Investigations showed that the particle size distribution - especially the fraction of particles less than 0.5 mm - has an influence on the slagging tendency. The optimizer took care of the primary, secondary and over fire air distribution, the load balance across the mills and the mill settings. The latter achievement was possible only with the availability of online information.
Authors: Max Starke, Dipl.-Ing. (M.Sc.), is responsible for process engineering and all practical aspects related to power plant applications at EUtech Scientific.
Hans-Joachim Schulpin is a physicist responsible for R&D, measurement technology and overall system integration at EUtech Scientific.
Michael Haug, Dr.-Ing. (Ph.D.) is Managing Director and co-founder of EUtech Scientific. He heads automation and testing applications
Michael Schreiber, Dr.-Ing. (Ph.D.) is Managing Director and co-founder of EUtech Scientific. He heads simulation and system control
Power Engineering April, 2007
Author(s) :
Max Starke
Hans-Joachim Schulpin
Michael Haug
Michael Schreiber
| 
Articles 
| View all PennWell sites | View all PennWell events
Add RSS Feed
