The Dawn of a New Era: Real-Time BOD Data

Background

In many areas of our life, manual monitoring has been replaced by automatic or online monitoring as and when the instrumentation technology has been developed; however there is one area that has fallen behind the times and has not caught up which is water quality. Many parameters are simply not possible to measure in the field or can only be done by grab sampling. In the last ten years or so we have seen a number of traditional parameters being measured continuously using ‘substantial’ sensor platforms often using wet chemistry/reagents or very complex and expensive semi-laboratory styled kiosk arrangements. The drive for new ‘online’ technology is directly linked to our ‘need’ to monitor and control our environment and processes.

It is without doubt that BOD (Biochemical Oxygen Demand) is probably the Holy Grail of water quality parameters in that in just one parameter it can indicate health and organic loading of polluted waters. Since its adoption at the beginning of the 20th century by the Royal Institute it is used worldwide for indicating the level of pollution in our rivers, oceans and water resource assets. The 5-day test comes from the fact that no river in the UK has a flow time of greater than 5-days. Today, it is used the world over to monitor and control the discharges of Wastewater Treatment Plants (WTPs) to receiving waters. However, in almost >99% of cases the measurement of BOD5 is done manually by sending off water samples to a laboratory which will typically take an extra 2 3 days longer than the 5 days of the test itself. Receiving a result of non-compliance or a pollution exceedance alert 7-8 days after the event is simply of little practical help when trying to control a process or pollution event. The drive for innovation and new technology is seen to be key to unlock this process cycle and to be able to monitor and control it 24/7 wherever it is required to be monitored.

Why do we need to measure BOD?

BOD is an incredibly important parameter because in just one parameter it provides a globally recognised assessment of the water quality. The higher the BOD value, the higher the organic matter in the water which is essentially food available for oxygen consuming bacteria. If you increase the organic load beyond that of the supply of dissolved oxygen (DO) from aquatic plants, algae photosynthesis and oxygen diffusion from the atmosphere, then the aquatic ecosystem becomes severely stressed making it unsuitable for aquatic life. By stressing this relationship further the environment can become hypoxic or anoxic. So it is critical that when we measure BOD that we get it right.

Spot samples or continuously monitor?

A grab sample reading is a reported reading from one moment in time and can only be used to describe the situation at that moment in time, not 30mins before or 30 minutes afterwards. Furthermore, it is not possible to tell where this reading is on the data trend or whether the trend is increasing, decreasing or simply stable. This situation just creates more questions; questions that cannot be answered as there is no more data. The only way to trend grab samples is to take multiple samples which is a very costly and labour-intensive technique.

In contrast, continuous monitoring provides a data trend with time and date stamps which can be used in real time to make active decisions about the conditions present. By way of example, if a WTP was discharging waters to a river and were nearing their permit then by immediately changing their process they could ensure that they don’t exceed their BOD limit. This is often done for easier parameters such as turbidity, ammonia or pH but not until now, for the likes of BOD, COD or TOC.

If you asked any water quality professional whether they would prefer continuous or grab samples then continuous would always be the preferred option but the lack of technology has meant that spot sampling has typically been the universal standard for BOD sampling.

So how good or bad is the BOD5 process?

In fairness neither word, good or bad, is the right adjective to use; the process is only as good as the process and the person(s) undertaking it; if those standards change so do the results. To many, this may come as no surprise but it is inevitable that when more variables are introduced into a process the greater the risk for error. The BOD5 process is no different and sadly it is unforgiving. There are multiple opportunities for error which can have a very significant effect on sample results if not carried out absolutely perfectly:

  • Sample preparation;
  • Sample collection;
  • Transportation & storage;
  • Laboratory test.

The graph below summarises the increasing uncertainty the further you get into the process. So although a laboratory may offer you an uncertainty of +/-15 or +/-20% of reading that is on top of the uncertainty already so far in the process.

