Optimization by computer - Optimization of Mineral Processing Plants with Simulation Software “NIAflow”

Summary: Years can pass from the first feasibility study through the project phase to the start-up of a mineral processing plant. During this time the originally defined technical and commercial framework could have changed significantly. It can be difficult to meet all defined parameters when putting a plant into operation. But even in the case of a successful startup the market can have changed. For the optimization of entire processing plants a simulation of the plant using software products like NIAflow® is advisable. During this simulation the process of the plant is modelled and evaluated in the software. Every piece of equipment is represented by an object in NIAflow®, the material flow through these objects is visualized and calculated.

1 Introduction

Years can pass from the first feasibility study through the project phase to the start-up of a mineral processing plant. During this time the originally defined technical and commercial framework could have changed significantly. It can be difficult to meet all defined parameters when putting a plant into operation. But even in the case of a successful startup the market can have changed. The defined products may not be in demand as before and the return of investment may be in danger. In general, any processing plant needs to be adjusted or optimized to varying situations.

Typical optimization targets include:

Increase of production

Improvement of plant availability

Adjustment of products to meet market demands

Products to meet specification

Reduction of circulated load

Reduction of energy- or water consumption

Producing new products by blending

While some of these targets can be achieved by improving the existing machinery, it is of utmost importance to look at the entire process layout. It is a waste of time and money to increase the capacity of one piece of machinery if that overloads an entire section of the plant.

For the optimization of entire processing plants a simulation of the plant using software products like NIAflow® is advisable. During this simulation the process of the plant is modelled and evaluated in the software. Every piece of equipment is represented by an object in NIAflow®, the material flow through these objects is visualized and calculated. The model and its results is then compared with the situation in the field and adjusted accordingly.

Once verified the model can be used to simulate production scenarios, look for bottlenecks and create solutions for certain optimization targets. This article describes the typical procedure of this type of optimization task.

2 Optimization steps

2.1 Modelling of the current situation, definition of the target

The first phase of a plant optimization project is the most demanding. Depending on the complexity and age of the plant and the quality of its on-site documentation the time and effort of data collection can vary greatly. It is of utmost importance to collect all information regarding the current process, the material, machinery and equipment. Aside from discussing the general procedure, in this first phase it is necessary to define how to collect the required information and what is the measurable target of the entire project.

After data collection the entire plant with its machines and materials and its connections is visualized in the simulation software NIAflow®. Using this model, the flow of material through the plant is calculated and displayed. Also an equipment list is produced as well as datasheets for each piece of equipment with detailed information according to the technical parameters and the materials being processed.

2.2 Verification

In this phase the model is tuned to real life conditions. With very complex plants a simplified model may be created. This simplified model focusses on the optimization target and the machinery contributing to this. Mainly transporting machines and equipment that do not modify materials can be removed from the model and replaced by simple connection lines.

Now an 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 including particle size distributions (PSD) meet the results of the sampling to an acceptable degree. This task is best carried out on site along with the personnel of the plant operator. It is also the best time to finalize a detailed optimization target.

2.3 Optimization

Once the model is verified it can be 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 to the machines or the process layout, the next set of bottlenecks can be found. This procedure is 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 into 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. Replacement of machines, new processing technology, modification of process layout

3 Modelling

3.1 Process layout

In the following, a plant optimization project is introduced based on the example of a fictive gravel plant for road building materials. The plant is supposed to consist of three stages.

In the primary crushing plant the run of mine (ROM) material is pre-screened and crushed down to <180 mm by means of a jaw crusher. Prior to crushing, material <20 mm is discarded due to contamination. The PSD of the crusher product is calculated by NIAflow® based on a closed side setting (CSS) of 120 mm (Fig. 1).

In the secondary crushing phase, the overflow of the vibrating screen is fed to the secondary cone crusher. Screen and cone work are in closed circuit. Products 22/32 and 32/45 may be stockpiled or sent to the tertiary crushing plant by means of belt conveyer 11.00. Material <22 is sent directly to production screen (15.00).

