Découvrez votre prochain livre préféré

Devenez membre aujourd'hui et lisez gratuitement pendant 30 jours
Mineral Processing Design and Operations: An Introduction

Mineral Processing Design and Operations: An Introduction

Lire l'aperçu

Mineral Processing Design and Operations: An Introduction

4/5 (9 évaluations)
1,631 pages
278 heures
May 2, 2016


Mineral Processing Design and Operations: An Introduction, Second Edition, helps further understanding of the various methods commonly used in mineral beneficiation and concentration processes. Application of theory to practice is explained at each stage, helping operators understand associated implications in each unit process. Covers the theory and formulae for unit capacities and power requirements to help the designer develop the necessary equipment and flow-sheets to economically attain maximum yield and grade.

This second edition describes theories and practices of design and operation of apparatus and equipment, including an additional chapter on magnetic, electrostatic, and conductivity modes of mineral separation. Basics of process controls for efficient and economic modes of separation are introduced.

  • Outlines the theory and practice in the design of flow sheets and operation of an integrated mineral processing plant
  • Introduces the basic magnetism, electrostatic, conductivity, and dielectrophoresis properties of minerals and related separation techniques
  • Describes automation in mineral processing plants allowing maximum yields and consistent high concentrate grades
  • Outlines problems and offers solutions in the form of various examples
May 2, 2016

À propos de l'auteur

25 years in Managerial Position in industry involved in operation, production and research and 18 years in academia (Teaching Undergrad, & Post Graduate courses and conducting Research). Retirement followed by work as Consultant

