NL8006289A - METHOD FOR PURIFYING AN ORGANIC SUBSTANCES CONTAINING WASTE WATER FLOW. - Google Patents

Description

Method for purifying an organic matter-containing waste water stream

The invention relates to the operation and maintenance of a recirculating methane-developing anaerobic filter of the type described in German patent application p27 ^ 831.32, made available in April 5, 1978 and the corresponding US patent application no. 57-5 ^ 5, filed July 13, 1979-

As is known in the wastewater treatment art, the action of methane-developing anaerobic filters relies on two types of bacteria living in symbiosis. The "acid formers" break down large molecules into molecules of 1-3 carbon atoms (acetic acid, propionic acid, etc.) which are mostly acidic in nature; they are fairly resistant to variations in pH, temperature and nutrient concentration. The "methane formers" digest the acids and small molecules by dismutation; one carbon atom is completely oxidized to CO 2 and a second carbon atom is reduced to methane. Carbon dioxide is partly soluble in water and part of the CO 2 remains bound to the alkaline compounds present (alkali time) while the rest escapes to the gas phase. Methane is practically insoluble and is completely removed in the gas phase. It has been found in the literature that the relative amounts of carbon dioxide and methane in the evolved gas can vary during the process. For example, Jennett and Dennis, in "Anaerobic filter treatment of pharmaceutical waste", Journal WPCF U7, 10-121 (1975) (in FIGS. 9 and 10), report methane contents of about 60 to 85 #; Youngen McCarty in "The Anaerobic Filter for Waste Treatment", Journal WPCF Vl_, R160-173 (1969) report methane levels in the range of about 70 to 80%. The methane content of the gas depends on the nature of the feed material; for example, the theoretical CH 2: CO 2 ratio in the products of the process is 1: 3 in the case of formic acid, 1: 1 in the case of acetic acid and 3: 1 in the case of methanol.

The action of anaerobic filters is often described in terms of the "organic load" (hereinafter 5 "OB") of the filter, ie the rate at which COD

(chemical oxygen demand) per unit volume of the filter is supplied. That load can be expressed as "lbs COD / cubic feet / day or in kg COD / cubic meters / day; 0.5 lb COD / cubic feet / day equals 8 kg COD / cubic meters / 10 day.

As described in the above-mentioned German Patent Application p27 ^ 831.32, the recirculating anaerobic filter is operated at a load (OB) of generally more than 3.2 (and preferably at least 8) kg COD / cubic foot / day 15 amounts.

The figures illustrate some aspects of the invention.

Fig. 1 schematically depicts an anaerobic filter and control devices for that filter.

Figures 2, 3 and ^ show graphs of the OB and KB values and of the pH values (measured once a day) over a number of work periods. Figures 2 and 3 relate to Example I (Figure 3 is a continuation of Figure 2). Fig. k relates to example IV.

FIG. 5 shows a collection of graphs showing the changes in the amount of eluent gas, the percentage of methane in the eluent gas, the production of methane and the methane efficiency over a period of several hours in response to a change in the feed rate.

FIG. 6 is a graph of the variations in the OB and KB in response to temperature changes in an anaerobic filter controlled to operate with a methane efficiency of 85%.

Fig. 7, 7A, 8 and 9 show registration cards 35 of the OB and KB values which are automatically recorded each time with an 8006289 3 interval of 1h. Figures 7 and 7A relate to a period of about 3i day until the end of the test of Example IV (Figure 7A is the continuation of Figure 7). Figures 8 and 9 apply for a two-day period 5 during the tests of Examples II and 5, respectively. V.

Figures 7 to 9 show the instantaneous values of the QB measured at intervals of 1h and the varying mean values of KB (for the previous hour) also measured at intervals of 1h 20 minutes later than the 0B 10 values. For example, if a current 0B value is recorded at 8:00 am, the varying average KB value (averaged over the period of 7-20h to 8:20 am) is recorded at 8:20 am; at 9.00h a current 0B value is registered again and at 9.20h the varying average KB 15 value (average over the period from 8.20h to 9.20h), etc.

In some cases the indications "out of scale" are shown in figures 7A and 9. This, of course, means that the values went beyond the upper limit of the scale of the map in question and that the following values must be read 20 in a new scale, which is also indicated, (in Fig. 7A there are two recorded points lying between 8 and 10 o'clock in the afternoon, where it is not clear which scale applies to the points in question).

