Industrial Wastewater Treatment...

Introduction...

There are numerous problems associated with the operation of any activated sludge treatment plant and these vary from site to site and process to process. It is true to say that when applied to the treatment of industrial waste the problems and complexity increase. It is not only influenced by the actual incoming waste stream but also the engineering of the plant, the skills and knowledge of the process manager and in particular the operator - all can affect the final plant performance. The way in which a plant is operated can result in poor effluent quality, over-expensive operation and premature investment in new plant. These three main issues can result in additional expenditure. In some ways paying the fine on a process effluent excursion can be the lowest cost option. However this option is politically more unacceptable and from the point of view of a manufacturing industry could be suicidal, especially if the effluent excursion forces the closure of the whole production facility.

Over-expensive operation can, in the case of industrially based waste treatment plants, be due to over-aeration or excessive chemical addition - be it nutrient or hypochlorite to control bulking / foaming. In the case of the former, attempts have been made to generate a formula for the addition of nutrients to feed liquors which is in most cases based on the incoming volumetric flow and not the actual mass of 'biologically oxidisable load'. Over-aeration can occur when 'too much' capacity is available resulting in excessive aerobic oxidation taking place.

Premature capital expenditure is quite common and is often due to poor operating practice giving the impression that the plant is under capacity. A poor sludge disposal policy can also contribute to this problem. An example of this is when highly polluted decant liquor from the dewatering process is reintroduced into the feed flow to the primary tank or into the activated sludge system direct. In the case of the former practice, this can cause overloading of the primary tank and in some cases solublisation of the solid waste, thus further increasing the organic load into the aeration system. If this solublisation creates readily-biodegradable substrate then this can cause serious oxygen demand problems which can in turn generate Poly Phosphate heterotrophic species which have the ability to take up a higher proportion of the added phosphorus nutrient than expected. This phenomenon is well documented, but its significance to industrial treatment plant is not fully appreciated by their operators.

The Key Issues of Operating an Industrial Wastewater Treatment Facility...

Some of the key issues in operating an Industrial wastewater treatment plant are presented. The issues take the form of common questions and points raised by process managers and operators.

Why is it Necessary to Know the Load Entering the Plant ?...

The nature and strength of the load entering an oxidation system such as activated sludge, will of course affect the performance of the biomass and influence the nature of the species produced. Low loaded plants are notorious for producing poor settling sludge. Under-dosing key nutrients can show similar symptoms on some facilities. Controlling the nature of the waste stream is very difficult in some industries, particularly those where a range of products are produced in no particular regular quantity. This is the case for specialist paper mills and pharmaceutical manufacturers. Sometimes poor housekeeping in the main production area can lead to unnecessary waste being discharged to the treatment plant. The discharge of known toxic materials as 'slugs' of waste can have a dramatic effect on the whole treatment process. Expecting a treatment facility to handle a very wide range of loads is asking for trouble. By its very nature an activated sludge plant performs best if operated using load balancing.

In the case of a facility where the discharged waste stream is quite variable in both strength and toxic content, then some form of flow or load buffering is very important. To achieve this requires a monitoring system that can both assess the load of the incoming waste stream as well as quantify its potential toxicity. The only satisfactory on-line method is respirometry as this approach actually uses the biomass that is treating the waste to measure the biologically oxidisable impact of the feed. Also by adding a known synthetic substrate in the same respirometer, toxicity in the feed stream can be confirmed. Should a toxic or high load waste be detected then the waste stream should be diverted to a balancing tank whose contents can be reintroduced into the feed stream in a more dilute form.

One of the most difficult problems to quantify on any treatment plant is the discharge of chemicals that slowly accumulate in the biomass, inhibiting its ability to treat the feed and eventually causing an effluent excursion or an unexpected change in the properties of the biomass. In the case of a treatment plant that nitrifies, then this more subtle form of toxicity can of course be more readily noticed as nitrifiers are more sensitive. In the case of treatment processes that discharge chemicals whose concentration and load is already suspected to affect the biosystem but not identifiable as toxic, then additional specialist monitoring equipment may be required.

Is the Feed Liquor Toxic to the Biomass ?...