Lié à Mineral Processing Design and Operations

Livres associé
Articles associés

Catégories liées

Aperçu du livre

Mineral Processing Design and Operations - Ashok Gupta

Mineral Processing Design and Operations

An Introduction

Second Edition

Ashok Gupta

Retired Head of Metallurgy Department

WA School of Mines, Kalgoorlie

Curtin University of Technology, Western Australia

Denis Yan

Consulting Metallurgist

Minerals Engineering Technical Services

Perth, Western Australia

Table of Contents


Title page



Preface to the Second Edition

Symbols and Units

Chapter 1: Mineral Sampling


1.1. Introduction

1.2. Statistical Terminology

1.3. Mineral Particles Differing in Size – Gy’s Method

1.4. Mineral Particles of Different Density

1.5. Incremental Sampling

1.6. Continuous Sampling of Streams

1.7. Sampling Ores of Precious Metals

1.8. Sampling Nomographs

1.9. Problems

Chapter 2: Particle Size Estimation and Distributions


2.1. Introduction

2.2. Methods of Size Estimation

2.3. Particle Size Distribution

2.4. Combining Size Distributions

2.5. Problems

Chapter 3: Size Reduction and Energy Requirement


3.1. Introduction

3.2. Design of Size Reduction Processes

3.3. Energy for Size Reduction – Work Index

3.4. Estimation of Work Index for Crushers and Grinding Mills

3.5. Factors Affecting the Work Index

3.6. Approximation Methods for Work Index

3.7. Work Index and Abrasion

3.8. Problems

Chapter 4: Jaw Crusher


4.1. Introduction

4.2. Design of Jaw Crushers

4.3. Jaw Crusher Operation

4.4. Jaw Crusher Capacity Estimation

4.5. Critical Operating Speed

4.6. Power Consumption Estimation

4.7. Problems

Chapter 5: Gyratory and Cone Crusher


5.1. Introduction

5.2. Design of Gyratory Crushers

5.3. Gyratory Crusher Circuit Design

5.4. Gyratory Crusher Operation

5.5. Capacity of Gyratory and Cone Crushers

5.6. Power Consumption of Gyratory and Cone Crushers

5.7. Problems

Chapter 6: Roll Crushers


6.1. Introduction

6.2. Design of Roll Crushers

6.3. Operation of Roll Crushers

6.4. Capacity of Roll Crushers

6.5. Power Consumption of Roll Crushers

6.6. High Pressure Grinding Rolls (HPGR)

6.7. Operation of HPGR

6.8. Capacity of HPGR

6.9. Power Consumption of HPGR

6.10. Problems

Chapter 7: Tubular Ball Mills


7.1. Introduction

7.2. Design of Tubular Mills

7.3. Operation of Tubular Ball Mills

7.4. Estimation of Mill Capacity

7.5. Mill Power Draw-Mechanical Methods

7.6. Problems

Chapter 8: Tubular Rod Mills


8.1. Introduction

8.2. Design of Rod Mills

8.3. Operation of Rod Mills

8.4. Rod Mill Capacity

8.5. Rod Mill Power Draft

8.6. Rod Mill Drive

8.7. Problems

Chapter 9: Autogenous and Semi-Autogenous Mills


9.1. Introduction

9.2. Design of AG/SAG Mills

9.3. Operation of AG/SAG Mills

9.4. AG/SAG Mill Power

9.5. Choice of Options between AG and SAG Mills

9.6. Problems

Chapter 10: Stirred Mills – Ultrafine Grinding


10.1. Introduction

10.2. Vertical Mills

10.3. Horizontal Disc Mill – IsaMill

10.4. Design Testwork

10.5. Problems

Chapter 11: Mathematical Modelling of Comminution Processes


11.1. Introduction

11.2. Basis for Modelling Comminution Systems

11.3. Mathematical Models of Comminution Processes

11.4. Modelling Crushing and Grinding Systems

11.5. Problems

Chapter 12: Screening


12.1. Introduction

12.2. Basic Design Features of Screens

12.3. Operation of Straight Screens

12.4. Capacity and Screen Selection of Straight Screens

12.5. Operation of Curved Screens

12.6. Modelling of the Screening Process

12.7. Screening and Crushing Circuits

12.8. Problems

Chapter 13: Classification


13.1. Introduction

13.2. Design Features of Mechanical Classifiers

13.3. Designing the Pool Area of Mechanical Classifiers

13.4. Design Features of Centrifugal Classifiers

13.5. Operation of Mechanical Classifiers

13.6. Capacity of Mechanical Classifiers

13.7. Operation of Centrifugal Classifiers

13.8. Hydrocyclone Models

13.9. Hydrocyclone Capacity

13.10. Hydrocyclone Circuits

13.11. Problems

Chapter 14: Solid – Liquid Separation – Thickening


14.1. Introduction

14.2. Design Features of Thickeners

14.3. Thickener Design-Batch Process

14.4. Thickener Design-Continuous Thickeners

14.5. Operation of Thickeners

14.6. Thickeners in Circuits

14.7. Problems

Chapter 15: Solid Liquid Separation – Filtration


15.1. Introduction

15.2. Design Features of Filters

15.3. Operation of Filters

15.4. Capacity of Continuous Vacuum Filters

15.5. Washing of Deposited Cake

15.6. Drying of Deposited Cake

15.7. Optimum Thickness of Cake

15.8. Filtering Media

15.9. Filtering Aids

15.10. Filtration in Mineral Processing Circuits

15.11. Problems

Chapter 16: Gravity Separation


16.1. Introduction

16.2. Particle Settling Rates

16.3. Gravity Separation Operations

16.4. Jigs

16.5. Differential Motion Table Separators

16.6. Flowing Film Concentrators

16.7. Dense (or Heavy) Media Separation

16.8. Gravity Separation Performance

16.9. Problems

Chapter 17: Magnetic and Electrostatic Separation


17.1. Introduction

17.2. Atomic Theory of Magnetism

17.3. Types of Magnetism in Minerals

17.4. Magnetic Properties of Some Selected Commercial Minerals

17.5. Industrial Roll Design and Methods of Magnetic Separation of Minerals

17.6. Electrical Conductivity of Minerals

17.7. Electrostatic Forces and Mineral Separation

17.8. Practical Separation Units

Chapter 18: Flotation


18.1. Introduction

18.2. Flotation Reagents

18.3. Flotation Equipment

18.4. Flotation Circuits

18.5. Flotation Kinetics

18.6. Factors Affecting the Rate of Flotation

18.7. Application of Kinetic Equations

18.8. Other Flotation Models

18.9. Problems

Chapter 19: Metallurgical Process Assessment


19.1. Introduction

19.2. Analyses of Constituents

19.3. Definition of Terms

19.4. Material Balance

19.5. Circulating Load

19.6. Problems

Chapter 20: Process Control


20.1. Introduction

20.2. Controller Modes

20.3. Signals and Responses

20.4. Input and Output Signals of Controllers

20.5. Integration of Processes and Block Diagrams

20.6. Setting and Tuning Controls

20.7. Complex Advanced Controllers

20.8. Dead Time Compensation

20.9. Instrumentation and Hardware

20.10. Controls of Selected Mineral Processing Circuits

20.11. Advances in Process Control Systems

20.12. Expert Systems

20.13. Mechanics of Digital Process Control Systems

20.14. Problems





Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK

50 Hampshire Street, 5th Floor, Cambridge, MA 02139, USA

Copyright © 2016 Elsevier B.V. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).


