1 Introduction 

Mineral processing plants are complex systems. Their products need to meet certain ridged criteria, be it for use as construction material or for further processing as ore concentrate. 

The properties of a material entering the plant however may vary considerably. This is due to the location in the deposit it originates from as well as climate conditions. So one has to cope with varying particle size distributions (PSD’s), moisture content and mate-rial contaminations. 

During its way through the individual processing steps each piece of equipment will con-tribute to change the material towards the intended final product. Even though pro-cessing equipment is usually carefully selected and operated the outcome is effected by its wear condition and operating setup. 

With all this influencing parameters it takes a lot of experience to run a mineral pro-cessing plant and keep the product within specification limits. 

Plant simulation is a great help to achieve this goal. With NIAflow^® an entire processing plant can be simulated. Each piece of equipment is setup with its parameters and will apply the same effects on materials as in reality. Once the model is verified against the existing plant it is fairly easy to forecast certain changes on equipment or material to the overall outcome. So a reliable tool is provided to make production forecast, evaluate equipment changes or optimize the entire plant. For different operating modes of a plant a cost-benefit-analysis can be carried out to run the plant profitably. 

This article describes plant modelling with NIAflow^® and the use of its models in various scenarios. 

2 Modelling of machines and equipment 

2.1 General 

In order to simulate an entire processing plant relevant machines and equipment have to be modelled in the software. These are especially those machines that affect the mass flow or the parameters of the materials in any kind. Following typical processing machinery is grouped into classes by their effect on material. 

With each of the classes the effect on mass balance and or material composition in-creases. Transport machines like mining trucks do hardly effect the material. On the other hand sorting machines have a massive impact on basically all material properties. 

Aside from cumulating material according to its sorting property they also effect the par-ticle size distribution (PSD) as well as the tonnage of each of their products. 

2.2 Transportation 

For simulating a processing plant, transport machines play a lesser role. They do not affect the mass balance and aside from little changes in moisture content and/or tem-perature they have no effect on material composition.

However as material transport is a necessary evil these machines produce cost and have to be included in the accounting balance sheets. From the class of transport machines NIAflow^® currently calculates technical parame-ters of belt conveyers like required width and power.

For each transporting device maximum feed rates by volume and tonnage can be de-fined and errors will be raised if limits are exceeded. 

Transportability is also monitored to e.g. avoid transport of slurry on belt conveyers 

2.3 Storage 

Storage objects define the throughput through a plant or its sections by their output set-tings. Each storage object can keep any number of materials with different tonnages and/or PSD’s as ‘alternative products’. These can be assigned to ‘operating modes’ intended to run the plant model with different throughputs and PSD’s. 

For a cost-benefit-analysis storage objects play a special role. The price setup for the materials in a storage object will be used to calculate the revenue for the complete plant if the storage object is set to contain a sellable product. 

2.4 Distribution, blending 

Equipment of this group controls the material flow through the plant. Typical objects are splitters and reversible belt conveyers. The settings of these units can also be stored in operation modes. 

2.5 Comminution, agglomeration 

Machines in this group have a lasting effect on the PSD of their products. The results of various several processes can be corrected within limits. So out-of-spec products can be re-screened or blended. While coarse crusher products can be re-crushed this can-not be applied to product that is too fine. The product is final. 

Comminution machines like crushers and mills effect the material on a very characteris-tic manner. According to the method being applied in the machine (pressure, shear, im-pact) each fraction of the product will respond characteristically. 

As for the storage objects comminution devices can store different product PSD’s. The following figure shows a series of different PSD’s corresponding to the various closed-side-settings (CSS) of a cone crusher. 

One of these product PSD’S can be selected to represent the active product. All prod-ucts tan be assigned to various operation modes and activated automatically when the mode becomes active. 

2.6 Grading 

A large quantity of grading processes in mineral processing is being done by means of screening machines. One knows the screen at the end of a chain of processing steps to produce the final sellable products like road building aggregates or concrete sand. 

Common are screens in cooperation with crushers. Here the screen (in front of the crusher) separates the material already smaller than the crushers CSS. 

The range of cut sizes of screens reaches from 100 μm to heavy duty scalping opera-tions with openings of about 300 mm and particle sizes up to 1500 mm. 

Following figure shows a little selection of typical screens. 

Traditionally a screen (here vibrating screen) is calculated using a specific screening capacity retrieved from laboratory values. The feed material is seen as fixed input. Sev-eral correcting factors are used to fine tune these numbers to current conditions. The result is the screening surface required. Products are estimated based on cut and usu-ally assuming a 100% efficiency. 

NIAflow^® uses a slightly different approach. The required surface is just an intermediate number and being used (along with other parameters) to compute an internal cut func-tion. 