This is a very important question when dealing with the treatment of industrial waste. The conventional definition of toxicity is a substance that reduces the viability of living cells such as biomass. This same phenomenon could also be observed if the plant is short of nutrients. It is very important that toxicity in the waste is identified and not confused with nutrient limitations as the necessary action to be taken is quite different. Most activated sludge processes can accommodate toxic compounds in limited concentrations and in fact can become acclimatised to them. The key to the handling of toxic compounds is reducing their use where possible, but more significantly, being capable of detecting their presence in the feed stream and taking appropriate action to minimise their impact on the ecosystems. The design of the aeration system can also help reduce the impact of toxicity. A plug flow system is more susceptible to a toxic discharge than a totally mixed system such as an oxidation ditch. Fortunately, in the case of most industrial plants that primarily handle carbon-based waste with sludge ages of 2/5 days, the impact on the biomass, which contains mostly heterotrophic species, can be minimal as the system can recover quite quickly provided the process is protected against high doses of toxic waste.

How Much Nutrient Should be Added ?...

This is one of the most difficult problems associated with the treatment of any industrial waste stream that requires the further addition of key nutrients such as Nitrogen and Phosphorus. Unfortunately the bulk of research work on the activated sludge process has been centred on municipal facilities where nutrient limitation is rarely experienced. Currently, until better methods are developed, the suppliers of such nutrients recommend a formula based on the estimated load entering the treatment facility linked to the flow. The same companies usually offer a service to check that the addition regime is working correctly by carrying out simple tests plus microscopic examination to identify that any biological indicators of nutrient limitation are not present. Also most operators assess whether the nutrient addition has worked by analysing the effluent either continuously or daily for Amm-N and Phosphate. This very simple approach should on paper be sufficient to test that the dosing regime is correct. Unfortunately MSL has found that this qualitative approach can still result in insufficient nutrient being added at the right points in the system which can lead to poor sludge settling and a general deterioration in treatment.

The problem of accurately knowing the level of nutrient addition required becomes even more critical when the plant has a tight effluent consent for both total nitrogen and phosphate. This is particularly the case for Paper & Pulp Mills in North America where the effluent constraints force the dosing of nutrients to be finely controlled but without the appropriate tools to help control and monitor it effectively, let alone continuously. The only solution is to accurately control the addition of nutrients to the feed stream using a combination of on-line respirometry and continuous analysers for phosphate and Amm-N. The technique, although novel, proposes using well established technology linked together in a single integrated unit. The actual monitoring unit would provide information on the load and toxicity entering the treatment stream. In this way the respirometer can determine the amount of treatment required in the presence of excess nutrients. Once the respirometer has detected 'endogenous respiration' the Amm-N and Orthophosphate content of the biomass can be determined accurately.

The graph termed as 'respirograph' shows the way in which the respiration rate of a biomass changes as the oxidiseable substrates present are used up. Depending on the ratio of RAS to inlet liquor, the maximum respiration rate and time to treatment will be affected. In the example the time for the sample in the respirometer to reach endogenous respiration was 65 minutes. The time to reach the 'endogenous state' is a very accurate measure of the ideal minimum retention time in the treatment facility to ensure complete treatment. In the case of this particular sample, nutrient limitation was not a problem so the information obtained was very accurate. The 'total area under the curve' provides a measure of the total oxygen requirement of the sample, i.e. the minimum total amount of oxygen required to achieve complete treatment. In addition to this the concentration of oxidisable material can be obtained by subtracting the demand due to endogenous respiration. This parameter is termed the 'area under the curve'.

In hardware terms, the system would consist of an MSL respirometer, Amm-N and Phosphate analysers, together with a suitable membrane filtration system for filtering the biomass. The respirometer would be injected with a known amount of excess nutrient comprising Amm-N, Phosphate and other required nutrients in sufficient quantities not to affect the respirometer's performance, but just enough to ensure excess nutrient was present throughout the test. The respirometer would carry out its normal multiple reaerations and decays until the biomass reached the endogenous state. The sample of biomass would then be analysed for the residual Amm-N / Phosphate content. The result of this automatic analysis regime briefly described here is to : (1) Accurately determine the 'Biologically Oxidisable Strength' of the feed knowing that the nutrient concentration was not limiting. (2) Accurately determine the amount of nutrient to be added to the feed liquor / Return Sludge (monitoring the requirement of the feed liquor for nutrient in isolation can result in the wrong information due to phosphate release occurring in the return sludge contributing to the phosphate requirements. It is essential that a mixture of the RAS and feed liquor are used for the determination of nutrient requirements). (3) Accurately determine the potential toxicity of the incoming waste.