It has been found that when operating a recirculating methane-developing anaerobic filter, the rate at which methane is formed is very sensitive to and reacts quickly to even small changes in conditions. Fig. 5 illustrates some features of the response obtained with an open volume recirculating anaerobic filter of 201 operating at a recirculation rate of 12 l / h. Feed rate to this filter was initially about 0.28 L / h from a feed containing about 25 g / L COD; this corresponded to a 0B of approximately 8.3 kg COD / cubic meter / day and a residence time of approximately 71.5 h. Initially, the methane efficiency was about 6% and the COD concentration in the mixture 8006289 h (from fresh feed and recirculation stream) fed to the base of the filter was about 6.0 g / l COD. The feed rate was then suddenly greatly changed so that the OB was increased to approximately 9.9 kg COD / cubic meter / day; this caused the COD concentration in the mixture fed to the filter to change from the initial value of about 6tk g / l COD to a value of about 6.5g / COD just after the change in the feed rate. From FIG. 5, it can be seen that this change in concentration initially less than 2% caused a significant and practically immediate change in the rate at which methane was formed.

It has been found useful to express the rate of methane generation in the weight of COD removed as methane per unit volume of the filter per day (hereinafter referred to as "KB" or "kinetic load"). , which, like the OB, is expressed in kg COD / cubic meter / day); one kg of methane and COD of U kg. Then KB: 0B indicates the methane efficiency of the filter.

As can be seen from Figure 5, the methane efficiency calculated in this way immediately fell when the change was made and then rose again; this was, at least in part, a continuation of the calculation method, because the change directly resulted in a new, 20% higher OB (the denominator in the calculation of the efficiency) while the actual direct increase in the COD concentration in the combined feed flow to the filter was less than 5%, as noted above.

According to one aspect of the invention, the methane efficiency of the filter is continuously monitored and the load (OB) is continuously modified (eg increased) to keep that efficiency at a significantly lower value than the anaerobic filter can achieve without a such increase in tax. According to a preferred embodiment, these load changes are effected at times that are less than Hh apart, and preferably less than 2h apart, for example, 0.5 or 1/12 h apart; the changes in the OB, when made so frequent, are generally each small (for example, less than 10%, but, as shown in Figures 6 and 9, 5, the changes are sometimes such that in a system where the OB is changed every 6 minutes, the OB can be approximately doubled in an hour, especially if the OB is increased from a relatively low value of, for example, 0.¼). By working in this way, one is able to increase the OB of anaero-10 be filters to previously unheard of high values higher than 32 kg COD / cubic meter / day (for example to values of approximately 48 to 64 kg COD / cubic meter / day or more).

We have also succeeded in producing very rapid increases in the load of anaerobic filters that operated under / relatively low loads (for example from 3.2 to 4.8 kg COD / cubic meter / day), which makes it possible over time that is required to bring a newly installed filter (or a previously damaged or deregulated filter) to a desired relatively high capacity.

By applying the new method according to the invention it is possible to raise the filter relatively quickly to previously unknown conditions in which the volume of biomass in the filter takes up a large part of the open volume of the filter. Previously, the known filters contain only an under-appropriate (volume) amount of biomass; For example, the aforementioned U.S. Patent Application 57.5 ^ 5 of July 13, 1979 states: "In the operation of the anaerobic filters it has been found that the amount of biomass adhering to the packing, 30 even after extended operation, is such as to occupy only a small fraction of the void volume of the filter, eg some 20% or less of the void volume (this being measured by allowing the contents of the filter to drain out and measuring the volume of the liquid thus removed), even though there is a significant recycle 35 of biomass to the filter, While the reasons for this are not 8006289 6 clearly understood, it is believed that the recycled bacterial serve, in part, as food for the bacteria inlhe filter. after lengthy (eg 6 months or more) operation under recycle that the biomass attached to 5 the filter packing is distributed on the packing substantially throughout the filter, visually this distribution appears to be substantially uniform from top to bottom »The biomass is gelatinous or slimy ..."

In an article entitled "The Anaerobic Filter 10 for the Treatment of Brewery Press Liquar Waste" in "The Brewer

Digest February 1972, biz. 66-73 Lovan and Foree hold a line of reasoning along a corresponding line; this article describes spheres with anaerobic filters with a height of 1.8 m with a volume without packing of 33.1 * 1 and with an open volume (in packed condition) of 15.2 1; On page 72 it states: "After 100 days of operation significant biological solids had accumulated to a depth of about 12 inches in Filter 1 and about 2k inches in Filter 2".

