GB2230619A - Chemical dose control - Google Patents

Abstract

Feedback-controlled chemical dosing systems are widely used in water- and sewage treatment, for example in the removal of phosphates from treated sewage by complexing with added iron (III) sulphate, where the progress of the reaction is monitored by following the redox potential of the sewage. However, error can arise in such control due to factors such as changes in background redox of the sewage or variation in the response of the probe measuring the redox values. As shown, a control unit 5 operates a pump 6 to add ferric sulphate to fermented sewage 1 in accordance with the redox potential measured by a probe 4. A data logger 7 records the redox value every 10 minutes, takes the average and, every 24 hours, makes any necessary adjustment to the measured value/dose rate correspondence by means of a potentiometer 8 and thereby compensates for any drift in the values not resulting from the dosing operation. <IMAGE>

Description

Chemical Dose Control This invention relates to optimisation of the performance of feedback-controlled chemical dosing systems, and concerns in particular such systems employed in the treatment of effluent - for instance, the removal of phosphates from sewage.

The fundamental requirement of the treatment of effluents like sewage and industrial waste is to remove objectionable solids from the raw material, and especially to lower its organic load (the content of organic matter), because direct disposal of such effluents to surface waters (such as rivers and streams) introduces disease organisms, and causes the prolileration of micro-organisms capable of fermenting, or in some other way digesting, organic material. This so dramatically increases the rate at which oxygen is removed from the water through respiration that even replenishment by aeration can no longer maintain normal aquatic life. In addition, the ensuing anaerobic conditions generate disagreeable odours.

Biological treatment of organic wastes frequently involves the fermentation of the organic matter therein by mixed populations of bacteria, protozoa and fungi.

During this fermentation phosphorus is released, in the form of relatively simple inorganic phosphate molecules.

This phosphate can stimulate plant growth - it may act as a co-factor in the plants' photosynthetic reactions (whereby energy from light is stored in high energy phosphate bonds during the chemical conversion of carbon dioxide to carbohydrates) - and unfortunately those plants able to respond most rapidly are the algae.

Accordingly, if ever this thus-fermented waste is disposed directly into surface waters the algae naturally present therein grow and produce turbid waters and unsightly surface scums. As a result, the more desirable vascular plants are shaded from light, and eventually die: these dead plants, plus sedimenting algae, gather on the bottom, gradually reducing the depth of the water body. In addition, oxygen may be depleted overnight by algal and bacterial respiration, which may debilitate or kill fish.

Algal proliferation can, however, be controlled by removing sufficient phosphate from the waste to ensure that, in the receiving water, phosphate never reaches growth-promoting concentrations. Such a technique is typical in the treatment of sewage, where the desired phoJph-.e removal can be achieved by dosing the biologically treated (fermented) sewage with ferric sulphate (or some other coagulant), which rapidly reacts with the phosphate to form insoluble hydroxo/phosphate complexes. These settle, and are drawn off as "sludge" in specially designed settlement tanks. When sufficient ferric sulphate has been added to the sewage to complex all the phosphate in this way, any further addition results in an excess of the ferric (Fe3+) ion, which causes a marked increase in the redox potential of the solution.This increase provides a convenient means of controlling the rate of ferric sulphate dosing. A redox probe, positioned downstream of the dosing point, and which forms part of a feedback-control system, can detect this increase and cause the system to respond by reducing the rate of dosing of the ferric sulphate. In this way, phosphate levels in the treated sewage may be maintained within acceptable limits, and algal growth is checked.

Unfortunately, factors other than the concentration of ferric ions may also affect the redox value apparently sensed by the probe, and thus erroneously influence the dose rate. There might, for example, be a variation in the background redox of the sewage (arising from changes in the chemical composition of the sewage itself, or perhaps as a result of the addition, in hot weather, of oxidants to the sewers), or there might be a change in the response of the probe itself (probably as a result of ferric and other ions becoming attached to the sensory surface thereof). Both these factors can cause the measured redox value to diverge from the "true" value, and give rise to incorrect dosing, and over a period may give rise to a significant drift in sequential values.Naturally, underdosing will occur 1L the measured value drifts upwards, overdosing if it drifts downwards.