By determining the residual Amm-N and Orthophosphate in the biomass after the test, the level of these constituents remaining represent the amount of excess added. Using this information the addition of nutrients can be dynamically controlled on an hour to hour basis. A suitable control regime would gradually increase or decrease the dosage of nutrient based on the respirometric data being generated. Should the instrument detect a high load entering the facility it would automatically recommend that the nutrient dosage be increased pro-rata to match the load, only reducing the ratio when the load drops or is partially diverted to a buffer facility.

The benefits from this type of inlet monitoring regime would be to optimise the costs of treatment as well as protect the activated sludge plant from potential toxicity and extreme loads entering the process. This type of monitoring and control system would also eliminate some of the variables that are blamed for a change in the characteristics of the biomass and subsequent poor treatment performance. Also by continuously and dynamically monitoring the actual nutrient requirements, changes in the demand will also identify problems in the system and act as an early warning of a possible process failure.

The capital cost of such a system would be approximately £ 60,000. The cost savings by adopting this approach would manifest itself in more economic use and consumption of the nutrient chemicals, plus providing the plant operator a quick system to identify excess loads entering the treatment well ahead of seeing their effects on the final effluent. The ultimate financial saving gained is to further minimise the possibility of complete production close down due to effluent failure. The benefits are in also knowing that the full capacity of the biomass can be utilised subject to engineering constraints. It would also be possible to offset some of the costs of such an investment by the use of the Amm-N and Phosphate equipment to monitor the final effluent leaving the facility as well. The cost and logistics of this additional use of the equipment would be subject to site topography and distances from sample points etc. In this case the additional costs would be for sample presentation and preparation.

How much aeration capacity is required ?...

The simple answer to this question is 'sufficient to ensure adequate treatment'. This is a very difficult question to answer rigourously. The best answer is 'sufficient to ensure that the plant meets its license'. If the load is quite varied, then sizing the aeration system is difficult, as is the control regime that is employed to achieve these objectives. In the case of a plant that nitrifies, producing nitrate in its effluent, operating zones of zero dissolved oxygen is acceptable as denitrification can take place. In the case of a purely carbonaceous oxidative system then low dissolved oxygen can result in the generation of Poly P species, which can demand more phosphate than is normally expected with some unusual results. In South Africa, where biological nutrient treatment plants have been engineered to remove phosphate biologically by generating Poly P Heterotrophic species, it is well documented that foam and bulking sludge can be a serious problem at times. Also in the case of an Industrial waste treatment plant whose waste contains sulphate, zero dissolved oxygen zones in the absence of nitrate can also generate levels of sulphide as well as impact the quality of the biomass.

The peak on the 21^st September shows the sudden influx of highly biodegradable waste into this oxidation ditch. This particular event could be the accidental discharge of some alcohols or the waste from the end of a batch sequence. In this case the respirometer is located in the main body of the ditch which meant that by the time the instrument had computed this information the whole ditch was loaded. If the instrument had been located on the inlet stream then this high strength waste would have been detected quite quickly and possibly diverted to a holding tank. Also it has been found that provided the MLSS (Mixed Liquor Suspended Solids) is maintained at a constant level, that simply monitoring the initial OUR values are a very good measure of the relative strength of the feed liquor entering the system. This only applies accurately to a plant that is primarily treating carbon waste and once the OUR curve has been characterised. In the case of this facility, this sudden increase in oxidisable carbon may have taken the facility outside its operating consent.

The total area under the curve and the instantaneous OUR are also quite valuable as these variables help to quantify the amount of oxygen required in total, plus the immediate oxygen input requirements to ensure that treatment is complete in the time available. In the case of very high OUR's, it is very unlikely that normal oxygenation will meet this demand, hence a reason to consider pure oxygen as a supplement or alternative to air. Obviously the optimum time computed by the respirometer will not apply if the aeration system cannot meet the oxygen demands at all times. Fortunately in the case of a fully nitrified facility where nitrate is present in the feed mixture of waste liquor and RAS, then some of these high demands are met by this alternative oxygen source.