In contrast, the Applicant operated anaerobic filters under conditions such that, upon draining, the amount of free (drained) liquid was clearly less than 70% of the open volume, for example 10, 20, 30 or 50% of the open volume which also included filters with such a high biomass content that after draining it appeared that the packing rings (and even the recirculation pipes) were almost completely filled with a gelatinous biomass. It would have been expected that such filters would not work properly because the streaming of liquid through the filter will be greatly impeded, leading to a reduced capacity, but that expectation has proven to be incorrect because these filters work very well with little back pressure, little biomass is flushed out, if there is any flushing out and with an unusually high load. These 35 filters are also extremely resistant to damage causing disturbances which would render ordinary anaerobic filters inoperative. The reasons for this are not clear. It is believed that the large bodies of biomass can be highly permeable to the organic compounds in the water and this may be due to the presence of a gelled secretion surrounding the individual bacteria or groups (eg layers) of bacteria, which secretion is permeable to those organic compounds, so that in any case the bacteria within the biomass bodies have access to the food that is formed by those (organic) compounds. The way in which the liquid moves through such filters is not entirely clear. It is possible that it passes through a network of cracks and channels. Upward flow of liquid is enhanced by the percolation effect of the relatively strong development of gas bubbles in the filter and it appears that a high degree of back-mixing occurs in the filter. For example, in a test using a tracer using a small amount of Indian ink injected into the liquid, it was (at a rate of about 2 ml / min.) At the bottom of a cylindrical recirculation filter (15 cm in diameter, an open volume of about 20 liters and a height of 1.2 m) was introduced, the effluent coming out of the top of the filter about 10 min. after the injection (ie after a total flow of about 2, hi) dark colored and pitch black when the effluent was viewed 15 min after injection (after a total flow of about 3.6 L). The flow rate of approximately 2ho ml / min through the filter (cross-sectional area approximately 180 cm) corresponded to an average flow rate (over the entire cross-section) of only 1.33 cm / min, but the ink reached the top with a speed of more than 12 cm / min. It is possible that, despite the fact that the structure of the gelatinous biomass is strong enough to bridge the openings of Prall rings with a diameter of 2.5 cm and despite the very low pressures in the anaerobic filter, the biomass may be distributed 8006289 8 into smaller flowable aggregates under the influence of the flow of liquid and gas during operation of these high biomass filters.

5 An article by Bellegem et al on "The

Anaerobic treatment of waste water of the potato starch industry "in the December 197 issue of TNO project mentions that:" It is clear that with an anaerobic filter the major object is to reach a concentration of biomass that is as 10 high as possible because this will increase the stability, the efficiency and the capacity of the system ".

Speaking of the highest achievable capacity, these authors say: "In our experiments it appeared that a load

O

15 of about 8kg COD / m / day could be maintained continuously, but that higher loads can lead to process instability and failure ". This OB of 8 kg COD / m / day is much smaller than the OB values' that are applied when applying the invention.

It seems likely that in the large bodies of biomass attached to the packing, a concentration gradient of the components of the nutrient solution occurs from the outside of each of those bodies of biomass to the interior thereof. This may be due to diffusion through a gelatinous separation that is permeable to such components. The outside of any of such bodies may be in contact with the relatively concentrated mixture of recycled material and fresh food; if the concentration in that mixture reaches an inhibitory or toxic value, the bacteria on the outside can be adversely affected by exposure and that concentration, while bacteria inside the biomass bodies can act more or less as before because the concentrations they are exposed to are significantly lower than those on the outside. This can help to explain the increased resistance of the filters to disturbances.