Knowledge and best practice in this fi eld are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

ISBN: 978-0-444-63589-1

For information on all Elsevier publications visit our website at https://www.elsevier.com/

Typeset by Thomson Digital


Dedicated to all students interested in the Science and Technology of Mineral Processing

And especially to the memory of late dear Elmie Yan and also to Chitra Gupta for her patience and forbearance while preparing this manuscript.

Preface to the Second Edition

Demand for the first edition of this book prompted us to revise, enlarge and update it. The format of the second edition is in keeping with the first edition. The logic of presentation remains the same.

This edition includes additional chapters on stirrer mills and magnetic separation. The stirred mills for ultrafine grinding of minerals which were developed in the last three decades are described and discussed in detail. These are now being extensively used in large-scale operating circuits for enhanced recovery of minerals from their ore body. Magnetic, conductive and electromagnetic forces for the recovery of minerals with natural or induced magnetic, conducting and semi-conducting properties are a common beneficiation technique particularly in the mineral sands and iron ore industries. The basic elements of atomic theory that help to understand magnetic forces in minerals are explained and the designing of equipment as well as separation processes of magnetic from non-magnetic minerals based on these theories are described.

The objective of the book remains the same as that of the first edition which is to help students interested in processing minerals to economically liberate and concentrate them for down-stream extraction processes. To understand the unit as well as the integrated processes the subject has been treated somewhat mathematically with the view to apply them in actual process designs and operations. In so doing it is expected that the book would suit students from the disciplines of Metallurgy, Chemical Engineering, Process Engineering and to a limited extent Electronics Engineering, who are engaged in the beneficiation of minerals and who are at under-graduate, graduate and post-graduate levels of study. Some data provided in the appendix is expected to aid in calculations of designing and plant operations. Solutions to simple and common plant problems are provided.

In writing this second edition of the book we offer our thanks to the reviewers who offered help and guidance to improve on the first edition. Also our renewed thanks to Dr Lutz Elber and Dr Halit Eren for their help and contribution in writing Chapter 20 on Process Control. We would also like to thank the various state and university libraries who helped in supplying up-to-date information. We would especially like to thank our publishers who prompted us to write this edition; especially Dr Kostakitas, Anitha Sivaraj and Christine McElvenny who helped us to produce this edition.

Ashok Gupta

Denis Yan

Perth, Western Australia, November 2015

Symbols and Units

A general convention used in this text is to use a subscript to describe the state of the quantity, for example, S for solid, L for liquid, A for air, SL or P for slurry or pulp, M for mass and V for volume. A subscript in brackets generally refers to the stream, for example, (O) for overflow, (U) for underflow, (F) for feed, (C) for concentrate and (T) for tailing. There are a number of additions to this convention which are listed later.

Chapter 1

Mineral Sampling


Proper sampling of tectonic deposits of minerals is essential to assess its economic viability of extraction. The sampling techniques used to attain this objective are described in this chapter for varying conditions such as incremental and continuous analysis for particles of different sizes, shapes, densities and for precious metals. Statistical approaches of analytical data are given to interpret results as accurately as possible.


Random sampling

incremental sampling

continuous sampling

statistical sampling by Gy’s method

1.1. Introduction

A processing plant costs many millions of dollars to build and operate. The success of this expenditure relies on the assays of a few small samples. Decisions affecting millions of dollars are made on the basis of a small fraction of the bulk of the ore body. It is therefore very important that this small fraction is as representative as possible of the bulk material. Special care needs to be taken in any sampling regime, and a considerable effort in statistical analysis and sampling theory has gone into quantifying the procedures and precautions to be taken.

The final sampling regime adopted is, however, a compromise between what theory tells us should be done and the cost and difficulty in achieving this in practice.

1.2. Statistical Terminology

A measurement is considered to be accurate if the difference between the measured value and the true value falls within an acceptable margin. In most cases, however, the true value of the assay is unknown, so the confidence we have in the accuracy of the measured value is also unknown. We have to rely on statistical theory to minimise the systematic errors to increase our confidence in the measured value.

Checks can be put in place to differentiate between random variations and systematic errors as the cause of potential differences. A random error (or variation) on average, over a period of time, tends to zero, whereas integrated systematic errors result in a net positive or negative value (see Figure 1.1).

Figure 1.1   Representation of a Random and Systematic Error.