Changes in the background redox of the sewage have normally been allowed for by utilising not one but two redox probes. One, as before, is placed downstream of the dosing point, and measures the redox of the treated waste, while the other is placed upstream thereof, and measures the redox of the untreated sewage - the ambient redox. Any change in background sewage redox is then eliminated from the first signal before it is sent to the control system by subtracting the signal produced by the second probe. However, a divergence in the responsiveness of the two probes due to their different environments may occur. For example, the performance of the downstream probe may change under the influence of continued exposure to dosing chemical, while that of the upstream probe may alter because of its position in phosphate-laden untreated material.The probes must therefore be regularly cleaned and calibrated if they are to respond similarly.

The provision of two probes instead of one is not, therefore, a totally satisfactory method. An alternative which has been adopted reverts to the use of only one probe, but additionally employs a human operator, who looks back over recently-recorded redox values and, either mathematically or visually, determines the average value thereof. As regularly as possible, he then fixes the centre of the response range of the control system - commonly referred to as the "zero point" - to coincide with the average value, typically by adjusting an analogue potentiometer within the system. This ensures that any progressive change either in the background redox potential or in the res#o#viness of the probe is compensatea tor, and will not result in incorrect dosing.The approach may, however, have a problem of its own - namely, that the apparent redox value may drift out of the response range between the times at which the operator makes the adjustment.

The present invention seeks to provide apparatus for the control of water treatment chemical dosing systems, especially those used in sewage phosphate removal by the addition of ferric sulphate, which apparatus regularly and automatically adjusts the correspondence (or relationship) between the value of the measured control parameter (say, the redox value of the treated sewage) and the dose rate (the rate of addition of ferric sulphate) upon the basis of the average of several recent such values, thereby continually compensating for any drift in the values whether caused by changes in the chemical nature of the material (say, the mineralised effluent) to be treated or by changes in the responsiveness of the sensor which measures the parameter.

In one aspect, therefore, this invention provides apparatus for controlling the dose rate of chemicals in a water treatment process, which apparatus comprises: control means, for adjusting the dose rate; measuring means, for continuously determining the value of some parameter of the water quantitatively representative of the need for treatment, and sending the measured value to the control means for thereby adjusting the dose rate in correspondence therewith; recording means, for storing at first intervals the instant value of the parameter; and calculating means, for evaluating the average of a pre-determined sample of the most recently recorded parameter values, and at second intervals sending the present average value to the control means for thereby re-determining the correspondence between the measured parameter values and the dose rate so as to re-establish the original correspondence between the "true" parameter values and the dose rate, and thus compensate for any drift in the calculated average.

In a second aspect, this invention provides a method of controlling the dose rate of chemicals in a water treatment process, which method consists of: continuously determining the value of some parameter of the water quantitatively representative of the need for treatment, and sending the measured value to the control means for thereby adjusting the dose rate in correspondence therewith; and at first intervals storing the instant value of the parameter; evaluating the average of a pre-determined sample of the most recently recorded parameter values; and a; second intervals sending the present average value to the control means for thereby re determining the correspondence between the measured parameter values and the dose rate so as to re establish the correspondence between the "true" parameter values and the dose rate and thus compensate for any drift in the calculated average.

Though the apparatus and method of the invention may be of use in the control offialmost any kind of water treatment by chemical dosing, it has been developed primarily with the treatment of sewage (and other similar organic effluents) in mind, and specifically for the subsequent treatment of fermented sewage with ferric sulphate to remove excess phosphate therefrom. In that sense it represents a significant improvement in the similar sewage phosphate removal arrangements discussed above.

The principal components of any apparatus used in this type of dosing treatment will be those conventionally relevant to the treatment being carried out. Thus, for sewage phosphate removal they include some type of redox electrode to monitor, via the redox potential, the progress of the dosing, and a recording device to store the measured redox potentials. In addition there is naturally a dosing pump of some description, which supplies the ferric sulphate to the sewage in response to a signal from a controller, which signal is in turn generated in response to the magnitude of the measured redox potential.