8006289 i "9

It is possible that the rapid increase in load observed when the invention is applied can be explained as follows. The amount of "nutrition" for the bacteria is maintained at a controlled (controlled) high concentration that promotes the growth of the bacterial population, but is always kept below an inhibitory concentration. (The inhibitory concentration depends on the compounds which feed; methanol does not inhibit before its concentration is relatively high; the inhibitory concentration is lower for acetic acid than for methanol; for acrylic acid, the inhibitory concentration is much lower). Fig. 2 (considered with Example I) indicates that with a feed containing 15 g COD / 1, growth was promoted if the efficiency was set to 87 a 88 Ji, but was not promoted for a period of H days, when the efficiency was set at 92%. At the efficiency of 88%, with this feed, the COD concentration prevailing in most of the recirculation fliter will be approximately 1.8 g COD / 1 (i.e. 12% of 15 g COD / 1) while the efficiency of 92% they will be only about 1.2 g COD / 1. At an efficiency of 80% tzalt as illustrated in Fig. 3, the COD concentration in the filter will be even higher (i.e. about 3 g COD / 1 for the feed of 15 g COD / 1 and about 5 g COD / 1 for a feed with 25 g COD / 1), but it was still not large enough to cause this feed (containing formaldehyde and acrylic acid in significant amounts) to have a significant inhibitory effect on the action of the large amount of biomass in the filter, as can be seen from the results shown in Figure 3.

The examples given below serve to illustrate the invention.

In Examples I - V, the anaerobic filter consisted of a circular cylindrical corrosion resistant tray 5 (Fig. 1). The container was 1.2 n high and had a diameter of 15 cm and was filled with plastic rings randomly poured into the container 8 0 06 28 9 10 (Pall rings, rings for biological applications from plastic with a diameter of 2, 5 cm and a height of 2 3 2.5 cm). This packing gave a specific surface area of 16U m / m and an open space of 90%. The packing rings consisted of polypropylene with a density of about 0.9, so that they tended to float in the liquid in the tray and butt against the top wall of the tray, so that a free, non-packing rings filled space at the bottom of the bin; a horizontal grid was located 5 cm above the bottom of the tray 5; the 10 zone below the grille contained no gasket. The total volume of the box was 22.2 L without packing; with the gasket in it, the free volume was 20 1. The acidic wastewater supplied to the anaerobic filter was fed from a hopper of raw wastewater by means of a positive displacement pump (equipped with an adjustable speed motor) via a line 9 into a stream of recirculation liquid with a bacterial solid material introduced therein, which was contained in the recirculation loop 10 and came from a separating vessel at the top of the filter and was circulated by means of a recirculation pump 12. The combined feed stream and recirculation latent stream entered the bottom of the tray 5 under the horizontal grating through a central opening with a diameter of 16 mm.

Just upstream of the hopper supply point was a non-return valve (not shown) which acted as safety 25 to prevent the contents of the hopper 5 from draining in the event of a pipe breaking.

Liquid and gas escaped from the top of the tank 5 (through a central tube with a diameter of 16 mm) and flowed to a separating vessel 11 in which an amount of practically stationary liquid 1¾ was present, the level of which was kept practically constant by means of an overflow control in the form of an open siphon that opens into an overflow pipe 18. The mixture of liquid and gas flowed upwards from the tank 5 through a pipe 19, the outlet of which was located above the surface 16 (approximately 2.5 to 5 cm above the 8 0 06 28 9 11 level 16). This liquid and gas mixture also contained solid biomaterial, mainly in finely divided suspended form, but also contained somewhat larger particles of solid biomaterial which may be loosely attached to gas bubbles.

When exiting narrow tube 19 (a circular tube with a diameter of 16 mm), those bubbles tended to detach from these larger particles (which are rounded floccular or spheroidal with a diameter of about 1 to 5 mm). Those particles then settled in the liquid mass 1k and were discharged with a liquid from the bottom of the vessel 11 and returned to the bin 5 via the recirculation cycle 10. The excess (non-recirculated effluent) discharged via the overflow 17 was often slightly cloudy. because it contained finely divided biomaterial (e.g., about 50 to 500 mg VSS / 1, usually about 200 to 300 mg / 1, such as when filtering the liquid through a 0.5 µm pore membrane, which individual organisms). The settling vessel 11 was sized such that the average residence time in that vessel (i.e. the flow rate of 20 fluid and solid from line 19 divided by the volume of the fluid mass I) was clearly less than 0.2 h, e.g. 6 min. (This residence time was so short that the settled biomass particles have practically no tendency to generate sufficient gas to drive much of it up and make it disappear via the overflow). The gas evaded a pipe 2b at the top of the vessel 11.

In the recycle circuit 10, a heater 26 was provided to heat the recycle material to the optimum temperature.