The bias is the difference between the true value and the average of a number of experimental values and hence is the same as the systematic error. The variance between repeated samples is a measure of precision or reproducibility. The difference between the mean of a series of repeated samples and the true value is a measure of accuracy (Figure 1.2).

Figure 1.2   Difference Between Precision and Accuracy.

A series of measurements can be precise but may not adequately represent the true value. Calibration procedures and check programs determine accuracy and repeat or replicate/duplicate measurements determine precision. If there is no bias in the sampling regime, the precision will be the same as the accuracy. Normal test results show that assays differ from sample to sample. For unbiased sampling procedures, these assay differences are not due to any procedural errors. Rather, the term ‘random variations’ more suitably describes the variability between primary sample increments within each sampling campaign.

Random variations are an intrinsic characteristic of a random process, whereas a systematic error or bias is a statistically significant difference between a measurement, or the mean of a series of measurements, and the unknown true value (Figure 1.1). Applied statistics plays an important role in defining the difference between random variations and systematic errors and in quantifying both.

1.2.1. Mean

The most important parameter for a population is its average value. In sampling and weighing the arithmetic mean and the weighted mean are most often used. Other measures for the average value of a series of measurements are the harmonic mean, and the geometric mean. Mode and median are measures of the central value of a distribution. The mode forms the peak of the frequency distribution, while the median divides the total number of measurements into two equal sets of data. If the frequency distribution is symmetrical, then its mean, mode and median coincide as shown in Figures 1.3 and 1.4.

Figure 1.3   Normal Distribution.

Figure 1.4   Asymmetric Distribution.

For a binomial sampling unit of mixed particles, the average percentage of mineral A is calculated by adding up all measurements and by dividing their sum by the number of measurements in each series.



 = sample mean (arithmetic)

xi = ith measurement

n = number of increments

The weighted percentage is calculated, either from the total number of particles in each series, or by multiplying each incremental percentage with the mass in each corresponding increment, and by dividing the sum of all products by the total mass for each series. However, the small error that is introduced by calculating the arithmetic mean rather than the weighted average is well within the precision of this sampling regime. The following formula is used to calculate the weighted average for a sample that consists of n primary increments:



Mi = mass of ith increment

M = mass of gross sample

Due to random variations in the mass of each primary increment the weighted average is a better estimate of v, the unknown true value, than the arithmetic mean.

1.2.2. Variance

The variance, and its derived parameters such as the standard deviation and the coefficient of variation, are the most important measures for variability between test results.

The term range may be used as a measure of variability.

Example 1.1

Consider a binary mixture of quartz and hematite particles with approximately 10% hematite. Samples are taken and the number of hematite particles is counted to obtain the percentage of hematite in the sample. Table 1.1 gives the result of 10 samples. For a binomial sampling unit the range is (maximum value – minimum value) = 12.6 − 5.7 = 6.9%.

Table 1.1

Sampling of a binary mixture with a binomial population (quartz and hematite).

If each series of measurements is placed in ascending order, then the range is numerically equal to xn x1 so that the range does not include information in increments x2, x3, …, xn−l. For a series of three or more measurements the range becomes progressively less efficient as a measure for variability as indicated in Figure 1.5.

Figure 1.5   Range of Experimental Values.

For two samples, the range is the only measure for precision but this is not sufficient to estimate the precision of a measurement process. The precision of a measurement process requires the mean of absolute values of a set of ranges calculated from a series of four or more simultaneous duplicates. This is the variance.

The classical formula for the calculation of the variance is



n = number of measurements

n − 1 = degrees of freedom

The standard deviation, σ, is the square root of the variance. The coefficient of variation (CV) is a measure of precision and is numerically equal to


Example 1.2

Variance values from a sampling procedure with a binary mixture of mineral particles are given in Table 1.2.

Table 1.2

Variance values from sampling a binary mixture of quartz and hematite.

The physical appearance of a sample that consists of 50 primary increments of 5 kg each is similar to that of a sample containing five increments of 50 kg, or to that of 250 kg of a bulk solid. However, the difference in intrinsic precision (as indicated by the variance) may be dramatic, particularly if the variability within the sampling unit is high.

, the sample mean as the best estimate for v, the unknown true value. If all increments are contained in a single gross sample, we have no information to estimate the precision of this sampling regime. If we want to know more about the precision of samples, systems and procedures, it is essential that duplicate or replicate measurements be made, from time to time, to determine the coefficient of variation for each step in the chain of measurement procedures.

