Automation irons out phosphate compliance

Installing an automated feed-forward monitoring system has reduced dosing costs and improved phosphate discharge compliance at Keynsham water treatment works, writes Angus Fosten of Partech


Keynsham sewage treatment works between Bath and Bristol serves a population of around 22,000, so in the general scheme of things it is not a particularly large plant.

In recent years, Wessex Water has invested in various systems at Keynsham in order to automate the operation of plant and to bring it up to unmanned status. One aspect of the treatment process that has been the subject of much attention has been the automatic control of ferric sulphate dosing required to keep the level of phosphate discharge within the Environment Agency’s consent level of 2mg/l total phosphate for the works.

For many wastewater treatment companies, natural removal both of nitrate and phosphate is, where applicable, the preferred method. But in many cases this is not possible and it becomes necessary for the works operator to resort to chemicals, with ferric chloride and ferrous sulphate being the most commonly used.

Where the efficiency of iron is much reduced, aluminium salts have been used successfully, normally in conjunction with an iron salt. In all cases, the amount of chemical used is critical for the performance of the works, cost control and meeting the metal discharge consent.

The removal of phosphate from the effluent of a sewage works is an essential part of the Urban Wastewater Directive. Many treatment works have phosphorous discharge consents, and there is anecdotal evidence that effluent containing phosphate is making the challenge of staying within consent levels more of a problem.

Dosing iron salts at the front end of a works requires a level of control to ensure that the pH influent is not made too acidic, as this has a detrimental effect on the nitrifying process. Historically, the iron dosing rate has been calculated by taking a series of samples to establish an iron:phosphorus dosing ratio, then a known volume of iron is applied accordingly, dependent on the incoming flow to the works.

At Keynsham, ensuring compliance is achieved using ferric chloride. In keeping with many other sewage treatment plants, Wessex Water has seen the costs of ferric salts increase substantially in recent years. As a result, the company has set out to find a means of reducing ferric sulphate costs while maintaining discharge consent levels.

For a number of years, ferric dosing at Keynsham was controlled using a flow-proportional control unit. But long-term monitoring of the system revealed that, during the winter months, the volume of ferric sulphate dosed would increase due to the rise in water flow through the plant. Clearly, this was not dosing in line with the phosphate load. Typical ferric sulphate use throughout the year was between 250tonnes and 300tonnes.

“We believe that we are very good at monitoring our use of ferric,” says Wessex Water’s Andy Preece. Historically, the company had monitored consumption on a monthly basis by measuring the holding tank levels. This enabled it to calculate ferric usage in terms of tonnes per year, and how much it was using in terms of iron milligrams per litre dosing. “From there, we can work out the iron to phosphate ratios, enabling us to calculate ideally what we are dosing,” says Preece.

At Keynsham, a phosphorus removal system was first introduced back in 2003.

Wessex Water commenced initial development work with Partech on a prototype controller in 2005, and initial results gave useful and repeatable information on the incoming phosphorus in the crude sewage. In addition, there was sufficient evidence from this prototype that it could improve the dosing control of the ferric salts by relating to phosphorus concentration and flow (or load) for Wessex Water to look into this method further.

The sampling system

The Partech feed-forward system comprises a sample pump, filter unit, filtrate pump and sample pot, dosing control unit and MicroMac C phosphate monitor. The sample pump, which has a flow rate of 2,700l/min, contains an impeller designed for use in raw sewage and wastewater-containing solids up to 7mm.

The pump is suspended in the intake channel and linked to the main kiosk. The self-cleaning filter unit comes with a 20µm stainless-steel wedge wire filter element in a specially designed PVC body, incorporating compressed-air inlet, sample inlet/outlet, filtrate outlet and cleaning port. The self-cleaning filter is supplied with a control system linked to the MicroMac C. This instigates a sample-on-demand cycle that controls the sample pump and the filtrate pump in an operator/site programmable manner. As part of the cycle the air clean is operated.

A small peristaltic pump is installed on the filtrate side of the filter unit, this supplies the sample to the sample pot, which has been designed to hold a minimum amount of sample. The analyser draws its sample from this pot and the excess flows to waste via an overflow.

The dosing control unit uses an electronic control system for signal input and output. This involves the input of a flow signal and a phosphate concentration signal. The unit has fail-safe defaults, that can be set by the operator, and alarm set points developed with Wessex Water. It links to the dosing system via an analogue input to the pump via an existing cable, and transmits a combined flow/phosphate milliamp output.

The MicroMac sampling system is based around analysing ortho-phosphate (ortho-P), as opposed to total phosphorous. This is because the ortho-P chemistry is simple and fast, unlike the total phosphorous chemistry, which requires additional processing within the MicroMac C and has a longer cycle time.

Iron reacts directly with the soluble phosphate, while the total phosphorous is largely in the bound form with the solids, and is therefore flocculated with the solids. It is also to be noted that in order to analyse for total phosphorous, it would be necessary to analyse an unfiltered sample, which, on the inlet, would be impossible.

The MicroMac C single-chemistry colourimetric analyser determines ortho-P in the range 0-20mg/l as phosphorus. The analyser uses the standard molybdenum blue chemistry. The analyser uses a loop flow analysis (LFA) system, which gives it the flexibility to match the performance of a laboratory analyser with much less attention. LFA is also flexible in programming and it has been possible to incorporate much of the routine maintenance of the colourimeter in the operating cycle.

“Since first trialling and then fully commissioning the Partech system, which extracts a sample directly from the crude at intake, we have been able to produce a plot so we can see what the ortho-P is doing over a 24 hour period,” says Preece. “It gives a more or less continuous value of P, and it shows at Keynsham that there is no significant influx of phosphorus in the crude sewage from any industrial users at present.”

Seasonal adjustments

Through the summer months, the ortho-P intake value can rise up to as much as 11mg/l. But in winter, it can be as low as 2-5mg/l as it is diluted by heavier rainfall. “Because we were flow proportional,” continues Preece, “we needed to dose accordingly with slight allowances for storm conditions. Before the introduction of the Partech system, when we had a ferric dose that was not according to the P content or loading, it was predominantly down to the increase in flow through the intake channel.”

With the Partech system, the monitor is on the front end, so staff can see what phosphorus is actually doing. This is a great benefit to management for it assists them in evaluating ferric dosing at other treatment plants which have a phosphorus consent. At the much larger Saltford works a few miles away, the annual amount of ferric sulphate used is significantly higher, so based on the savings already made at Keynsham, it is likely further savings can be made at Saltford.

Because Keynsham is an unmanned site, it is important that the system was reliable enough to only require infrequent inspection and servicing. Preece reports that the monitor cabinet needs refreshing with reagent and water every two months, and the sample pump and filter unit in the channel needs cleaning once a month.

Based on the success achieved at Keynsham, Wessex Water says it is planning to install three Partech systems at the Saltford WTW where the consent level is 1mg/l total phosphate.

“What we now have is a system that allows us to determine in real-time how much ferric sulphate needs to be dosed in relation to P content, which is particularly important during storm conditions, when historically, the dosing level has been much higher,” says Preece. “Although it is still early days, the indicators reveal that during the first year of operation, the consumption of ferric has been reduced between 10% and 20%. In terms of percentages, the best case scenario at the moment is 20%.

“Until we put it on a bigger works, we will not be able to put a figure to the actual cost saving. However, having said that, even a reduction of 10% on a plant using in the region of 1,500 to 2,000tonnes a year would be impressive, and more than pay for the cost of the feed-forward control system.”

Action inspires action. Stay ahead of the curve with sustainability and energy newsletters from edie

Subscribe