In the examples, the vessel 11 had an inverted conical shape and contained liquid up to a height of 20 cm with a free gas space of 7.5 cm above it. The surface of the liquid 16 was 30 cm above the tip of the inverted conical trough 5. The open space above the surface 16 was under atmospheric pressure or slightly higher, 8006289 * 12

The volume rate of the feed through pump 8 (a positive displacement pump) was measured continuously (by means of a tachometer giving an electrical signal proportional to the pump speed and thus proportional to the liquid flow). The evasive gas leaving the installation through line 2h flowed through a conventional liquid and mist separator 26 and was then continuously analyzed by passing it through a standard infrared methane analyzer 27 and then through a wet gas meter 28.

The methane analyzer was a Beckman infrared analyzer (Model 865) as is well known in the art using a methane-filled detector. The instrument was regularly calibrated so that the electrical signal delivered was directly proportional to the methane concentration of the evasive gas.

The wet gas meter 28 was provided with an optical detector and a 250-tooth gear which gave 250 electrical pulses with every liter of revolution of the meter.

The "output" from the wet gas meter and from the 20 CH 2 analyzer were (by means of suitable instruments) combined into an electrical signal indicating the rate of methane evolution (and thus indicating the "kinetic load" or "KB"); This rate can be expressed in the number of kg COD in the form of methane per unit time per unit volume of the filter (1 kg methane has a COD of ^ kg).

The "output" signal for the pump speed of the feed was combined with the known (or regularly measured) COD value of the feed to an electrical signal 30 which indicates the COD feed speed and thus the 0B.

The motor driving the positive displacement pump for the power supply 8 was controlled by a controller comprising an arithmetic unit or microprocessor (described below) to maintain a predetermined 800628913 relationship between the OB and the KB.

Further measuring instruments included thermocouples for measuring the temperatures (a) in the geometric center of the filter and (b) in the material being recycled, as well as a pH meter for measuring the pH of the recycled liquid just upstream from the point where the fresh food was added.

Example I

The acidic wastewater used in this example was an aqueous solution containing a mixture of 375 parts by weight of acetic acid 37.5 parts by weight of formaldehyde 115 parts by weight of formic acid 15 parts by weight of butyric acid 30 parts by weight of acrylic acid

At the start of the test, the concentrations of these components were such that the COD of the water was 15 g / l, which means that the concentrations were: 20 acetic acid 9.38 g / l COD 10 g / l 100% formaldehyde 0, 9 ^ g / K ^ COD 1 g / 1 formic acid 2.88 g / 1 'C5COD 1 g / 1 butyric acid 1.10 g / 1 COD 2 g / 1 acrylic acid 0.75 g / 1 ^ COD 1 g / 1 25 Later in the experiment, the feed was made more concentrated (without changing the ratio of the ingredients) so that the COD was 25 g / 1, meaning the concentrations were: acetic acid 15.83 g / 1 ^ COD 16.67 g / 1 30 100% formaldehyde 1.56 g / 1COD 1.67 g / 1 formic acid ^, 79 g / l-oCOD 1.67 g / 1 butyric acid 1.8 g / 1 "O COD 3.33 g / 1 acrylic acid 1, 25 g / 1 COD 1.67 g / l

Furthermore, the food mixture contained 1 water per 2b. 35 6.3 g urea, 2.2 g 85% phosphoric acid (the concentrations were} 8006289 11 * such that the COD: N: P ratio was 1000: 5: 1), 2k ml of an aqueous solution ie iron ( ll) sulfate and cobalt (II) sulfate (1 mg / 1 Fe and 1 mg / 1 Co) and 2k ml of an aqueous solution of sodium sulfate (S concentration in that solution 10 mg / l) and 8 g / 1 Na HCO ^. These food mixtures were acidic; the pHs were likely close to 1 * to 5.

At the beginning of this test, the filter had already been started up and had been working for 2 \ months with the same feed mixture (COD 15 g / l) with the feed rate of the acidic waste water being such that a load (OB) was obtained from

D

8 kg COD / m day (calculated on the basis of the open volume of the 20 l container).

The progress of the test is shown graphically in Figures 2 and 3.

At the start of the test, the controller became so

set to obtain a practically constant difference ("Δ B") of approximately 1.28 kg COD / m day between 0B and KB. That is, if KB + 1.28> 0B, the controller increased the feed rate to 0B to maintain the desired difference 20. Conversely, at the start of the experiment, the number 1.28 corresponded to an efficiency of approximately 86% (that KB

i.e. ^ = 0.86), calculated from the value of OB at the beginning of the test. Over a period of 16 days of operation under these conditions, the KB rose to practically 25 9.6 and the efficiency increased to approximately 89%. In the following days, the KB (and of course the 0B) decreased slightly.