1.2.3. Confidence Intervals

Other convenient measures for precision are confidence intervals (CI) and confidence ranges (CR). 95% confidence intervals and 95% confidence ranges may be used, although if concern over sampling precision is high, then 99% or 99.9% confidence limits must be considered.

That is, if we repeat a particular experiment 100 times, then 95 times out of 100 the results would fall within a certain bound about the mean and this bound is the 95% confidence interval. Similarly, confidence limits of 99% and 99.9% mean that 99 times out of 100, or 999 times out of 1000 measurements, would fall within a specified or known range.

In the draft Australian Standard, DR00223, for estimating the sampling precision in sampling of particulate materials, a confidence interval of 68% is chosen [1].

Fortunately, we do not need to repeat a measurement 100 times if we want to determine its 95% confidence interval, either for individual measurements or for their mean. Applied statistics provides techniques for the calculation of confidence intervals from a limited number of experiments. The variance between increments, or between measurements, is the essential parameter.

The most reliable estimate of σ² is var(x), the variance of the sample. The reliability of this estimate for σ² can be improved by collecting, preparing and measuring more primary increments, or by repeating a series of limited experiments on the same sampling unit.

The 95% confidence interval for a normal distribution is equal to ±1.96σ from the distribution mean. In practice, we often use the factor 2 instead of 1.96 to simplify calculations and precision statements. The 68% confidence interval is equal to ±0.99σ, for infinite degrees of freedom and if the number of replicate results exceeds 8 then a factor of 1.0 is an acceptable approximation.

1.3. Mineral Particles Differing in Size – Gy’s Method

Representing large bodies of minerals truly and accurately by a small sample that can be handled in a laboratory is a difficult task. The difficulties arise chiefly in ascertaining a proper sample size and in determining the degree of accuracy with which the sample represents the bulk sample.

In each case, the accuracy of the final sample would depend on the mathematical probability with which the sample represents the bulk material. The probability of true representation increases when incremental samples are taken while collecting from a stream, like a conveyor belt for solids and off pipes for liquids or slurries.

Several methods have been put forward to increase the probability of adequately representing the bulk minerals [2–5]. One such method involving both the size and accuracy of a sample taken for assay was developed by Gy and is widely used [6,7]. Gy introduced a model based on equiprobable sample spaces and proposed that if

dMAX = dimension of the largest particle

MMIN = minimum mass of sample required

σ² = variance of allowable sampling error in an assay (in the case of a normal distribution this equals the standard deviation)



where K is usually referred to as the sampling constant (kg/m³).

In mineralogical sampling the dimension of the largest piece (dMAX) can be taken as the screen aperture through which 90–95% of the material passes. As ±2σ represent the probability of events when 95 out of 100 assays would be within the true assay value, 2σ is the acceptable probability value of the sample. The sampling constant K is considered to be a function of the material characteristics and is expressed by



PS = particle shape factor (usually taken as 0.5 for spherical particles, 0.2 for gold ores)

PD = particle distribution factor (usually in the range 0.20–0.75 with higher values for narrower size distributions, usually taken as 0.25 and 0.50 when the material is closely sized)

PL = liberation factor (0 for homogeneous (unliberated) materials, 1 for heterogeneous (liberated) and see Table 1.3 for intermediate material)

m = mineralogical factor

Table 1.3

Liberation factors as a function of liberation size [8].

The mineralogical factor, m, has been defined as



α = the fractional average mineral content and

ρM and ρG = the specific gravity of the mineral and the gangue, respectively

The liberation factor, PL, is related to the top size dMAX and to the liberation size dL of the mineral in the sample space. It can be determined using Table 1.3. In practice, PL is seldom less than 0.1 and if the liberation size is unknown then it is safe to take PL as 1.0.

When a large amount of sample has been collected, it has to be split by a suitable method such as riffling. At each stage of subdivision, samples have to be collected, assayed and statistical errors determined. In such cases, the statistical error for the total sample will be the sum of the statistical errors during sampling (σS) and the statistical error in assay (σ) will be


When the sample is almost an infinite lot and where the proportion of mineral particles has been mixed with gangue and the particles are large enough to be counted, it may be easier to adopt the following procedure for determining σP.


PM = proportion of mineral particles

PG = proportion of gangue particles

N = number of particles

Then, the standard deviation of the proportion of mineral particles in the sample, σP, will be


The standard deviation on a mass basis (σM) can be written in terms of the percent mineral in the whole sample, provided the densities (ρ) are known. Thus if ρM and ρG are the densities of the mineral and gangue, then the mass percent of mineral in the entire sample, consisting of mineral and gangue (the assay), will be


assuming that the particles of mineral and gangue have the same shape and size.