Than the PSD of the feed material is fragmented into small fractions. Each of those frac-tions then will be evaluated according to its probability to arrive in the oversize product or in the fines. At the end of this process PSD’s are cumulated. So this procedure rather calculates products then machines. The machine with its selected surface and setup is used as input data. 

This approach delivers reliable product data including percentages of misplaced parti-cles. When fine tuning machinery settings like surface area and media setup NIAflow^® updates results instantly, so one can evaluate if products remain within specification limits. 

Below figure shows the calculated products of a screen where the 5/8 product is out of spec due to high content of oversized particles. 

For calculation and monitoring of the results following data is being used: 

• Feed material 
  • PSD 
    • Oversize content 
    • Content half cut 
    • Maximum particle size 
  • Density 
    • Specific gravity 
    • Bulk density 
• Screen media
  • Opening shape
  • Opening direction
  • Media type
  • Media thickness
  • Open area
• Process
  • Required efficiency
  • Effects of wet screening
• Machine
  • Inclination
  • Screen aid
    • Banana deck
    • Ball tray
    • Ultrasonic device
  • Layer on deck 

2.7 Sorting 

Sorting processes are very complex. In each fraction of the feed material sorting proper-ties can be distributed differently. Therefor each PSD-fraction is calculated independent-ly. As the result of a sorting process the sorting property will accumulate in one of the products while the other is displaying significantly lower contents. Should sorting proper-ties be distributed unevenly within the fractions also a change of the PSD will result. 

Below figure shows a very simplified color sorting process. It is assumed that each frac-tion of the feed material has the same even color distribution (e.g. from white = 0 to red = 1). 

The user has to define a cut function which then is applied against the color distribution of each fraction of the PSD. Resulting one can find a light colored product (left) and a heavily colored product (right). 

3 Plant modelling 

3.1 Definition of a plant 

3.1.1 Operating schedule 

Each NIAflow^® project may contain any number of plants which themselves consist of objects connected to each other by lines representing the material flow. Plants are de-fined by a number of parameters where the most important is the operating schedule. Defining 

• Working hours per day 
• Work days per week 
• Plant availability 
• Shut down days 

provides the basis for calculating annual production as well as for running the cost-benefit-analysis. Plant specific properties are furthermore cost for infrastructure, climate conditions or paint specification for machines and equipment. 

3.1.2 Operating modes 

Function ‚Operating modes‘ allows for recording certain plant conditions that then can be re-applied by the push of a button. Following settings can be stored within an operat-ing mode: 

• Storage objects 
  • Output tonnage 
  • PSD of active product 
• Transport objects 
  • Splitter settings 
  • Active output of reversible belt conveyer 
• Comminution, agglomeration 
  • Active product PSD 

3.2 Example project: ‚John Doe Aggregates‘ 

Following figure shows a typical processing plant in three stages. Objects, material and other data do not refer to any existing processing plant.

Connection lines between machines and equipment define the material flow. During calculation runs products and/or machinery settings for each object are being recalcu-lated. During calculation all defined limits are monitored. 

Closed circuits are calculated until the set precision or the number of iterations is reached. 

3.3 Plant optimization 

3.3.1 Model verification

With NIAflow^® one can build models of very complex existing plants. To use these mod-els e.g. to analyze production or for plant optimization it is necessary to verify the model against the real conditions on site. 

For very complex plants a simplified model may be created that is focused on the main target of the project and does not contain machines and equipment with neglectable effects on the materials (e.g. transport units). 

Now one operation mode is being sampled on site and compared against the results of the model. The model is verified when its simulated tonnages and material data includ-ing particle size distributions (PSD) meet the results of the sampling to an acceptable degree. Additionally the model must respond to a change of input parameters (e.g. feed tonnage) the same way the existing plant does. This task is best carried out on site along with the personnel of the plant operator. 

3.3.2 Sampling 

According to requested precision and project target 4 groups of sampling points can be defined (figure).

 Essential sampling points are marked in red. These objects define or modify product PSD’s. In order to verify a model samples have to be taken from the positions in green. Sampling at these positions does not increase the overall workload as they are usually included in quality management systems. 

Additional sampling on collecting belts help to improve the model precision. 

3.4 Optimization

Once the model is verified it can used to simulate certain operating conditions of the plant. These operation modes have to be defined by its machinery setup (tonnages of storage objects, setup of splitters, crusher products). In the model the plant throughput can be increased to a point where the first bottleneck is reached. This is the case when a product is out-of-spec or certain machine min-max-limits are reached. 

After eliminating these first bottlenecks by adjustments on the machines or the process layout the next set of bottlenecks can be found. This procedure is being repeated until the optimization target is met or the plant capacity is reached. 