The apparatus of the invention incorporates control means which adjusts the dose rate of the treatment chtmie > . This control means may be any convenient commercially-available standard control hardware - for example, a pHOX 40-00/144C with proportional output, available from pHOX Systems Ltd. - capable of converting an input signal (the value of the control parameter) into a suitable output signal which it sends to the dosing pump. In the embodiment of the invention mentioned hereinbefore which controls the addition of ferric sulphate to sewage, the primary input signal is the redox potential value'of the sewage and the output is an inversely-related current.Thus, the dose metered by the dosing pump (which is preferably a pump with an adjustable speed and stroke, the general range of the pump's dose rate being set by adjusting the stroke, and the output signal then being used to adjust the speed, and thus the specific rate, correspondingly) is regulated so as to increase the dose rate of ferric sulphate whenever the measured redox potential decreases (indicating a deficit of ferric ions in the sewage), and vice versa. The control means also receives a secondary input which serves to adjust its output to compensate for drift in the measured parameter value average; this is described in more detail hereinafter.

The signal fed to the control means (conveniently as an electric voltage supplied along conventional wires) is determined by some measuring means the precise nature of which naturally depends upon the parameter to be measured. For many applications a suitable parameter will be the redox potential of the treated aqueous liquid, which changes depending upon whether there is too much or too little dosing chemical. An example of a dosing chemical having this effect is the coagulant ferric sulphate, which is used both for the removal of phosphate from sewage and for the flocculation of clay particles in river water (which cause turbidity and must be rer ed before the water can be used fcr public supply).

Redox potential may be measured by any of several suitable commercial redox probes. These vary in type and sophistication, but a preferred such device is the dip-type Model D0040 Redox Electrode sold by pHOX Systems Limited, which measures the unknown potential, using a platinum electrode, against a calomel reference electrode.

The instant value of the cQntrol parameter is at intervals recorded (stored). The recording means is advantageously that type of standard hardware, combining both meter and data recorder, known as a data logger.

The preferred logger for the purpose of the control of sewage phosphate removal is a slightly modified version of the Squirrel (1986) model logger produced and supplied by Grant Instruments (Cambridge) Limited. It stores the value sent to it in semiconductor memory and in digital form. In addition to storing instant redox potential values at pre-determined intervals, this unit may be programmed to evaluate the average value of a certain selection of those stored values, which average is required (as is described in more detail hereinbelow) to determine the correspondence between the measured redox potentials and the dose rate.

The time interval chosen after which the control parameter is stored will in part depend upon the rate of divergence of the apparent (sensed) and the "true" control parameter values. For example, in the case of sewage phosphate removal the redox potential may actually change relatively slowly as a result of natural variations in the background redox potential as the composition of the sewage changes, or may apparently change, more slowly still, if the responsiveness of the probe changes due to clogging of the sensory surface thereof by the deposition of ferric and/or other ions from the dosed liquid. The interval is also ultimately limited by the capability of the equipment employed. In the sewage phosphate removal treatment of the invention's primary embodiment, the aforementioned Squirrel logger has been found to perform particularly satisfactorily when an interval of 10 minutes is left between each consecutive recorded redox potential - this normally being enough adequately to trace the essentially slow drift of the sensed redox value due to the two factors just mentioned.

The calculating means of the apparatus of the invention evaluates the average of a sample of recentlyrecorded parameter values, and periodically sends that average to the control means so that the correspondence between the measured control parameter value and the dose rate is re-determined and, if necessary, re-set.

The evaluation of the required average is, as just indicated, conveniently carried out by the data logger which records the control parameter data. The data sample of which the average is taken (the term "average" here will in most applications denote an arithmetic mean) will normally consist of a convenient number typically around 100 - of consecutive values up to and including the most recently-recorded figure. For figures recorded every ten minutes, the average may usefully be calculated taking the data of the immediately preceding 24 hour period (which thus consists of 144 values).

Though the calculation can be done by analogue circuitry, it is generally more convenient to perform the computations digitally, and then (if necessary) convert the result to an analogue signal to send to the control means.

At intervals the present average value is sent to the control means. This interval may be of any convenient length - perhaps corresponding to that period over which the average was determined (24 hours, for example), or perhaps being equivalent to that other (first) time interval at which the control parameter values are recorded (10 minutes, for example) - but it will naturally be dependent upon the expected drift rate, and the shorter the interval - the more often the average value is used to re-determine the parameter/dose rate correspondence - the more closely will the dose rate follow the actual, true, needs of the liquid being treated.