Example 1.3

Regular samples of the feed were required at a copper-processing plant, having a copper content of about 9%. The confidence level of estimation was required to be 0.1% Cu at 2σ standard deviation. The liberation size of the Cu mineral (chalcocite) in the ore was determined to be 75 μm. The top size of the ore from the sampler was 2.5 cm. Determine the minimum mass of sample required to represent the ore, given that the density of chalcocite is 5600 kg/m³ and the density of the gangue is 2500 kg/m³.


From the data, dMAX = 2.5 cm.

Since the confidence interval required is equal to ±0.1% of a 9% assay,

Again from the data dMAX/dL = 25/0.075 = 333.33; hence from Table 1.3, PL = 0.05.

As the ore contains chalcocite (Cu2S) assaying about 9% Cu, it can be considered to contain [159.2/(63.56 × 2)] × 9 = 11.3% of Cu2S. Thus α = 0.113.

That is, the copper content of Cu2S is given by the ratio of atomic masses

The chalcocite content of the ore is then given by

The mineralogical composition factor, m, can now be calculated from Equation (1.7),

then from Equation (1.6)

Thus, the minimum sample size should be 131 kg.

Note the importance of not rounding off the numbers until the final result. If σ is rounded to 0.0055, then MMIN = 133 kg and if σ is rounded to 0.0056 then MMIN = 128 kg.

Example 1.4

A composite sample of galena and quartz was to be sampled such that the assay would be within 0.20% of the true assay, of say 5.5%, with a probability of 0.99, i.e. the sample assay would be 5.5% ± 0.20%, 99 times out of 100. Given that the densities of galena and quartz were 7400 kg/m³ and 2600 kg/m³, respectively, and the average particle size was 12.5 mm with a mass of 3.07 g, determine the size of the sample that would represent the composite.


Step 1

Determine σT in terms of N from Equation (1.11).


To determine N it is necessary to find σT

Step 2

To determine the value of σT to satisfy the deviation limits so that the area under the curve between the limits will be 99% of the total area, use Table 1.4.

Table 1.4

Probability P vs. deviation X, relative to unit standard deviation [9].

From the table, corresponding to a probability of 0.99, the value of the deviation from the standard unit is 2.576.


Substituting into Equation (1.12), N = (37.089/0.07764)² = 228,201.9.

Hence, the mass of sample necessary to give an assay within the range 5.5 ± 0.20%, 99 times out of 100 would be = 228,201.9 × 3.07 = 700,579 g ≈ 701 kg.

1.4. Mineral Particles of Different Density

Where variations in density of individual particles and their composition occur, the following considerations may be adopted to provide the sample size [9]:

1. Divide the material into n density fractions, ρ1, ρ2, ρ3, …, ρn varying from purely one mineral to the other, e.g. copper and quartz in a copper ore, and consider n size fractions, d1, d2, d3, …, dn.

2. Consider the mass percent of the ith size fraction and jth density fraction as Mij.

3. Consider Aij and Pij as the assay and the proportion of particles in the ith size and jth density fractions.

For sampling a mixture of two components (mineral and gangue) the proportion of particles in the ij fraction would be


And the assay of the mixture, Aij, will be approximately equal to



ρM = density of the mineral and

ρG = the density of the gangue

and the standard deviation of the ijth fraction will be


For a multi-component system the principles developed in Equations (1.13)–(1.15) may be extended. Their solution can be achieved easily by a computer. The general equation for overall sample assay is


Example 1.5

A nickel sulphide ore mineral (pentlandite) has an average particle size of 1 mm. It is separated into three fractions and the properties of each fraction are as follows:

The lot has to be sampled so that it would assay ±0.15% of the true assay having 5% with a probability of 99%.


Step 1.

As the stipulated probability is 0.99, Table 1.4 may be used.


Step 2.

Estimate Pij from Equation (1.13) in the following manner (Table 1.5).

Table 1.5

Calculation of proportion of particles in the ijth fraction.

Step 3.

With known P and σ, N

Now follow Example 1.4 to determine the sample size.

1.5. Incremental Sampling

Theoretically, unbiased gross samples can be obtained by collecting a sufficiently large number of single particles from a sampling unit. If each particle has a finite chance of being selected for a gross sample, and if this probability is only a function of its size, then this collection of particles will constitute an unbiased probability sample.