The controller was then adjusted (point A of the graph) 3 to obtain a Δ B of 1.92 kg / ra day.

Initially this corresponded to an efficiency of 83%.

30 As can be seen from the graph, the KB had risen to approximately 1l * U within 11 days of this change and the efficiency had increased to 87%.

The controller was then (point B of the graph)

O

adjusted again to obtain a ΔΒ of 1,28 (kg COD / nr 35 day), corresponding to that point of an efficiency of 92%.

80 06 28 9 15

As the graph shows, the KB remained practically constant for the next H day period.

Then the controller (point C of the graph) was set to a Δ B of 2.56 kg COD / no day) corresponding to 5 at that point with an efficiency of approximately 86%. Within a few days this led to a clear increase in the KB to approximately 16 kg COD / m2 day.

In point D the controller was set again so that

O

a Δ B was obtained from 1.5 kg COD / nr day increasing the efficiency to 92% and keeping the KB practically constant.

In point E the controller was reset to a Δ B of

O

2, k (kg COD / m day), corresponding to an efficiency of approximately 88% \ within a few days, the KB had increased to more than 1 and the efficiency was approximately 90 1.

In point F (fig. 3) the regulator was set to obtain an efficiency of approximately 80%. that is, the ΚΒ.ΌΒ ratio was kept at about 80% (with Δ B allowed to vary accordingly, i.e., Ü = ° E = 0.80; so that ΔΒ = 0.20 x 0B).

20

In point G the feed was made more concentrated i.e. the fresh feed fed to the filter was now the feed solution which had a COD of 25 g / l instead of the pre-used solution with a COD

of 15 g / 1. Efficiency was maintained at 80%. Within 25. ...

less than a month after this direction was set to an efficiency of 80%, the KB had risen to practically 32 (kg C0D /

O

no day) and the 0B was over 38, U; these were values never reached before.

During the run, the recirculation pump was driven at a practically constant speed to provide 12 1 / h of recycled material. Since the rate at which fresh feed was fed during this period varied, the recirculation ratio varied in the following manner: 35 80 06 28 9 16 feed rate recirculation ml / h ratio early February 480 25: 1 point A * 620 19.4 : 1 5 "W * 889 13.5: 1" C * 940 12.8: 1 "D * 9éT 12, U: 1" E * 1030 11.7: 1 "F * 1329 9: 1 10" G * 994.3 (richer) 12.1: 1 end * 1286 (richer) 9.3: 1

The average upward flow rate (over the entire cross-section of the 2 180 cm 'filter) can be easily calculated from the feed rates and the recirculation rate of 12 l / h; in all cases she was little more than. 1 cm / min.

Fig. 2 shows that if ΔΒ was too low (or if the set efficiency was too high) no growth occurred.

In Fig. 3, ΔΒ was above that low level and gave an unprecedented 20 growth to a 0B of 38.4 and a KB of practically 32.

Example II

The test described in Example I was then varied by feeding another nutrient solution to the same device containing the same biomass. This novel nutrient solution contained the water-diluted waste effluent from an acrolein, acrylic acid and acrylate ester preparation plant. From time to time, the concentration of the feed varied between the COD of about 25 and 35 g / l ^ rv J on average, the concentration (as COD) was 30 g / l, of which roughly a COD of 10 g / l came from acrylic acid and a COD of 8-10 g / l originated from acetic acid while the other components were ethanol, ethyl acetate, ethyl acrylate, diethyl ether (low content) butanol, butyl acetate, sec-butyl acrylate, n-butyl acrylate and some low acrylate polymers.

i 80 06 28 9 17

In the course of the test described in this example, the efficiencies were set to different values between roughly 65 and 75%, whereby OB values were obtained

O

of more than 32 (for example Ho) kg COD / m3 daily KB values

O O

5 of 28.8 kg COD / m-3 day (in practice 30, ¼ kg COD / m3 day or more).