Quite how the present average value influences the control means following receipt from the calculating means depends naturally on the nature of the control means - and, specifically, on how its response range (or zero point) is determined and controlled. In principle, however, it is merely a matter of adjusting the zero point up or down, so changing the correspondence between the measured control parameter value and the dose rate - thus, in effect maintaining the range of the control output to the pump despite the different range of outputs from the measuring means.

Typically, the secondary input to the control means, from the calculating means, is in the form of a voltage which is adjusted, in order to effect the necessary shift in the level of the zero point, by way of an associated variable resistor. This is in practice achieved by employing a potentiometer, conriected between the calculating means and the control means. Most conveniently the potentiometer is a digital device, directly controlled by the average value derived by the calculating means.

The calculated average measured parameter value is used to re-determine the correspondence between the measured parameter values and the dose rate, so as to compensate for any drift in the calculated average value. In this way there is re-established the original correspondence between the "true parameter value and the dose rate. In this context, of course, "true" means the actual value directly attributable to the reason for carrying out the treatment process. Thus, for example, in the described sewage treatment process, the "true" value is the actual redox value of an undamaged redox probe, or - where the background redox level is changing - is the proper proportion of the actual value attributable to the treatment per se.

The invention extends, of course, to water treatment equipment, especially for the treatment of sewage or other organic effluent, whenever using apparatus or a method as described and claimed herein.

A specific embodiment of the invention is now described, though by way of illustration only, with reference to the accompanying Drawings in which: Figure 1 shows schematically the components of an apparatus of the invention used to control phosphate removal from sewage; Figure 2 shows in graphical form the output of the redox probe of Figure 1, and the corresponding adjustment made to the response range of the control hardware; and Figure 3 illustrates pictorially the principles of zero point adjustment.

Figure 1 shows a stream of fermented sewage (1; the source is not shown) being treated at a dosing point (2) with ferric sulphate, Fe2(SO4)3, in order to remove phosphates therefrom. Downstream of the dosing point 2 are a staggered plate mixer (3) and a redox electrode (4). The redox potential output of the electrode is fed into dose control hardware (5), which in turn sends an inversely-related output signal (9) to a variable stroke dosing pump (6) which supplies ferric sulphate (from a source not shown) to the dosing point 2. In addition, the control hardware 5 transfers the electrode output as a signal (10) to a data logger (7) which records the instant value of the redox potential after some fixed interval.After another fixed interval the logger calculates the average of the several values recorded, and this average is then sent to a digital potentiometer (8) which in turn sets the centre (zero point) of the control hardware's response range.

The input signal received by the control hardware 5 thus has two components: the first - the electrode output - serving to measure and display the redox potential values within a specific range; the second the calculated average sent from the logger - serving to provide a movable range of values over which the inversely-related output signal 9 is generated. This range of values, being moved by a variable resistor, is positioned such that its zero point coincides with the ambient redox value of the liquid being treated. This produces a current output with a value mid way between the maximum and minimum, which of course causes the pump to operate at half speed.

When the apparatus is first put into operation, the > ta c. upshould ideally be carried out -.;l;en flow thrcx the sewage works is about average (i.e. not during low flow at night or during flow surges just after dawn).

The pump stroke is set at its zero or "neutral" setting, so that no chemical is delivered even though the pump's motor may be running (the pump's plunger - or diaphragm - is not moved by the motor).

The modified Squirrel logger is then switched on line to set the zero point of the control range at the ambient redox value. This ensures that there is a response range of 100 mV above and below the ambient redox value to accommodate changes in phosphate load.

The setting also produces an output current of 12 mA (half way between 4 and 20 mA) thus causing the pump to run at half speed (there is, however, no delivery of chemical as the pump stroke is set at zero).

The pump stroke is gradually increased, causing chemical to be added to the sewage, and phosphate concentration is simultaneously measured in the dosed effluent. The stroke is increased in increments until sufficient chemical is being added to the effluent stream (the flow rate of which is about average) to remove 95% of the phosphate load. The pump stroke need not then be altered further unless the average daily flow through the works changes. Henceforth the system, by responding to changes in measured redox values, will automatically adjust the speed and therefore the output of the pump (dose rate of chemical) to the phosphate load of the effluent.