In the tertiary crushing plant aggregates for road building are manufactured. The overflow of  ‘Product Screen A’ are in closed circuit with the tertiary vertical shaft impact crusher (VSI). Overflow of the middle and bottom deck may be stockpiled or recirculated to the VSI. Finally, ‘Product Screen B’ (20.00) produces the remaining specifications.

3.2 Particle Shape

In this project the particle shape is also evaluated. With particle shape mode activated, NIAflow® requires the particle shape of each fraction of a PSD and not only within the limits of the specification. Missing values outside of the specification are assumed to be 100 % cubical. Mathematically this approach is identical to measuring 300 individual particles inside of specification limits and ignoring the outsize particles. To define particle shape in NIAflow® two classes are created representing the width to length ratio of 0,16-0,33 and 0,33-1. The first class contains elongated particles the second cubical particles (Fig. 2).

3.3 Operation modes

Under real production conditions, a processing plant needs to be tuned to current market conditions. Aside from throughput, the material flow through the plant must be controlled and modified. As a result, several operation modes arise. To work with operation modes NIAflow® collects data concerning splitter settings, throughput through storage objects like silos and stockpiles and the active crusher product. In the following 3 operation modes are defined.

3.4 Crusher products

Depending on the feed rate, the product PSD of the VSI will shift. This has to be accounted for when running different scenarios during optimization. The reduction ratio of a vertical shaft impact crusher is usually bigger with low feed rates. During modelling field data will be used to set up alternative crusher products (Fig. 3).

The PSD ‘Normal Operation’ will be used for throughput rates between 125 and 275 t/h. Feed rates above or below are modelled with one of the other curves respectively.

3.5 Screen setup

Vibrating screens control product quality and material flow significantly, which puts them in focus for most plant optimization projects. In this example product screen A and B of the tertiary screening plant will be investigated more closely (Table 3).

3.6 Target Definition

For this example the following targets shall apply:

Maximize throughput in a mixed production of the three operation modes

Maintain particle shape requirements for road building ­materials

Avoid capital investments

4 Verification

4.1 Relevant measuring points – input data

In this article we are dealing with a virtual plant. Model verification therefore is obsolete. However, the general procedure will be discussed.

The properties of material-changing machinery play a major role in a real processing plant but also when calculating the NIAflow® model. Therefore, relevant data have to be carefully compared against reality. Especially screening machines have a lot of relevant properties that effect how material is being processed. NIAflow® uses all data entered to automatically create a cut-curve for each cut size of the screen which actually splits the feed material into fine and coarse material.

After modelling an operation mode is defined from which to take samples. Ideally this operation mode is the most common and easy to sample. During sampling the focus should be on the following relevant sampling points:

1. All raw material-entry points in the plant,

2. Crusher and mill output,

3. Product bins and stockpiles

4. Products of material-changing machinery, e.g. screens

It is usually impossible to sample all points of interest. Therefore, some areas have to be looked at like black boxes.

4.2 Model calibration

After sampling, the NIAflow® model is calculated using the same machinery setup. Now PSDs and tonnages from the sampling results are compared to the calculated values.

The theoretical NIAflow® model can be fine-tuned by adjusting the calculated cut-curves in position and slope. In doing so, the production performance of the individual machine can be matched by the NIAflow® calculation. Usually a range of 85  % to 105  % is used to adjust the position of the cut-curve. For values smaller than 100  % the real separation is finer than in the model prior to calibration.

The cut–curve increase is usually adjusted in the range from 80  % to 110  %. Smaller values here indicate that the screening efficiency in reality is lower. Should an adjustment of less than 80  % become necessary, it is an indicator of an improperly operating screen machine.

After calibration is completed, calculated products will match real products closely. Alterations to machines or the process layout performed in NIAflow® will now yield the same results as those being performed in the plant operation. If a good compliance can be shown especially at sampling points of group 1 to 3 the model is considered verified.