In the practice of sampling bulk solids, such a sample collection scheme is highly impractical. Therefore, we collect groups of particles for primary increments.

Sampling experiments that are based on the collection of a series of small and large increments from a large set of particles demonstrates that the precision of a single-increment sampling regime is a function of the number of particles in a primary increment. As a matter of fact, the formula for the variance of a binomial sampling unit, var(x) = N·p·q, shows the fundamental relationship between probabilities (p, q), the total number of particles in an increment (N), and thus its mass, and the precision for the parameter of interest. Hence, the precision of a single increment is essentially determined by its mass.

The binomial sampling experiment in which p = 0.095 and q = 0.905 (q = alternate binomial probability = 1 − p) can be used to set up a table in which the precision of a primary increment is given as a function of the average number of particles in an increment. In Table 1.6, the 95% confidence ranges (CR) that are calculated from corresponding 95% confidence intervals (CI) for the expected number of value particles in each increment, together with their coefficients of variation, are listed.

Table 1.6

95% confidence intervals and ranges [10].

Since these ranges are based on properties of the binomial distribution, their values are obviously independent of particle size. For a bulk solid with a certain density, its top size determines the mass of a primary increment. Top size is defined as the 95% passing sieve size.

The mass of an increment must be such that it is large enough to include the large particles, and particles present in the sample should be in the same proportions as in the lot being sampled. The minimum mass of the increment is, therefore, dependent on the size of the particles being sampled.

The top size of a bulk solid is a measure of length. The mass of a particle is a function of its volume and specific gravity and ultimately a function of its mean diameter or length. The mass of a primary increment then can be defined in terms of the number and mass of particles from the top size range. If no other information is available, then an acceptable rule of thumb is to collect primary increments with a mass equal to 1000 times the mass of a top size particle.



M = mass of primary increment in kg

M(t) = mass of top size particle in kg

Experience and theory are embodied in a number of national and international standards on sampling of particulate materials where the sampling regimes are defined in terms of the total number of increments, and the average mass of a primary increment.

It is generally accepted that a primary increment should contain no less than 1000 particles.

In the standard on sampling of iron ore (ISO TC102), the minimum mass for primary increments is specified in relation to the top size, and in Table 1.7 these recommendations are tabulated and compared to calculations for hard coal.

Table 1.7

Minimum mass for iron ore and coal primary increments [10].

The minimum mass for a primary increment should preferably be defined in terms of the volume of a particle from the top size range.

Once the mass for a primary increment is selected either in accordance with applicable standards or on the basis of the previous guidelines, or determined by the critical design parameters of a mechanical sampling system, the required number of primary increments remains to be defined. Too few increments will result in too low a precision, whereas too many would unnecessarily increase the costs for sampling and preparation.

Most standards for bulk solids contain simple formulae to calculate the required number of increments for a consignment from a given number for the unit quantity (usually 1000 t), or from tables that list the minimum numbers of primary increments as a function of the mass of a consignment. ISO TC 102 for iron ore specifies this number on the basis of the three levels of variability which presupposes some knowledge of the expected variability.

Table 1.8 is extracted from ISO 1988 on sampling of hard coal, and lists the number of primary increments for a 1000 t unit quantity.

Table 1.8

Number of increments to attain a precision of ±0.1 of the true ash [11].

Table 1.9 gives the mass and numbers of primary increments for hard coal, as specified in ASTM D 2234, for consignments of up to 1000 t, and for an estimated precision of ±10% of the ash content.

Table 1.9

Number and mass of increments (ASTM D 2234).

For consignments larger than this unit mass of 1000 t, ASTM D 2234 and ISO 1988 use the same formula to calculate the required number of primary increments:



n = required number of increments

ΣM = mass of consignment in tonnes

n(t) = tabulated number of increments

The overall standard deviation of sampling, sample preparation and assay is a function of the variability of the particulate material, the number and mass of the increments and the random errors associated with sample preparation and assay. It can be expressed as



σL = standard deviation of a primary increment (from Equation (1.3))

σPA = standard deviation of preparation and assay

In a well-balanced sampling regime, the variance of sampling and the variance of preparation should be of the same order of magnitude.

Dividing a consignment of bulk solids into lots, collecting, preparing and assaying samples from each lot, and reporting composite assays for the consignment, affects significantly the precision of the final result. The variance of preparation and analysis may easily become a limiting factor for the precision of a sampling regime.