During one period, the regulator was set to decrease OB to about 3.2 kg COD / m day. The nutrient solution used during this period was found to contain 1 q a significant amount of a toxic component (chloraldehyde). The filter was then flushed with water containing 2 g / l ITaHCO 3 and then a nutrient solution practically free of said toxic component was fed to the filter at a fixed rate such that an OB

O

15 obtained from approximately 11.8 (kg COD / m day) over a period of 5 days, during which time the KB increased from 8.3 to 9.6 kg

O

COD / m day (indicating a gradual increase in efficiency to around 80%). Thereafter, the feed rate was controlled with an automatic controller, increasing efficiency

O

20 put at 75% \ the KB rose to more than 16 kg COD / m-3 day and the

O

OB increased to more than 20.8 kg COD / m-3 day in about 16 hours as shown in Figure 8.

In the course of the test described in this example, the KB rose at one point to 38.¼ and the OB to 25 56lig COD / m day. Some mechanical difficulties then arose (described below) and after a few more weeks the bin 5 was drained. Only 1.8 L of water came out of the branch, although the original empty volume was 20 L.

The draining 1.8 L of liquid was collected 3q (in an open vessel free of air) for use in Example III.

This liquid contained a significant amount of suspended biomass. Pouring out the packing from the tray taught that the rings and the recirculation pipes and other parts of the plant were almost completely filled with biomass. The original total weight of the packing rings was about 2 kg, while the weight of the rings containing biomass from the tray was 16.6 kg, from which it follows that the rings contain 1U, 6 kg of biomass. Drying the biomass at 115 ° C, a treatment that removes the "" free "" or extracellular moisture surrounding the bacterial cells, taught that this "free water" constituted approximately 89.5% of the biomass attached to the rings . The weight of the dried biomass (dried at 115 ° C was 1.53 kg; calcination (at 600 ° C) gave an ash amount of 23.1% of the dry weight; qualitative analysis of the ash (by means of atomic absorption) taught that iron and small amounts of copper and zinc were present.

The rings on which biomass was attached floated when placed in water; an average density was slightly less than that of water. The biomass could be easily removed from the rings, for example by keeping it moving in water.

Regarding the aforementioned mechanical difficulties, after the KB had risen to 38.1 * and the OB to 56 (kg COD / m 2 day) ruptured by a mechanical effect a drain line so that all the liquid drained from the filter. The filter was refilled and within a few days it worked with

O

a KB of 28.8 and an OB of 1 * 3.2 (kg COD / πτ day) (the efficiency was set at approximately 66 ί). The KB then began to decline with a corresponding reduction in the DB, so that after a few weeks the KB was approximately 19.2 and the DB was approximately 28.8 (kg COD / m ^ day). Inspection learned that an outlet from the wet gas meter had become somewhat clogged so that that meter gave gas development rate readings that were significantly lower than actual values; As a result, the controller had decreased the feed rate accordingly. At that time, the tray was drained in the manner described above.

It is noted that in this example the operation of the regulator protected the filter from the input of 3 80 06 28 9 19 an unforeseen high toxicity of the power supply and allowed the filter to remain in a state where it could resume operation and quickly reached a high load with a feed containing no dangerous levels of toxic components.

Example III

After the test of Example II, practically all the biomass was removed from the rings, which were then returned to the bin 5. The trough was filled with 1.8 L of the 10 drained liquid (see Example II) together with water containing (calculated on the total refill) 50 g of methanol and 30 g of sodium acetate, after which the plant was operated with a recirculation of 100 % without fresh feed being added until a working temperature of 37 ° C was reached.

Then, a fed feed consisting of a mixture of 90 parts of water and 10 parts of the aforementioned diluted effluent with a COD of 30 g / l (as used in Example II) was supplied for 2 days. Then, a similar mixture was fed in a ratio of 80:10 for 1 day. And then the fresh feed was changed to consist exclusively of an aforementioned diluted effluent with a COD of 30 g / l (containing 8 g / l of supplied NaHCO2). The controller was set to obtain an efficiency of about 70%, starting with an initial feed rate corresponding to an OB of about k (kg COD / m ^ day). Within 3 weeks, the 0B had risen to approximately H6, U and the KB to approximately 32 kg

O

COD / m day. The efficiency was then adjusted to 75% for 5 days and then to 80% for 1 day and to 85% for another day; the KB was still approx

O

30 32 kg COD / m day. The efficiency was then set to 90% o and the KB decreased to approximately 28.8 kg COD / m day. The installation was then drained; 12.9 l of liquid were obtained (compared to a total open volume of 20 l in the absence of biomass).

