The main results of a research program directed to the investigation of fly ash cohesiveness and adherence to ESP plates and their relation with coal and ash properties are presented.

A total of 58 coal and fly ash samples coming from 6 pc boilers have been extensively characterized by means of standard and non-standard analyses to identify physical, chemical, morphological, mineralogical, and ESP performance related characteristics, like resistivity and dust layer adherence to the collecting plates. A measuring method of cohesive forces on electrostatically precipitated fly ash and a test apparatus have been developed. The ash cohesion under approximate real conditions in industrial ESPs has been measured. Correlation's between the cohesive force and the coal/fly ash characteristics have been established, and also the relation with ash resistivity.

The primary characteristic of fly ash which is generally included in specifications for electrical precipitators for coal-fired boilers is the content of various minerals in the particles. Together with the additional information included in the Coal Classification Code such as moisture and sulfur content, mineral content enables prediction of the electrical resistivity of the fly ash particles suspended in the flue gas when the boiler type and operating conditions are properly considered. Additionally, fly ash characteristic specifications usually include information about the expected particle size distribution of the fly ash.

Although it is recognized that, in some cases, particles cohesiveness and adhesiveness may require special attention in the electrostatic precipitator design, these characteristics are not included in standard fly ash specifications. Currently, the lack of such specifications may, in practice, be Justified since the rapping devices incorporated in modem fly ash electrostatic, precipitators are generally enough efficient to secure that dust layers are maintained with an even distribution and acceptable thickness over the entire collecting plate surface. It is well known, however, that fly ash produced from some Australian and South African coals, can exhibit strong cohesiveness which gives rise to a thick, uneven dust layer on the collecting plates, and thus seriously impedes electrostatic precipitator performance. Similar characteristics have been observed for fly ash from coals mined and burned in Puertollano, Spain.

Under this perspective, development and implementation of measuring methods to quantify cohesiveness and adhesiveness of collected fly ash would be beneficial. Likewise, obtaining the correlation between cohesiveness and coal characteristics would also be worthwhile so that the need for special design precautions could be identified in the planning stage. These objectives aimed the research project which is presented in this paper whose main intention was to establish the background for an extension of existing Coal Characterization Codes with respect to the ash adherence to ESP collecting plates and the cohesiveness of the fly ash particles.

The research work has included the following tasks:

• Selection and sampling of several different coals and fly-ashes.
  Samples from 6 boilers belonging to 5 Spanish coal power plants were taken. Selected coals cover the entire range from lignite to anthracite, including local coals and a bituminous coal from South Africa. Sampling points for fly ash were ESP inlet, plate curtains, ESP hoppers, and ash silos.

• Physical, chemical, morphological, and mineralogical samples characterization.ISO analysis, immediate and ultimate analyses, and ASTM D-3682-78 elemental composition of coal ash analysis have been applied to coal samples.Fly ash specific gravity, bulk density, Blaine specific surface, moisture, carbon content, loss of ignition, sulfates content, and elemental analysis by atomic absorption (SI, Al, Fe, Ca, Mg, Na, K, Ti, P) were obtained by using standard methods. Additionally, in-situ resistivity measurements and some advanced analyses, like Coulter-Counter size distribution, X-Ray diffraction (XRD), scanning electron microscopy (SEM), EDAX microprobe, and X-Ray photoelectron spectroscopy (XPS), were performed.

• Development and implementation of a representative measuring method for the cohesive forces in electrostatically precipitated dust layers.

• Correlation of dust layers adherence/cohesiveness with the coal/fly ash characteristics, including the relation with ash resistivity.  A statistical regression analysis between all the physical and chemical coal and ash properties and the cohesive forces measured with the test apparatus has been performed. The objective of this analysis was to identify the variables that are able to explain the observed cohesive strength of fly ashes and to produce predictive techniques for cohesiveness estimation.

Measuring Method for Ash Cohesive Forces

The forces that keep compacted an ESP dust layer have different origins. The whole of them is known as cohesive force, and it is made up of. Van der Waals forces, capillarity forces, forces due to solid bridges between particles, and electrical forces. The cohesive force mainly depends on the nature of particulate matter, its granular composition, surface characteristics, and the compacting and internal structure of the layer. The environmental conditions where the la er is formed are decisive in its genesis and hence, in its physical properties.

In order to reproduce the influence of the environment in the cohesiveness measurement, a laboratory apparatus has been developed in which the layer is built up in a small electrostatic precipitator, where the charge, migration, and deposition of ash particles is reproduced. The fly ash cohesion grade can be measured on this layer by means of the tensile strength evaluated from the ash effusion voltage, according with the method proposed by Pontius et al. [1]. Also the thickness of the stable layer can be measured as a result of cohesive and adhesive forces in the layer.

Due to the small quantity of ash that is normally available from the fly ash samples, a small point-to-plane based apparatus was chosen to be the most suitable laboratory corona chamber to be developed. The apparatus was aimed to fulfil the following requirements:

• To obtain fly ash deposits electrostatically precipitated, instead of manually extended.

• To be able to operate using a small quantity of fly ash.

• To allow examining the form of the adhesion and cohesion characteristics at which normal electrostatic precipitator are found to operate.

• The characteristics produced should be repeatable.

• The electric conditions should be as close as practicable to those produced in a large scale precipitator, particularly in terms of current density and electric field level.

• The nature of the fly ash layer in the apparatus and the conditions surrounding it should model, as closely as possible, those which occur in a full scale electrostatic precipitator.

The working conditions of the test apparatus simulate the conditions present in a real scale ESP. The working gas used is hot humidified air and the magnitudes that can be controlled are: gas temperature (up to 150^OC), gas moisture, flow velocity (up to 1.5 m/s), fly ash loading, and electric field (up to 6 kV/cm).

       where T is the cohesive force (tensile strength) and h is the interellectrode distance.

Due to the uncertainty about the independent variables that could be the most significative, a stepwise regression model has been chosen. This model works as a pure multivariable analysis, but it is possible to put in or take out modelling variables at every step of the execution, depending of its significance in the resultant regression. The statistical base to accept a variable requires that the global indicator of correlation adjustment be higher than the value of the F-Snedecor distribution function for the corresponding degrees of freedom and the desired trust level (a=0,05).

The best obtained correlation is as follows:

The same statistical method was applied using the measured dust layer thickness as independent variable. In this case, the regressions present very low coefficients of determination and any kind of dependence can not be significatively concluded.

To completely understand the meaning of obtained results it should be taken into account that the results of the used method are strictly limited to those relations between variables that can be mathematically expressed by means of a polynomial equation. This implies that any other existing qualitative or complex relation is not identified by the method.

The variables included in the final regression are generally considered as very important parameters in relation with ash behavior in ESPs. In this sense, ash surface enrichment on sulfates has been proven to be related with ash agglomeration, and both sodium and sulfates content in the ash have a substantial influence on fly ash resistivity in the same direction indicated by cohesive forces regression. Therefore, it becomes apparent that ash resistivity should be related to ash cohesiveness, although no regression has been found by statistical models. Figure 3 shows this relation. Evidently there is not possible to obtain a mathematical correlation between the two properties but a threshold in ash resistivity seems to exists at about 10" Ohm-cm, and ashes with a higher resistivity should be very cohesive.

This paper is based in the work performed in the Phase 3B of the CARE research program. The program was financed jointly by the European Coal and Steel Community (ECSC), Cia. Sevillana de Electricidad (Spain), and FLS Miljo Environmental Management (Denmark).