Ideally the aeration system, be it a plug flow process or a totally mixed reactor such as an oxidation ditch system, should be capable of providing oxygen at ALL times to ensure excess dissolved oxygen is present. The ability of the system to achieve this level of oxygenation is down to oxygen demand of the biomass and its operating temperature. The higher the temperature the more difficult to achieve the correct level of oxygenation. At 38 centigrade the volume of air required to maintain a dissolved oxygen level of 3 mg / L could be so high that foaming could be a serious problem just caused by the volume of air that has to be introduced.

Until recently the only recognised method, be it quite crude, of controlling a totally mixed reactor such as an oxidation ditch, was by use of one or two single dissolved oxygen sensors. Extensive work in the Netherlands by MSL has demonstrated that respirometry is a far better tool for controlling an oxidation ditch than dissolved oxygen on its own. This respirometer is located at the outlet of the ditch so that it captures a representative sample of the biomass / settled sewage within the reactor but is also reflective of the quality of the final effluent. In fact it has been shown that provided the liquid velocity circulating in the ditch is at a normal level, then except for the inlet zone area, the location of the respirometer is not important as the plant can be treated as a totally mixed reactor. By simply capturing a sample, the respirometer determines the amount of treatment required to achieve the endogenous state. The longer it takes for the respirometer to take a sample to the 'endogenous state' the more capacity that is required for treatment in the plant.

In a control regime the respirometric information would be used to regulate the addition of more or less oxygen. This could be achieved by either turning on more zones or by increasing the dissolved oxygen set points for the fixed dissolved instruments. In the case of an oxidation ditch where the dissolved oxygen is added by use of a rotating brush system, then the set point of the local dissolved oxygen probes should be adjusted appropriately. This has the effect of increasing or decreasing the proportion of the ditch that is measurably aerobic. The control of a conventional plug flow plant is not as simple but follows a similar control philosophy.

The key to respirometric control of any aeration system is to know that the dissolved oxygen is not the limiting factor in the treatment of the waste. If the dissolved oxygen level is high enough but the respirometric data is indicating that treatment is not complete, then the operating concentration of active biomass must be increased or the size of the treatment plant needs increasing. The importance of biomass activity is covered next.

What is the Biomass Activity ?...

In any activated sludge system the aim of the process is to convert soluble organic material together with some inorganic chemicals into biomass, water and carbon dioxide. The speed of this process depends on the biomass active components, temperature and nature of the feed material (waste). Conventionally the Mixed Liquor Suspended Solids ( MLSS) has been considered to be the best guide to the actual level of active species. Whether the MLSS is determined gravimetrically or optically using on-line instruments, this assumption that the viability of the biomass is directly and reliably linked to the level of suspended solids is incorrect. The only way to measure viability, be it on-line or analytically, is by use of a respirometric technique. Fortunately doing the measurement on-line will give a more accurate and useful information than any laboratory based measurement. This is because in transporting the sample even a short distance will cause its nature to change.

When biomass is sampled and analysed by the MSL respirometer a number of parameters are computed. One parameter is the 'last respirometric rate' this corresponds in most cases to the respiration rate when the biomass is endogenous. Changes in the endogenous rate should reflect any changes in the concentration of the viable species in the biomass. In the case where the ratio of optical suspended solids (MLSS) to the 'endogenous ' respiration rate changes, then this suggests more solids or less solids entering the treatment plant. The ratio of the optical suspended solids to the standard gravimetric suspended solids is also a very useful indicator, suggesting that the biomass is becoming more or less bulky. Another advantage of respirometry for the monitoring and control of the biomass level, is that it is self-regulating in so much that as the liquid temperature increases so does the oxygen uptake rate. Controlling the plant on the basis of actual biomass activity would trigger an increase in wasting whereas the MLSS value would not. Obviously if the degree of treatment in the plant appeared to drop, then the actual level of activity would be increased and the wastage rate reduced.

The significance of this data is that it shows that although at first glance the patterns of initial OUR (decay 1 rate) would suggest that this plant is receiving a wide range in strength of food, when in fact the trace is showing a more serious problem, that of a very wide range of solids entering the system via the feed liquor and from the final clarifiers as RAS.

