The odour control system installed at Hallett Cove consists of a biological filter followed by an activated carbon filter to enable “polishing” of any residual odour remaining after the biological stage. SA water has asked Bioaction (the manufacturer of the system) to review the performance of this filter in an effort to determine the operating efficiency of the biological stage and report on any effects that the current performance would have on the lifetime of the secondary “polishing” activated carbon system.
SA water has provided data detailing, over a period of 19 days, the levels of H2S present in various stages of the filtration system. The data gathered took place over a summer period between 20/12/2016 and 9/01/2017 and as such includes both the Christmas and New Year’s break. All the supplied charts have been placed in Appendix A. It is typical of this period that a pump station will experience it highest loads and in turn any filtration system may also experience higher than average loading. The key data points from these charts has been summarised in the below table:
Activated Carbon Inlet
This data shows results that are typical of a dual stage system with the first stage biofilter removing a large percentage of the H2S and as a result reducing the average ppm(v) from 3 down to 0. However again in a typical result we can see here that the biofilter has allowed for a peak of 6ppm(v) to pass through. While the activated carbon filter removes the remaining H2S.
Biofilters require a population of biomass in order to treat an incoming contaminate, in this case H2S, in order to be effective. If the biomass has not built up to a level required for an incoming contaminate to be fully removed, then additional biomass is grown. Similarly, if there is not enough contaminate to sustain the biomass it will die back. Due to the time dependent nature of biomass growth biofilters in their nature are slow reacting and as a result generally do not handle “peaks” in the incoming contaminates as effectively as chemical and adsorption based filters.
It can also be seen in this data that the activated carbon vessel was effectively able to remove any additional contaminates remaining after the biofiltration stage. The 2.2 ppm(v) peak appears only during the initial 10min of Odalog data during a time which both the biofilter inlet and the carbon inlet are both reading 0ppm(v). The reason for this is unknown.
Base on the provided data it is possible to calculate the percentage reduction of H2S across the biofiltration stage. This data was then used to produce a plot showing the removal efficiency of the biofilter over time. However due to the rapid changing of data the results are noisy and as such a noise filter utilising the moving average technique was implemented. A moving average filter smooths data by replacing each data point with the average of the neighbouring data points defined within a span. In this case a span of 20 has been used for all instances of filtration. A comparison of various spans has also been shown below for reference. In this comparison, it can be clearly seen that as the span increases there is a point where there can be a significant loss in data beyond just noise. The introduction of this filter allows for a more visual determination of the trends that occur within the data.
Across the board the filtered data clearly shows that the lower removal efficiencies are very sudden and do not last for a significant period of time.
The calculation of the removal efficiency uses the following equation:
Where C_i is the concentration in while C_o is the concentration at the outlet. In this case, we can see that the lowest removal efficiency occurs at the highest peak with an efficiency of 60%. In addition, there is an overall average of 98.1% removal efficiency during the period of testing.
One of the issues most evident throughout this data is the resolution of the instruments used is limited to ±1 ppm(v) which when reading the low values shown in this data there is a large number of 0 ppm(v) readings that could potentially skew the averages.
Based on the supplied data there is an average of 0.1495 ppm(v) H2S at the inlet to the activated carbon filter. This number is low enough that based on mass flow calculations of H2S entering the system at a flow rate of 666L/s an expectation can be made for a carbon lifetime of 50.5 years. This calculation however only accounts for the presence of H2S and there are other compounds and contaminates that will certainly affect and reduce this lifetime.
Based on the provided data a conclusion can be made that the biofilter is operating at a reasonable efficiency achieving an average removal efficiency of 98.1% with a lowest efficiency of 60% occurring during the highest peak during the sample period. This performance is typical of biofilters due to their slow reacting nature and as such it can be said that this biofilter if performing sufficiently.
A conclusion can also be made that the activated carbon filter is able to remove all the remaining H2S contaminates with a resulting reading of 0 ppm(v) at the vent stack for the entire period of testing. In addition, the carbon lifetime should not have any appreciable reduction as a result of “peaks” in the incoming air stream as the average H2S loading is only 0.1495 ppm(v).