Two broad types of measurements can be used to control a wastewater treatment plant: laboratory analysis and on-line measuring. However, the latter method, which is still not widely used in France, is an essential means of improving the functioning of a treatment plant, in other words optimising operational parameters and achieving greater reliability of the treatment process.
Marc Viandon, Dr LANGE France Bruno Weiss, Dr LANGE France
Two broad types of measurements can be used to control a wastewater treatment plant: laboratory analysis and on-line measuring. However, the latter method, which is still not widely used in France, is an essential means of improving the functioning of a treatment plant, in other words optimising operational parameters and achieving greater reliability of the treatment process.
European Directive 91/271/EEC on the treatment of municipal wastewater, transposed into French legislation by the Law on Water of 3 January 1992 to set requirements for the operation of wastewater treatment plants and the quality of their discharges into the environment, is having a radical effect on the French approach to the size, design and management of wastewater treatment plants.
Not only does this directive change the permitted levels in treated water (a maximum concentration of BOD5 of 25 mg/l compared to the previous limit of 30 mg/l, for example), but it introduces two new concepts that were not included in the 1980 directives:
- - treatment reliability, since the parameters for treated water must be complied with 95 % of the time,
- - the cycle must provide for the treatment of storm water, “except in the case of exceptional rainfall”.
These two concepts will require not only in-depth studies to be made upstream to acquire a detailed knowledge of the flow to be treated but also a thorough reformulation of the management concepts governing a wastewater treatment plant.
Furthermore, operators should be aware that what is at stake in managing the municipal wastewater treatment system
(sewer system; wastewater treatment plant; receiving water) goes beyond the legal framework referred to above: the natural environment is being rendered more and more fragile by the ever-increasing volume of municipal discharges. There is now a need for denitrification and phosphate removal to be included in wastewater treatment. Integrating these processes in a treatment system adds to the complexity of the plant, consisting of one or more specific chains that include electromechanical machines combining hydraulic, physicochemical or biological processes, the latter being particularly difficult to manage and prone to malfunctioning. A modern wastewater treatment plant can be compared to a complex chemicals factory. A further complication is the fact that the plant never experiences stable operating conditions, either quantitatively or qualitatively: the profile of material entering the treatment cycle is constantly changing. What instruments are currently available to handle this complexity?
Instrumentation for wastewater treatment plants
Introduction
Two broad types of measurements can be used to control a wastewater treatment plant or the composition of its discharges into the natural environment. These may include laboratory analyses or on-line measuring. The former consist of analyses made after manual or automatic sampling using the appropriate sampling devices. In the latter case, captors are placed at the same strategic points and allow on-line measuring to monitor or regulate pollution levels.
These two approaches are not mutually exclusive: it is not a question of using one instead of the other, but rather of combining the two approaches to produce a wider range of information.
In France, wastewater treatment plants are on the whole poorly equipped with on-line measuring devices. Only the biggest plants use several sensors, mainly for real-time monitoring of rates of flow but also for the level of dissolved oxygen or the redox potential within the aeration tank.
The main reasons for reticence in acquiring such equipment for wastewater treatment plants are easily identified: generally speaking, it would appear that true or sequential on-line analysis has not yet gained official acceptance, leading to considerable differences in approach. In addition, on-line analysers cover only a limited range of parameters and, justly or unjustly, the quality of the measurements has not so far enjoyed a trouble-free reputation, and is in any case noticeably poorer than that obtained with laboratory analyses.
Complementarity of the two approaches
For a wastewater treatment plant to operate effectively, a precise knowledge of the composition of water at inflow and outflow and at various key points is required. To this end, a measurement strategy needs to be strictly defined, ultimately aimed at answering three questions:
- - What parameters should be measured?
- - Where?
- - How often?
Individual chemical analyses carried out in the laboratory on spot samples have, up to now, been the only response to this need for measurements, since they allow the required value to be precisely expressed and give access to a wide range of parameters. Furthermore, an analysis can always be repeated if there is any doubt.