The 'total area' together with the optical suspended solids (ftu) show that the actual organic load is more associated to problems with the handling and return of solids, than a dramatic increase in the 'biological oxidisable' content entering the treatment plant. This diurnal pattern is quite typical of a short retention mixed reactor where the bulk of biomass is forced into the final clarifier at high flows. It supports the earlier observations that buffering should be available for any treatment plant likely to receive a wide range of flows, organic loads or toxic discharges. The events presented for these two days caused the effluent quality to deteriorate quite badly.

What are the Settling Characteristics of the Biomass ?...

The settling characteristics of the biomass are affected by both physical-chemical as well as biochemical factors. In industrial plants, because they are expected to handle a wide range of waste material, it is not surprising that the activated sludge's settlement characteristics can change, or that suddenly the system starts to produce foam. As mentioned previously, some types of foam or bulking sludge can be caused by low dissolved oxygen levels, high sulphate, low loading, toxicity and factor X. Factor X covers all the events that result in bulking sludge or foam that have yet to be explained ! I would suggest that the list is quite large to fill this category. The most critical element to be aware of is any changes in the settling characteristics or a change in the quantity and type of foam in the system. Monitoring or being aware of a change in the plant is very important to help identify and understand these problems. With the best will in the world, these kind of problems affect every process plant at some time and it is best to be aware of the actions to stop it getting worse, or better still preventing it happening.

A very useful parameter and an early indicator of a potential change in the nature of the biomass is to monitor the ratio of the optical suspended solids to the gravimetrically determined suspended solids, as well as knowing the level of viability and actual biological load being treated. All these factors can be used quickly to identify any change in the characteristics of the biomass and system. An increase in the ratio of optical suspended solids to analytical solids suggests that the biomass is starting to become more bulky. In the case of the reverse, then this could be attributed to an increase in inert solids entering the system. i.e. partially digested solids carried over from a primary settling tank or over capacity / over aeration mineralising the biomass. However by monitoring the endogenous activity and maximum activity by introducing a readily biodegradable substrate at times, the actual reasons for a change in these factors can be pin-pointed quite quickly. For example, a fall in biological activity with an increase in optical and gravimetric suspended solids, confirms that the biomass is accumulating more inert material. Obviously the temperature is also important as a drop in temperature also reduces the activity of the biomass.

Monitoring the solids level at two critical positions in final settlement tanks will help to identify settling as well as potential washout or serious rising sludge problems. Careful monitoring of the nature of the interface between the sludge and the water is also a very useful measure of the health of the biomass. A well defined interface can indicate potentially bulking sludge, but can also yield very clear effluent provided the tank is not hydraulically over loaded. Also being able to measure the thickening characteristics of the sludge in the final clarifier allows the settlement characteristics to be continuously assessed.

Is the Effluent Quality within Consent ?...

The quality of the effluent leaving a activated sludge system is affected by the extent of aeration treatment as well as the settlement in the final tank. Ideally, knowing the extent of treatment in the biomass whilst still in the aeration basin offers the best early warning for potential effluent excursions. This is particularly important if the information can be used to prevent an effluent excursion. Again this is where respirometry is very useful as it can provide quite an accurate estimate of the effluent quality. The extent of treatment can be determined by either the time taken to reach endogenous, or the 'area under the curve'. Both these parameters are an accurate measure of the amount of 'food' still to be treated in the liquor. In the case of a nitrifying treatment plant this factor is very accurate for determining the level of Amm_N in the effluent. As the Amm_N content increases so the closeness of this relationship drifts apart. This is to be expected, as an increase in Amm_N will also mean an increase in the concentration of biologically oxidisable carbon which will be measured by the respirometer but not the Ammonia monitor.

The level of suspended solids in the effluent is partially due to the extent of oxidative treatment monitored by this technique, but also the quality of settlement and the sludge loading on the final tank. Identifying that the solids in the final effluent are derived from insufficient treatment or from poor settlement, offers an important aid in pin-pointing the nature of the problem and ensuring that appropriate action is taken as soon as possible.

What Instrumentation is Required to Operate the Plant Effectively ?...