If we solve n, the number of primary increments, from the simplified formula for a sampling regime of only one gross sample per consignment and a single measurement in the final analysis sample, it follows that [10]



var(c) = composition variance

var(d) = distribution variance

var(pa) = preparation and analysis variance

var(t) = total variance

The denominator of this formula may easily become a limiting factor for the total precision of sampling regimes since it shows that

Logically, the total variance cannot be smaller than the variance of preparation and analysis.

Example 1.6

Suppose that we have a consignment of 1000 t of iron ore with an assay of 65.0% Fe, and that we want to determine the assay with a precision of ±10%.

A precision of ±10% of 65.0% Fe is equivalent to ±6.5% of this iron content, and results in a standard deviation of 6.5/2 = 3.25 for a total variance of 3.25² = 10.56. Substitution of this total variance of 10.56, composition and distribution variances of, say, 1.66 and 0.71 and an average mass of 2 kg for primary increments into the formula for n, Equation (1.20), results in

In this case, a sampling regime of one primary increment of 2 kg would result in the required precision of ±6.5% of the iron content.

For a precision of ±1.0% instead of ±6.5%, how many increments must we collect to attain this precision?

For an iron assay precision of ±1.0% the standard deviation becomes 1/2 = 0.5 for a total variance of 0.5² = 0.25. Substitution into Equation (1.20) would result in a number of 7.5 (or rounded to 8) increments.

If a greater iron assay precision of ±0.25 % is required, then the denominator in the above equation becomes negative and it is not possible to obtain that degree of precision sampling from a single lot. The consignment must be divided into sub-lots in order to determine the assay with this precision. Table 1.10 shows a number of possible combinations.

Table 1.10

Precision of sampling regimes and number of increments required [10].

These results show that the assay of this particular type of ore cannot be determined with a precision of ±0.25% at affordable costs because the variance of preparation and analysis rapidly becomes a limiting factor for the total precision.

In terms of standard deviations, the number of increments may be calculated [1]:


When it is impossible to achieve the desired precision by testing a single gross sample from a lot (e.g. the number of increments is impossibly large), then it is necessary to divide the lot into a number of sub-lots, nS. Then


The total number of primary increments for the whole lot then becomes nS × n.

Example 1.7

Calculate the number of increments required in the sampling of silica in iron ore for a desired overall precision of 0.1, if the increment standard deviation is 1.5 and the preparation and assay standard deviation is 0.095.

If the lot is divided into, say, eight sub-lots, then

Therefore, 32 primary increments are required per sub-lot for a total of 8 × 32 = 256 primary increments in total.

When small amounts of sample are removed from a lot at random and combined to form a composite sample, the problem is to ascertain the incremental amounts and number of increments which would represent the bulk within specified limits. Each individual sample will have an assay equal to, more or less than, the true value. Assuming that a normal Gaussian distribution will be obtained if the assays were plotted against frequency, the probability that the assay will be within permissible limits will be given by


where x = deviation from the true assay.

Now if A is the average assay and A1, A2, A3, …, AN are the individual assays of the n sample increments, then


and the standard deviation of each increment sample will be


The standard deviation of the entire sample, σ, is related to σL by the relation


Substituting the values of σ and σL into Equation (1.23) and simplifying gives




The application of this method of estimating n is illustrated in Example 1.8.

Example 1.8

Ten samples were taken from a heap of copper ore each weighing 2 kg. The assays for copper show the following distribution:

The final sample should assay within ±5% of the true value with a confidence of 99%. Estimate the number of increments.


Step 1

Average assay = (A1 + A2 + A3 +  … + A10)/10 = 2.65

Standard deviation σL = 0.400, from Equation (1.25).

Step 2

For the probability P = 0.99 (given) and using Table 1.4 to simplify the solution of Equation (1.23),

h = 2.576

Step 3

Calculate the value of h in terms of n where K = 0.05, A = 2.65 and σL = 0.40, from Equation (1.28)


Step 4

Hence, 61 increments have to be taken whose weight would be 61 × 2 = 122 kg

1.6. Continuous Sampling of Streams

When mechanical samplers are employed, the samplers are designed to cut into and withdraw

Vous avez atteint la fin de cet aperçu. Inscrivez-vous pour en savoir plus !
Page 1 sur 1


Ce que les gens pensent de Mineral Processing Design and Operations

9 évaluations / 0 Avis
Qu'avez-vous pensé ?
Évaluation : 0 sur 5 étoiles

Avis des lecteurs