The result of the optimization is a set of measures to modify the real plant to produce the same results as the model. According to the complexity and required efforts these measures are grouped in three levels: 

1. Regular adjustments like change of feed rate or splitter setup 
2. Adjustments of machinery parameters like closed side setting on crushers or media change on screens
3. Change out of machines, new processing technology, modification of process layout 

4 Cost-benefit-analysis 

4.1 Data collection 

At the end of all efforts a processing plant is a means to generate profit. Processes in a plant as well as allocated cost of the individual machines are hard to collect manually. NIAflow^® combines generated cost with the flow of material through the plant thus creat-ing the basis for cost-benefit-analysis. 

For the following analysis three operating modes for ‘John Doe Aggregates’ have been defined. All cost on plant and machine level had been defined as well as the marked prices of the products. 

Plant level data: 

• Currency 
• Cost of infrastructure incl. depreciation timeframe 
• Annual running cost 
• Energy- and diesel cost per consumption unit 

Machinery level data: 

• Purchase price incl. depreciation timeframe 
• Electrical consumption 
• Diesel consumption per operating hour 
• Supplies consumption per operating hour 
• Wear cost per ton throughput 
• Maintenance cost per service cycle 

To evaluate all cost against revenue all sellable product must be defined with its market value. NIAflow^® considers a material as product if: 

• It is leaving the plant or 
• It is ending in a storage object that does not have further output connections 

Product becomes a sales product if it carries a price. Additionally it can be set up whether or not the product is being sold wet. 

4.2 Operating modes 

4.2.1 General 

In reality the switch from one operating mode to another is usually carried out by means of setting splitters and reversible belt conveyers and/or change the raw material input in the plant. NIAflow^® offers the same options. Following three operating modes are de-fined that are to be evaluated concerning their commercial results.

4.2.2 Mode ‚Aggregates‘ 

It is assumed that the raw material is of good quality and can be used to produce road building aggregates exclusively. All production of sellable product is carried out in the tertiary section of the plant. Stockpiles P.01 and P.02 (see Figure 5: Figure 5) do not have production. The only non-sellable product is contaminated fines from the primary scalper screen. 

In this operating mode a raw feed into the plant of 260 t/h is possible before product 11/16 develops too much carry over. 

4.2.3 Mode ‚Recrush‘ 

For this mode it is assumed that there is no market for product 11/16 and 16/22. There-for these ready to sell products have to be recrushed into finer product specifications. As in the previous mode only the tertiary section of the plant produces. 

Raw feed can be raised to approximately 215 t/h before the following limits are reached: 

• Product 2/5 with too much carry over 
• VSI at throughput limit 
• Products 11/16 and 16/22 out of spec. This however is neglectable as they are being recrushed. 

Limits reached or exceeded can be visualized using the PSD label option of NIAflow^®. To use this feature specification limits and/or curves have to be defined in the target storage objects of the materials. On the flowsheet the PSD can be shown along with the material percentages at certain limits. Limits that are exceeded are shown in orange. 

The figure below is to illustrate the approach as an example for all operating modes.

4.2.4 Mode ‚Mixed‘ 

All splitters are set to 50/50. Maximum raw feed is 290 t/h. The first limit reached is again carry over in product 11/16. This mode comes with the highest raw feed and does not utilize the crusher to 100%. Theoretically the feed could be increased further after removing the bottleneck at cut 11,2mm. 

4.3 Evaluation of operating modes 

Following table summarizes the results of the analysis for the three operating modes.

Large differences become visible at production rates. Operating mode ‘Aggregates’ shows a mid-range production level. With its lowest wear cost and best revenue it ar-rives at the best ROI and reaches break-even after 4 years already. 

The most commercially unfavorable mode is ‘Recrush’. This doesn’t come as a surprise as finished product is converted into finished product again with all the associated con-sumption. Cost per ton is at 10,6 € and significantly higher than in any of the other modes. Accordingly return-on-investment is much lower and break-even can only be reached after 8,5 years. 

Mixed mode has the lowest cost per ton at 9,63 EUR/t. However as the average sales price is also the lowest its position in the ranking of the three modes is a good average. 

5 Summary

This article described the modelling of machines and equipment of plants of the mineral processing industry using the simulation software NIAflow^®. Models can be used to evaluate production or to optimize entire plants. With the function ‘Cost-benefit-analysis’ processing and commercial data can be combined to create profitable structures and processes. 

The general procedure was demonstrated using a fictive plant model. Three operating modes had been defined and analyzed. 

Provided good input data is available for modelling, a NIAflow^® project will behave iden-tically to the real plant. It therefor provides an excellent tool for planning and operating mineral processing plants.