However, these measurements are not entirely free from potential sources of error, such as sampling, for example. Taking a representative sample for subsequent analysis is not necessarily a straightforward task: what sample can be considered perfectly representative of the medium? Yet the answer to this tricky question is crucial, since no analysing method, however precise, can make up for a sampling error. The sample may also undergo modification during transport, either of the parameter of interest or as a result of interaction between different parameters. Lastly, it seems that measurement can be influenced by the subjectivity of the operator: intercomparative studies have shown that this effect may exist despite the use of standard procedures. However, by their very nature, these measures are not perfectly representative of the system and can do no more than provide an instantaneous “snapshot” of the existing phenomena, a view that is “frozen” in time. The overview of the treatment system thus obtained will therefore be limited and partial, and the values obtained may well over- or underestimate the true situation. This is probably the main weakness of individual spot sample chemical analyses.
To remedy this situation, it is essential to measure several samples taken simultaneously from different points in the system, thus providing a representative view, and to carry out repeated sampling at the shortest possible interval, allowing discrete variations in the system to be taken into account. On-line measuring tends to meet both of these requirements.
A medical analogy clearly illustrates the importance of considering the time aspect: a doctor can make a diagnosis based on various types of clinical observations (blood tests, X-ray examination, scan, etc.). Yet this does not stop a “perfectly healthy” subject suffering a heart attack.
The complexity of living things implies abandoning the notion of absolute. In the many wastewater treatment plants using activated sludge techniques, the most sensitive part of the treatment is based on biological phenomena involving bacteria, protozoa, fungi...
algae and yeasts.
A reliable diagnosis will be based not only on spot tests but on the results of continuous monitoring. This is achieved by on-line measuring, allowing dynamic variations in the parameters to be recorded.
On-line measuring also has other advantages, roughly corresponding to the drawbacks of repeated individual measurements: the sample is immediately analysed, eliminating the risk of poor sample conservation. Moreover, the analysis procedure is carried out systematically, without human intervention, making it perfectly reproducible. Lastly, on-line measuring shares some of the advantages of repeated sampling:
- - standardised procedures can be applied, analysis by an on-line analyser being none other than a faithful and automated reproduction of the operations that would be carried out by a laboratory technician,
- - the measurements carried out are regularly calibrated/recalibrated or are based on standards, thus ensuring their reliability.
These two broad types of measurements are clearly highly complementary and the movement towards automated surveillance of a wastewater treatment plant, and more generally of the entire network, plainly requires the development of a methodological approach combining repeated analyses and on-line monitoring of certain parameters.
On-line measuring devices; examples of applications
On-line analyses, which are still not widely used in France, can be classified into three families according to the measurement principle on which they are based:
- - Measurements using chemical methods. The majority of these use electrical intensities or temperature variations which, in the presence of catalysts, induce or accelerate certain chemical reactions. These measurements may be correlated with manual analysis of pH and certain metals, or indeed with global parameters such as COD, TOC or BOD.
- - Measurements of biological reactions. These are based on the reaction of the biomass, which may have acclimatised and adapted to the types of pollution usually carried by the effluent. These measurements can be correlated with BOD and provide information on possible toxicity.
- - Measurements based on physical methods: these sensors, based on physical measurements related to the optical properties of the effluent, are designed to provide information that can be correlated with analytical methods for the detection of material in suspension and for certain parameters of nitrogenated or phosphatic pollutants.
Some potential applications of the last-named type of analysers in a biological or physicochemical treatment plant will be dealt with in the remainder of the article. They rely on the fact that these analysers can be coupled to remote data transmission or automation systems with a view to process integration.
It should be borne in mind that the need to protect the natural environment calls for ever greater reliability of treatment plants, as witnessed by the changes in the regulations. This is particularly true in the case of the European Directive of 21 May 1991. Wastewater treatment plant operators will soon be required to ensure that treated water complies with stipulated levels 95% of the time. For the many activated sludge treatment plants, this requirement means that the entire biological process will have to be monitored at all times. In other words, whatever the circumstances, it will have to achieve total reliability. Yet, while the microbial population present in the sludge can accommodate qualitative and quantitative fluctuations in the wastewater entering the treatment process, it is nevertheless sometimes subject to undesirable phenomena: poisoning when concentrations of toxic substances in the untreated water exceed a given threshold, or inability to cope with a massive influx of organic pollution. Measuring standard parameters in the traditional way may result in the incident not being detected until hours or even days after the event, when it is too late to repair the damage. What often happens is that the detrimental effect on the activated sludge, though temporary, leads to a general malfunction of the process, which may take a long time to stabilise, seriously compromising compliance with the regulations. On-line measurement of pollution parameters in the water at plant input and output, whether they are global (COD, TOC) or specific (nitrates, ammonia, phosphorus) allows an imminent problem or ongoing dysfunction to be rapidly detected, depending on whether the measurements are made before or after treatment. Where necessary, corrective action can be taken without delay: storage of the incriminated effluent, for example. On-line measuring can thus help to improve the reliability of a biological treatment plant by allowing the operator to make a prompt decision on whether an effluent can be treated.
