A low level sensor/detection on the feed hopper is required and the mass meter is an absolute requirement for the feed hopper.

A flowmeter on the dosing pumps is not a good to have, it is a must for proper and efficient dosage control.

Multiple compartment (3) steel tank system or three individual baffled tanks are required for mixing, hydration and dosing respectively. In a compact design these can be interchangeable for versatility. A mixing tank + a dosing tank only would not suffice - for the industrial scale application that is. It is imperative to have a hydrating tank for effective flocculation downstream. Compromised hydration renders the flocculants ineffective and can result in high reagent consumption and poor flocculation. Engage your reagent supplier to advice or the Process Engineer’s previous experience with the application.

The quantity and strength of flocculants required for delivery to the process is used to determine the size/model of the pumps and the number of units required. Emergency redundancy capability must be built in and this depends on the plant availabilities, and downstream process requirements. A 15-25% factor is widely accepted. Mono pumps are ideal for low pressure transfer duties and the dosing applications. They provide a smooth pumping action which ensures the integrity of the shear sensitive polymer-based flocculants.

The dosage rate requirement is determined by settling tests and test work results provided by the reagent supplier. Working in partnership with the reagent supplier has proven to yield much better results and optimum design applicable to a given process

Individual pump volumetric monitoring and dilution control system for dosing pumps is necessary and a fully automated control interfaced with thickener dynamic process controls (clarometer) preferably using a PLC's and SCADA system for optimum control and benefit..

The main advantage of automatic flocculent control is the considerable reduction in flocculent usage and significant contribution to thickener stability by assisting in the reduction of bed level and bed density variations. I dealt with flocculent plants more than a decade & a half ago and I trust that the principles still apply today.

Interesting points you mentioned there but I would like to focus on point 1 a bit more.

I read literature that detailed that several technologies used in low-level detection include:

1. Weight & Cable (plumb-bob, yo-yo, etc.)

2. Ultrasonic

3. Guided Wave Radar

4. Load Cells

5. Strain Gages

With that said, what works best for flocculants in powder form, bearing in mind the cost implication. Do you know of any specific product that you have experience with?

What works best for flocculants powders are load cells. If the facility is well protected and taken care of, the units are virtually maintenance free. They provide the highest accuracy and reliability in the industry. A scheduled calibration programme is necessary for process assurance and overall system integrity.

Check the following website, for assistance and further direction. You can request the type of load cells and model used in their plants

http:/http://www.senmin.co.za/

They handle both brown or greenfields projects and provide first line maintenance of the units. They provided us with the bulk storage flocculant facility.

I am curious that you think that floc flow meters are required. Do you have experience with mono-pumps or other progressive cavity pumps ever drifting in terms of the relationship between flow rate and pump speed? Similarly secondary dilution can be regulated at a suitable average point using a mechanical regulator such as a maric valve.

I agree with most of what you said, especially the requirement of good floc control to maintain a steady bed-level.

Note that schenck process can be used for a price check on weighing technology (both of the batch dose and the remaining floc in storage). I am unsure of their level of support in SSA.

We have Mono pumps here, C62M and C81M. With our setup, we can get up to 15m3/hr out at a 90% setting on the VSD.

The challenge with flowmeters is that the normal Doppler will not work well because there are no solid particles in the floc and depending on where it is installed, there may not be air bubbles. I think a MASSFLO meter would probably work but I am yet to install one

I know this is plant specific and depends on the wattage on the motor we have as well as the static/dynamic heads but to sum it up we pump the floc ~60m downstream.

Isn't the volume-metric flow directly proportional to the VSD speed?

I Have developed innovative lime dosing system, with 2m3 tank, hopper, domestic water pump, mechanical valve for dosing. It is 100% effective and in operation for the past 10 years. Similar type was installed 2nd time in another mine working 100% effectively. Simple and easy. For IRON ORE PROCESSING AND SETTLING TAILING IN THICKNET 1500 M3/hr water treatment and settling solids at 150 tph tailing, in 750 tph process plant.

Each dosing line is fitted with the Magnaflow Flowmeter - these are magnetic type flowmeters, they are reliable and with high level of accuracy for this application. The flowrate measurements are used to calculate the effective mass of flocculant used per stream - for accounting and performance monitoring purposes. The VSD on the other end is used in the PID control coupled with the clarometer for effective control and optimal dosing.

I have experience with the mono pumps and they have worked very well for the flocculant dosing application. There isn't so much a drift in the flowrate and the pump speed (they are directly proportional) All you need is to take good care of the units and reliability & precision are guaranteed

As for secondary dilution - the multi-stage dosing system along the feed launder through to the centre feed well where further dilution is provided proved to be most effective.

If the flow rate from the pump is directly proportional to the pump speed, why do you need to buy an maintain additional flow meters? Why can't you simply use the pump speed to indicate the flow rate (after calibration)?

Because you'd want to see when your life of pump is getting depleted. A HAZOP study on the plant would show the necessity of a flowmeter. It’s still something you can run without.

The design of a flocculent plant is well known and reinventing the wheel is counterproductive. Go buy one from a reputable supplier. (Senmin was good advice, but they would try and bind you into a supply contract for free supply and maintenance of the floc plant, but you most probably already know this.)

Lime dosing can be used when clay is present in the ore. The lime act as a coagulant that could increase the underflow density and improve over flow clarity, but you would still need flocculent for the thickener.

I still need to see the site where a clarometer is in operation after the thickener was commissioned and in operation for even a few months.

In theory the floc-dosing could be controlled through only the VSD changing the speed of the mono pump, but in practice this control has proven to be problematic.

The distance of the flocculent plant from the thickener and the flocculent flow rate to the thickener is more important than trying to try and do optimized automatic flocculent control.

Flocculent dilution is often used to reduce the effect of too high flow rate in the flocculent line, but is mostly not appreciated as an action / reaction effect.

Thickener control is a complex, much research has gone into this challenge, but it has not been completely solved for thickeners subject to feed variations and/or clay in the feed. However if you have a thickener with a stable feed and very little clay it is almost good enough to have the operator check the thickener once a shift - which is more often than not the thickener control applied in industry. The only proven reliable instrument on the thickener is a pressure transmitter for the bed mass.

Thickeners are more fascinating and challenging than what process people in general believe.

I had had a clarometer working for 4 years and it was running straight to the floc pump. Having said that it needed weekly attention from the met tech even in a pretty benign environment (sulphide float tails). Also, the control wasn't ideal due to problems inherent with sampling for a clarometer. But even with the cycling the consumption was 30% less than when left in manual. That thickener was run with a very high and stable bed mass (those loops worked well) and at about 130% of design throughput (it is a supaflo HRT thickener -- "floc miser" feed well). Any potential "pump wear" of a mono pump running on floc would be slow enough for the feedback loop to deal with.

I agree 100% that maintaining the inventory as indicated by bedmass (pressure gauge) is the best way to control a thickener. There are some details to actually doing this, but they were worked out in the 70's -- well documented by F.G. Shinksey for other applications. Stopping the thickener ever seeing underflow pumping disturbances is crucial -- this is where you need to put your flow meters (albeit more expensive ones).

Secondary dilution is best achieved with a floc box on the bridge -- these are made by Outotec so I'll leave the sales pitch out of this discussion!

Back to the original topic, you need to make a decision with floc plants between volume measurement weight measurements of the primary dose of powdered floc. The extra expense of actual weight measurement might be easy to justify if you expect poor maintenance practices or have a humid environment to deal with.