Elsevier

Minerals Engineering

Volumes 43–44, April 2013, Pages 154-158
Minerals Engineering

A model for pulverised fuel production in an air-swept tube mill

https://doi.org/10.1016/j.mineng.2012.11.005Get rights and content

Abstract

The grinding process of an air-swept tube mill can be modelled adequately using the selection and breakage functions approach to modelling comminution. However to model Pulverised Fuel (PF) production, all the other sub-processes that include: release of the ground product from the ball charge, internal air-classification within the mill and classification of an installed external classifier must be considered. Inefficiency in any one of these process steps can negate the overall mill capacity with little compensation from the other processes.

Using industrial data from Kendal power station and data from a scaled down pilot mill at the University of Witwatersrand, a system of sub-process models have been defined and a simulator developed. We were able to simulate mill performance under different loading conditions over the entire liner life. The model details and some of the results of industrial modelling are discussed.

Highlights

► Air swept mill model is developed based on PBM grinding model, virtual internal mill classifier and physical external classifier. ► A particle transport model from grind site that is dependent on liner condition is proposed. ► After parameterisation model remarkably predicts real industrial mill operation.

Introduction

For coal power plants, the main objective of grinding is to produce a Pulverized Fuel (PF) product that meets the boiler requirement in terms of size distribution and moisture content. This is normally accomplished in vertical roller mills or tube mills. The Mineral Processing Research Group at the University of Witwatersrand in South Africa has in collaboration with a number of Eskom power plants been gathering data to help develop an air swept tube milling model.

From the data that was available from the plants and the experiments conducted in a pilot mill, key process steps that affect throughput and product size distribution were identified. Each of these process steps were put together to develop a simulator that modelled the measured behaviour well.

A number of researchers have proposed coal models for tube mills (Austin et al., 1984a, Austin et al., 1984b, Kolacz and Sandvik, 1996) which however do not address how liner condition can affect PF production in a mill. It is known that a mill loses capacity by as much as 30% as liner profile changes due to wear and can even be more dramatic if the liner type is changed. The air flow rate, temperature and variation in coal moisture are other issues that have not been addressed before but their impact is critical to the PF production process, and this necessitated the development of a temperature-momentum balance model that interacted with the grinding model. This has been discussed elsewhere (Makokha et al., 2009).

Section snippets

Milling theory

The grinding model must address both the properties of the material being ground and the role of the equipment in the size reduction process. Traditionally, in the coal industry, the Hardgrove Grindability Index (HGI) has been used as an indicator of how easy or difficult it is to grind the coal. The main problem with this parameter is that the grinding mechanism in the test equipment does not represent what is typical in a tube mill. It was thus necessary to base the grinding model on the

Modelling processes involved in PF production

To model PF production we must track the path taken by the coal once it enters the mill and identify key processes involved before the coal exits as PF via an external classifier. It should be noted that any problem with one of these processes can seriously cripple mill productivity even if the rest are functioning well. For example, a soft coal with high moisture even if it grinds easily will still result in overall poor mill performance because the moist fine coal does not transport easily

Air system

Before discussing the application of the model to an industrial mill, attention is drawn to the air flow system shown in Fig. 3.

The hot primary air is introduced to the mill with the new coal feed and the gas rate is directly proportional to the coal feed rate. The recycled oversize material from the classifier also re-enters the mill with the new coal feed. A damper is in place to ensure sufficient drying air goes into the mill while the remainder bypasses the mill and goes directly to the

Conclusion

A model based on various processes of grinding has been developed and after obtaining suitable parameters, encouraging results have been achieved. It must be appreciated that this feat is achieved with the additional burden of modelling the mill temperature and pressure. There is a need to obtain more sampling data for the reject stream before considering ways of improving the model. Testing the model on other plants with different air configuration systems and different coal properties will

Acknowledgement

The financial support from Eskom of this research work is gratefully acknowledged.

References (9)

There are more references available in the full text version of this article.

Cited by (7)

  • The impacts of dry stirred milling application on quality and production rate of the cement grinding circuits

    2020, Minerals Engineering
    Citation Excerpt :

    Napier-Munn et al. (1996) stated that the outcomes of the approach were accurate in estimating both the size distributions and flow rates. Bwalya and Moys (2013) proposed a model structure for the tube mills by using population balance model and found that the model structure fits the experimental data with the least sum of error. Benzer et al. (2001) and Altun (2018) had applied this methodology to simulate closed circuited ball milling operation of cement grinding circuit for energy optimization purposes.

  • Developing a methodology to model and predict the grinding performance of the dry stirred mill

    2019, Minerals Engineering
    Citation Excerpt :

    Napier-Munn et al. (1996) stated that the outcomes of the approach were accurate regarding both the size distributions and flow rates. Bwalya and Moys (2013) proposed a model structure for the tube mills by using population balance model and found that the model structure fits the experimental data with the least sum of error. Benzer et al. (2001) applied this methodology on closed circuited ball milling operation to simulate the cement grinding circuit for energy optimization purpose.

  • Energy and cement quality optimization of a cement grinding circuit

    2018, Advanced Powder Technology
    Citation Excerpt :

    Benzer et al. [22] also focused on closed circuited milling operation and similar approach was applied to simulate the circuit for optimization purpose. Bwalya and Moys [23] proposed a model structure for air swept tube mill where population balance model was considered in defining the grinding process that also included the transportation action. The results concluded that the model structure fits the experimental data with the least error hence were all in good agreement.

  • Modeling of unburned carbon in fly ash and importance of size parameters

    2016, Fuel Processing Technology
    Citation Excerpt :

    During steady state combustion of PC particles, i size material fed and i size material combusted and i size material is becoming ash. Thus, modification of the grinding model from the study of Bwalya and Moys [8] can be used as a combustion model and the steady state combustion process for a specific size fraction (i) can be stated as: “Rate of combusted i-size material in product = rate of addition of i-size material in feed + rate of production of i-size material by combustion of all larger size − rate of combustion of i-size material”.

  • Exploring ball size distribution in coal grinding mills

    2014, Powder Technology
    Citation Excerpt :

    From the experimental data and the scale-up procedure, the dimensions and parameters that apply to the test and industrial mills are tabulated Table 3: After obtaining the scaled-up parameters we were able to run industrial scale simulations using air-swept mill software that we have developed and has been described in detail in another paper [13]. Using the dynamic wear model that can flexibly compute the BSD for various TUPs, we predict the expected steady state BSD if the TUPs were as given in Table 4a.

View all citing articles on Scopus
View full text