4.3 Working with the verified model

The NIAflow® model is used to simulate the operation modes of the plant and to forecast the products. While targeting the objectives defined earlier the model is adjusted until bottlenecks are reached. This has to be done with all of the defined operation modes as not every mode will produce the same results. After removal of the first bottlenecks the process is repeated until a final plant limit is reached.

At the end of the process, a list of measures to be taken is produced outlining each measure and the expected benefit.

5 Optimization

5.1 General

It has to be evaluated in which throughput range the plant can operate sufficiently. Aside from the obvious maximum throughput also a minimum value has to be considered. The latter is less interesting however and defines a limit under which the product will run out of spec. Within the scope of this article only maximum rates will be looked at.

5.2 Limits

Maximum throughput for the secondary cone crusher will be defined at 500 t/h while the tertiary VSI has a limit of 400 t/h. The primary crushing plant is not in the scope of the optimization. Furthermore, lower limits for product PSD’s (Table 4) and particle shape are relevant.

5.3 Optimization – particle shape

It quickly becomes obvious that the particle shape requirements are not met in any of the operation modes. Even in the mode with maximum re-crush, which should produce the best shape, percentages are below 80 %. The two other modes are significantly below that mark.

Prior to further steps, the layout of the process has to be adjusted to meet the particle-shape-condition. Hereby material -22, even it is a finished size product, is now fed to the tertiary crusher instead of directly onto ‘Product Screen A’. With this adjustment the advantage of saving on wear and tear in the crusher is gone but the particle-shape-condition is met in all operation modes. Fig. 5 shows “Normal Operation”.

5.4 Optimization – maximum throughput

In this calculation the maximum capacity of the model while adhering to all limiting conditions is calculated. Table 5 presents the results of the individual measures. The value “Limiting Factor” represents the currently reached bottleneck.

Already at fairly low throughputs the content of material <5.6 mm in product 5/8 increases to 15  % and marks a bottleneck. Adjusting the screen opening of the middle deck of “Production Screen B” from 6 to 6.5 mm eliminates this bottleneck and allows for a production increase in the mix of all three operation modes of 50  %.

A further production increase leads to an out-of-spec increase of -11.2 mm in product 11/16. Eliminating this issue would require increasing the opening however that would result in too many >11.2 mm particles in product 8/11. Improving the screening efficiency can only be reached by changing the media from polyurethane to woven wire cloth. The open area is significantly larger and allows for an increase in production of 14  %.

It is a safe indication to come closer to the overall plant capacity when several bottlenecks are reached at the same time. Now cuts 2 mm, 8 mm and 16 mm come into focus. Additionally the overall capacity of the secondary crusher is reached. While the bottleneck at 2 mm cut can be resolved by adjusting the opening, cuts 8 mm and 16 mm must be converted to wire cloth. These measures result in a production increase of another 12  %.

Increasing the total plant throughput further produces a bottleneck at 11.2 mm. The maximum gain would be 4  %. Due to the fact that this cut size has been optimized already and the secondary cone crusher is at its limit, the maximum plant capacity is considered reached and the optimization process is completed.

Fig. 6 shows the tertiary crushing and screening plant in operation mode “Normal Production”. Maximum production is 305 t/h including 75.7 t/h from product 32/45 and product 22/32 in the secondary crushing and screening plant.

6 Summary

Using a virtual crushing and screening plant for road construction materials the general approach to a plant optimization has been presented. In this example limiting factors were the particle shape and the product PSD’s. During the individual optimization measures the plant capacity could be gradually improved up to a point where new machines would have to be purchased.

With the simulation software NIAflow®, several optimization measures can be applied and their effect on the overall plant can be evaluated. One is not limited to target values like PSD and particle shape. There could also be scenarios in which the reduction of circulating load or meeting certain material properties is the objective of a plant optimization. Provided there are good field data for modelling and calibration of the NIAflow® model, it will perform closely to the real plant. It therefore provides an ideal planning tool for the operators of mineral processing plants.