The answer to this question depends on the complexity of the treatment facility, both in terms of its engineering construction, as well as the nature of the feed liquor and the effluent consent license. Every Industrial plant that is required to add nutrients should try to automate this element of its operation. We recommend respirometry as the best tool together with the appropriate Amm_N and Phosphate analysers. Obviously monitoring the inlet flow to the facility is necessary as is monitoring the Return Sludge flow.

In terms of the aeration system, again this depends on the choice and configuration of the process. Unfortunately too many consultants design treatment plant without paying any attention to the process, let alone the monitoring and control elements for the system. It is because of this oversight that additional instrumentation is required to operate and control the system correctly.

In the case of dissolved oxygen, it is absolutely fundamental that the equipment can operate unattended with minimal maintenance or attention but always guaranteeing the quality of the data. Over-aeration can lead to poor settling in a clarifier or cloudy effluent. Similarly low dissolved oxygen levels can result in insufficient treatment. The actual number and location of the dissolved oxygen probes is subject to individual plants configuration, but with the introduction of respirometry for better control the exact location is not so critical but their presence still remains important.

Monitoring the final effluent quality continuously is only important if the effluent consent is either very tight, or the concentration of Nitrogen or BOD or COD is linked to a charge by the organisation receiving the effluent. Continuous monitoring of the suspended solids in the primary and final effluent is well worth the investment, as it helps to alert the operator and process manager of potential problems in the plant. An increase in the suspended solids in the primary effluent can lead to problems of higher oxygen demand, as well as forcing an increase in the solids loading on the final tanks.

The most fundamental element in selecting instrumentation for the monitoring and control of any waste water facility, is to ensure that the equipment is auto-cleaning and self-calibrating thus minimising the amount of external intervention, but more fundamentally, maintaining the users complete confidence in the data at all times.

One final point about the selection and type of instrumentation used to monitor and control any facility is that the investment could be wasted if the technology to allow full use of the data is not available. It is important that ALL the information monitored can be readily examined and the nature of the information fully appreciated by all those involved. The next section will expand and discuss this in more detail.

Is the Operating Practice Correct or Not ?...

The answer to this question can at times be very painful. Sometimes it is the operating practice that can cause serious problems to the process, leading to effluent excursions. No two treatment plants behave quite the same. Even with the increase of reliable self checking instruments, the one factor missing from the smooth operation of a facility is a large database of experience. The key to success is to take ALL the experience in the operation of the treatment plant and compress it into a simple 'expert' computer-based system linked to the appropriate information sources, be they instruments, analytical tests and/or visual observations. Such a tailor made system is not available off the shelf. Any conventional Supervisory Control and Data Acquisition (SCADA) system might be seen as being equivalent to this expert system outlined. This is not the case.

When dealing with a biological system an effective system is one that provides rapid access to the information in a visual and graphical form. By adopting technology similar to that covered in this document, together with the pooled knowledge base from other process experts and operators, it is possible to produce a system that can provide guidelines and advice to the plant operator. This approach will of course reduce the risk of a plant operator taking action that may cause a moderate effluent consent failure becoming more serious.

Currently there is a move towards even more advanced systems based on mathematical models. It is seen that incorporating mathematical models together with an 'expert' system would yield the ultimate software tool. It is envisaged that the model would allow simulations of adverse conditions as part of a continuing training program. The idea being that by putting the operator through such training, it would allow him to be aware of problems on the plant before they become too serious and be prepared to take the appropriate action with more confidence in the out come.

Summary...

This document has attempted to focus on the important issues associated with the control and operation of an industrial treatment plant. It has discussed the importance of protecting the biosystem from shock loads, toxic discharges and potential operational mistakes. It has also described how it would be possible to monitor and control the addition of nutrients, leading not only to a possible reduction in chemical usage, but more significantly protecting the biomass from any practices that could lead to poor effluent quality. Although it has described how respirometry together with dissolved oxygen can be used to optimise the degree of aeration without putting the final effluent quality at risk, it has avoided producing a specific formula for the number and location of all these instruments as the design of the treatment facilities vary so much. They all need to be assessed individually.

The fundamental message from this document is that with the correct quality and quantity of technology linked to a dynamic 'expert system' it is possible to train operators to operate the activated sludge process without having to recruit a new generation of process personnel holding higher degrees in process technology. The key is having total confidence in the data and information available with the minimum of delay in getting the information to the operator.