Figure 2 shows a plot of the redox potential (in mV) measured by the probe 4 against time (in days), as produced by an apparatus which records the redox value every 10 minutes. The Figure also shows the pertaining constant zero point value as set by the digital potentiometer 8 every 24 hours.

Figure 3a shows an entirely fictitious situation of a substantially constant sewage throughput and a failing probe. More particularly, it shows the relationship which exists between the descending apparent redox value sensed by a probe which is gradually becoming clogged by deposition of ions onto its sensory surface, and the redox value derived from minor fluctuations in ferric ion (and hence phosphate) concentration. Figure 3b shows the associated pump control output signal.

The output of the dosing pump is confined to a fixed range, centred around the zero point, whatever the apparent redox value sensed by the probe. Therefore, in the case of this downward drift in the apparent value, the pump would normally attempt to compensate by adding more and more ferric sulphate to the sewage (as indicated by the pump output control signal in the initial part of Figure 3b). Eventually, the pump would permanently operate at full speed - even though the actual treatment required would not warrant this. The potentiometer stops a situation such as this occurring by moving the zero point - as indicated by the sudden drop in the signal of Figure 3. Hence, the pump output range is adjusted to negate the effects of the loss of probe responsiveness, and is always positioned around the "true" pertaining redox potential resulting from changes in ferric ion concentration as phosphate is removed from the sewage.

Claims (13)

CLAMS

1. A method of controlling the dose rate of chemicals in a water treatment process, which method consists of: continuously determining the value of some parameter of the water quantitatively representative of the need for treatment, and sending the nltasllred value to control means for thereby adjusting the dose rate in correspondence therewith; at first intervals storing the instant value of the parameter; evaluating the average of a pre-determined sample of the most recently recorded parameter values; and at second intervals sending the present average value to the control means for thereby redetermining the correspondence between the measured parameter values and the dose rate so as to re-establtsh the correspondence between the "true" parameter values and the dose rate and thus compensate for any drift in the calculated average.

2. A method as claimed in Claim 1, when used to control the addition of iron (III) sulphate to sewage in order to remove phosphates therefrom.

3. A method as claimed in either of the preceding Claims, in which the measured control parameter is the redox potential of the liquid being treated.

4. A method as claimed in any of the preceding Claims, in which the dose rate is adjusted by changing the speed of a dosing pump.

5. A method as claimed in any of the preceding Claims, in which the first time interval is 10 minutes.

6. A method as claimed in any of the preceding Claims, in which the second time interval is 24 hours.

7. A method as claimed in any of the preceding Claims and substantially as hereinbefore described.

8. Apparatus for carrying out a method of controlling the dose rate of chemicals in a water treatment process, which apparatus comprises: control means, for adjusting the dose rate; measuring means, for continuously determining the value of some parameter of the water quantitatively representative of the need for treatment, and sending the measured value to the control means for thereby adjusting the dose rate in correspondence therewith; recording means, for storing at first intervals the instant value of the parameter; and calculating means, for evaluating the average of a pre-determined sample of the most recently recorded parameter values, and at second intervals sending the present average value to the control means for thereby redetermining the correspondence between the measured parameter values and the dose rate so as to re-establish the original correspondence between the "true" parameter values and the dose rate, and thus compensate for any drift in the calculated average.

9. Apparatus as claimed in Claim 8, wherein the measuring means comprises a redox electrode.

10. Apparatus as claimed in Claim 9, wherein the control means converts the redox potential value of the liquid being monitored into an inversely-related output current which adjusts the speed of the chemical dosing pump so as to increase the dose rate as the measured redox potential decreases and vice versa.

11. Apparatus as claimed in any of Claims 8 to 10, wherein a data logger is used as both the recording means and the calculating means.

12. Apparatus as claimed in any of Claims 8 to 11, wherein the calculated present average value is supplied to the control means in the form of a potentianeter-adjusted voltage, which shifts the level of the zero point of the control means.

13. Apparatus as claimed in any of Claims 8 to 12 and substantially as hereinbefore described.