Chart showing the factors which introduce uncertainty into the BOD5 measurement
Figure 1: BOD5 uncertainty chart
BOD5 is globally recognised as notoriously inexact and if you ask any water quality specialist they will have different views on its accuracy; however, they will agree that the current 5-day BOD test is nothing more than an approximation that typically varies in accuracy from +/-15% to +/-50% of reading. So if its that bad why is it so widely used? Well in short there is no other practical way; to date the only alternative has been very expensive mini-laboratory styled analysers which require significant capital (ranging from £30,000-£100,000 each) or sending samples to a laboratory. Beyond that they also require regular maintenance, power and other resources and hence there has been no significant shift away from manual sampling. So how useful is a result that can only be used as an indication and that will take approximately 7 days to appear in your inbox? That part is easy; even if you could get a result within a few hours of an event it is often too late to allow a process to be changed or a pollution event contained. When you superimpose the inaccuracy of the BOD5 process then you may end up with a false positive or vice versa which typically leads to bad decisions being made based on bad data. There are obviously ways in which we can improve the BOD5 process but ultimately to get a reasonable result of +/-15 to +/-20% overall, the process above needs to be followed with a level of military precision which it rarely receives. Is there anything that can be done to improve the results beyond rigorous procedure? Well, by taking multiple samples (same sample sub-divided) and sending them to the same or different laboratories will provide multiple results that can serve to reduce the average mean error. The downside is that this almost doubles the laboratory costs which can vary between £5-25 per sample. Understandably not many people do this but more because they are not aware of the true uncertainty of BOD5 analysis.

REASONS FOR CONTINUOUS MONITORING OF BOD

The strengths of point sampling and continuous online monitoring have long been established for many parameters but there are several parameters including BOD/COD/DOC/TOC and faecal coliforms where technological boundaries have prevented their widespread adoption. There is a whole multitude of reasons for needing to measure BOD/COD/TOC continuously:
  • Resources: continuous monitoring is often required where there is a lack of resources and manpower to provide routine grab samples;
  • Responsiveness to Fluctuations: by detecting spikes or fluctuations in BOD it is possible to better respond to situations in sudden changes in oxygen demand whether that is in a WTP, process control or in the environment;
  • Compliance: Ensure that countries, environmental protection agencies and companies meet increasingly stringent legislation; many WTPs within Europe and in the developed world have to ensure they do not discharge any waters exceeding 15-20mg/l BOD. A BOD5 sample once per week will not ensure this!
  • Pollution Prevention: Real-time alert of pollution and indeed pollution prevention;
  • Real-Time Control: Continuous data can help identify where process conditions are distressed to either aid decision making or automatic control;
  • Process Optimisation: Monitoring & optimisation of Water Treatment Plants to improve performance and conformance; it is then potentially possible to directly match incoming load to WTP capacity;
  • High ROI: Cost vs ROI; once it is understood whether continuous monitoring is perceived as an investment or as a cost, continuous monitoring can pay for itself very quickly based on labour/resource savings, improved data quality and the better decisions that come with that plus the avoidance of any risk or fines that are associated with non-compliance.
  • Improved Accuracy & Reliability: Anybody with experience of BOD5 sampling will have a tale to tell where the returned results were implausible and led a series of questions or actions that led to nothing because there was nothing to explain the readings. This is an all-too-regular occurrence but the problem is that there has been very little choice and so there is no platform to challenge or have the readings verified. Hence you are put in the scenario that poor data leads to poor decisions…hence good data then becomes an asset; an asset based on which you can make good decisions or better still provide automation.
  • Cost/Energy Savings: – whether it’s a reduction in lost product or it’s an energy saving such as variable speed drives to control aerators at WTPs, continuous data can revolutionise process. For BOD this has never been an option (until now).
  • SMART Networks: by measuring BOD in real-time it is possible to join up and integrate with multiple systems together: sewer network, WTPs and water resources to provide a real-time overview of BOD.
  • Learned Behaviour: when only one or two samples are taken per week it is not possible to fully understand a process ie diurnal effects; by monitoring it continuously it is possible to how the system works with infinite detail. The data may reveal various data signatures which may indicate other issues.
  • Human Error: we’re not perfect but the reliance on getting the BOD sampling procedure perfect every time to provide reliable data is an under-statement.

WHAT CONTINUOUS MONITORING TECHNOLOGIES ARE AVAILABLE?