While on-line measuring can be used to rapidly detect any dysfunction, another field of application exists: the optimisation of certain operational parameters that are characteristic of the process. This can be illustrated by the following two examples:
- - At present, the most widely used method of eliminating phosphorus is to precipitate orthophosphate ions using metallic salts. The precipitating agent is often added to the aeration tank influent, in a process known as “simultaneous precipitation”. The optimal dosage can be determined by using on-line measurement of orthophosphates. There are two points at which this measurement can be taken:
- - at plant inflow: given the rate of flow and the concentration, one can estimate the load of orthophosphates to be treated and therefore the necessary quantity of metallic salts to be added in view of the stoichiometry of the reaction,
- - in the aeration tank or at plant outflow: a feedback control loop can also be considered. Compared to the first method, this has the advantage of taking biological removal of phosphorus into account.
Metallic salts can thus be optimised in terms of the efficiency of phosphorus removal, which can itself be estimated from the on-line measurement of orthophosphate ions at tank outflow.
• The number of wastewater treatment plants equipped with a specific treatment process for nitrogen (nitrification/denitrification) is continuing to increase. Nitrification (oxidation of ammonia to form nitrates) takes place in an aerobic tank, and denitrification (reduction of nitrates to nitrogen gas) under anoxic conditions, in other words without free oxygen. While it is possible to eliminate nitrogen by successive aerobic and anoxic phases, the anoxic and aerobic zones are generally separated in different tanks: the anaerobic tank is frequently placed downstream from the nitrification tank (“post-denitrification”). An important parameter for monitoring the effectiveness of the nitrification reaction is the ratio of organic carbon to nitrogen in the raw water. This ratio should reach 8 to 9 g of COD per gram of nitrates. The reduction in carbonaceous pollution achieved in the primary sedimentation tanks is often so effective that an external source of carbon, such as acetic acid or ethanol, has to be used. The dosage of these expensive products can be varied according to the residual nitrate concentration at tank outflow, which reflects denitrification efficiency.
A physicochemical type of wastewater treatment plant can also benefit from the implantation of on-line analysers. The correct functioning of this type of plant is based on the hypothesis that the pollution load is proportional to the rate of flow, a hypothesis that may prove wrong, as in the case of an influx of stormwater or heavily polluted water. Given the fact that the quantity of inorganic floc is itself proportional to the rate of flow, the dosage of reagent will in the first case be excessive and in the second case insufficient.
In each case, there would be a deterioration in the quality of the treated water. Determining the appropriate dosage of inorganic reagent, i.e. as a function of the quality and not the rate of flow of the wastewater, requires the reliable, reproducible and, more especially, continuous measurement of a pollution parameter such as the COD. Studies have succeeded in correlating flocculation with the efficiency curve for COD elimination (Marchandise, 1978).
COD can be estimated on line, by correlation with the absorption of ultraviolet light (Matsché et Stumwohrer, 1996). Most organic pollutants have, in fact, unsaturated chemical bonds or groupings of chromophores absorbing spectral bands between 200 and 350 nm. This method requires neither sample preparation nor chemical reagents.
Conclusion
Automated control of wastewater treatment plants is difficult but highly rewarding. The spread-out nature of the treatment plant, the relatively long retention time in the different sections of the plant and, in the case of activated sludge treatment, the complex kinetics of the biochemical processes involved, make the operation difficult to manage and therefore control.
It is nevertheless possible to establish control strategies (Ganczarczyk, 1983) based on the ability to measure the characteristic variables of the process on line. The firm Dr Lange offers a complete range of on-line analysers for COD, TOC, nitrates, ammonia, orthophosphates and total phosphorus.
Measurement is based either on colorimetry —spectrophotometry after a chemical reaction has been induced—, or direct spectrophotometry for nitrates and COD. The latter two types of analysers, the only available European models in the form of a probe, can be immersed in a pipe or tank, are easy to use and require little maintenance (see photo).