There are a number of different technologies available that claim to measure BOD but they typically are unable to measure BOD in real-time or unable to measure it accurately and reliably. Technologies include photometric, manometric, biosensors, microbial fuel cells & UV-VIS spectrometers. Until now UV-VIS technologies have been the most widely adopted sensor for indicating BOD in real-time but cleaning, path length selection and accuracy have all shadowed their reliability. There has been a notion for some time that fluorometers could be used to measure organic loading of waters but there has been little development or scientific research in this area. Proteus Instruments were the first to incorporate the learnings and apply the technology and pioneer the patented Proteus multiparameter sonde that can measure BOD/COD/DOC/TOC and coliforms in real-time based on the use of in-field fluorometers. The development of this instrument has highlighted the high uncertainty (>30%) in BOD5 measurement. Unlike other instruments, the Proteus is a sensor that can be installed in any water body, river or effluent channel and measure instantaneously and without the issues that plague UV-VIS technologies. With an integral cleaning system and lower power consumption, the range of applications for the Proteus is unparalleled.

BOD5 VS PROTEUS – AN EXAMPLE

A Proteus was installed at a final effluent discharge point for a major UK water company to measure BOD/COD/TOC in real-time with astounding results. The client arranged to take regular water samples for BOD5 to be analysed at their chosen laboratory. The BOD measurement range was relatively low varying between 1-10mg/l with a typical cyclic diurnal pattern of loading. When the results from the laboratory were returned there was no clear correlation between the samples and the Proteus readings highlighting the issues with laboratory and spot readings . On further investigation it is possible to see the marked variability in BOD between the laboratories and within a single laboratory (figure 2). It should be stated that all four samples came from one fully agitated sample (which was subsequently divided) which should have provided four identical laboratory BOD5 measurements.

Figure 2: Box and whisker plot of the four replicate BOD5 measurments for each bulk (2 L) sample collected. The coloured circles denote each sample replicate (i.e. lab or sample replicate within lab).
Figure 2: Box and whisker plot of the four replicate BOD5 measurments for each bulk (2 L) sample collected. The coloured circles denote each sample replicate (i.e. lab or sample replicate within lab).

When undertaking an optimised field calibration of the Proteus against a single set of samples (i.e. from a single lab or replicate number) the problem with laboratory uncertainty is clear with no significant calibration achieved (Figure 3). When a mean of all laboratory BOD5 samples was assessed a strong linear relationship between the Proteus and the laboratory data was observed (Figure 4). This equated to a minimal quantification error for the Proteus (Mean error = 0.07 mg/l). This is clearly a very strong argument for taking multiple samples and also for the use of continuous real-time measurement technology.

Figure 3: Calibration of the Proteus based on single sets of laboratory samples - i.e. those from one laboratory or replicate run.
Figure 3: Calibration of the Proteus based on single sets of laboratory samples - i.e. those from one laboratory or replicate run.
Figure 4: Mean laboratory BOD data with 95% CI (black dot and line) recorded for the six sample times and the corresponding Proteus BOD measurement (red triangle)
Figure 4: Mean laboratory BOD data with 95% CI (black dot and line) recorded for the six sample times and the corresponding Proteus BOD measurement (red triangle)

An example of the data from this installation can be seen below along with the captured single-point BOD5 sample readings. Note that these are not averaged readings and hence there are deviations between the Proteus and the laboratory readings. Obviously manual samples taken once per day or per week would not provide the level of detail that can be seen here. Real-time measurement of BOD will ensure that no unexpected events are missed and will allow organisations to react quickly to environmental pollution or process issues, long before they exceed threshold status.

lab-vs-proteus-datas

Summary

The main reason why we still use BOD5 as a technique is because it is (1) enshrined in environmental legislation and environmental/industrial monitoring practice but (2) also it has historically been the most accessible approach to measure BOD. The technique itself is widely regarded as being potentially very unreliable often causing concern or alarm when there is none due. Technological advancements such as the Proteus will allow the global community to monitor and control water quality accurately and in real-time. The step-change in technology has truly global implications for traditional BOD sampling and could serve as a dramatic step to improve the water quality of our water resources, protect our aquatic ecosystems and improve our drinking water supplies. Its easy to see that this technology could literally reinvent the way we monitor and control our environment; it will generate new questions and logically new answers. There is a very true saying that ‘good decisions are based on good data’; getting real-time BOD data will allow organisations and communities to make quick and better decisions which will ultimately improve our environment and our health.