In this example, the use of the regulator 8006289 20, when releasing a relatively small amount of biomass acclimatized to the same type of feed, caused a rapid increase in the KB to a value of about 32 kg COD / m day. and then brought about a rapid 5 increase in efficiency while maintaining this high KB value.

Example IV

After draining referred to at the end of Example III, the plant was again filled with water containing 2 g / l NaHCO2 and a 100% recirculation was maintained overnight (without feed) at a temperature of 3T ° C reach. It will be clear that the anaerobic biomass in the installation was exposed to air during the period (approximately one hour) between the draining and the refilling. The fresh feed was then changed to an aqueous mixture, each of four different organic compounds, namely Hik n-butanol, sec-butanol, n-butyraldehyde and n-butyric acid, each making the same contribution to the CODgpven. This mixture was diluted with water as needed to provide a feed stream at a concentration (expressed in COD g / l) as noted below. For example, to obtain a nutritional strength of (COD) 20 g / l, the mixture contained 1.93 g / l n-butanol; 1.93 g / l sec-butanol; 2.05 g / l n-butyraldehyde; 2.75 g / l n-butyric acid, which concentrations for each 25 component contributed 5 g / l to the CO2. The fresh feed also contained U g / l NaHCO3. Pig. H graphs the results. It is seen that within a month the KB increased to 3 3 more than 51.2 kg COD / nr day while the 0B was approximately 6U kg COD / nr day.

30 8 0 06 28 9 21 day conc set feed efficiency% 1 2.0 10 2 2.0 50 5 3 5.0 65 U 5 65 5 10 50 6, 7, 8 9 10 ko 10 10 10 65 11 20, T0 1U 20 70 15 30 75 16 30 75 15 17 30 75 18 30 75 21 30 80 22 30 80 23 30 80 20 2k 30 80 25 30 80 28 30 80 29 30 80 30 30 80 25 31 30 80 35 30 80 36 30 80 37 30 80

Notes: 30 On day 15, a single amount of activated sludge (from an aerobic process) was added to the feed to provide an additional bacterial population.

The figures in this table and the OB, KB and pH values shown in Figures 2, 3 and h are not mean values for the settings for the respective days. The numbers are 8 0 06 28 9 22 once per day (usually at 8:00 AM); the KB values are averages over 1 hour, as set out below.

There were some accidents during this 5 period. On day 5, it was found that the wet gas meter had leaked, so the information about the gas production rate (and thus KB) supplied to the regulator was less than the real speed.

On days 22 and 25, it appeared that the instrument panel clock (described below) had not worked properly, so that the information about KB supplied to the controller was again lower than the real value; the effect of this can be seen in the valleys of the OB and KB curves of fig. 1 * at 22-2b. On day 37, the * 3 filter operated with an OB of 62, h kg COD / m day with the set efficiency 80%; the efficiency setting was increased from 15 to about 85%. As shown in Fig. 7 at "X", this meant

O

that the OB had fallen to approximately 59.2 kg COD / m day which in turn caused a decrease in the KB and the OB and KB continued to decrease after one day of operation of the installation at 85% set efficiency, the efficiency setting 20 again at 80% and the KB (and OB) increased at a very high rate at which values may be reached even higher than will follow from the slope of the OB and KB curves, if the efficiency is at 80% would have been kept without the 1 day break with an efficiency of 85% · At that time, the feed line became clogged and the feed pump forced air into the filter for about 11 h; during this period the automatic controller operated so that the OB became

O

reduced to about 20.8 kg COD / mJ day in 1h, to about ll *, h kg 3 COD / m day in the next hour and eventually to about

O

30 1 *, 5 kg COD / m day. After the defect had been repaired, the automatic controller worked so that the OB and KB rose rapidly; within 12 hours the KB had risen to more than 38, H and the OB to 1 * 9.6 (kg COD / m ^ day). The behavior of the filter during this period is shown in Fig. 7A.

In this example IV, the filter (which still contained a relatively large amount of biomass that had been exposed to air) was subjected to various feeds; the use of the regulator made it possible to achieve a very high load in a short time.

5 Example V

Example IV was continued, reducing only the NaHCO 2 content in the feed to 2 g / l (from U g / l in example IV). The OB at the start was 62, k and

O

KB was approximately b9.6 (kg COD / m day). Efficiency was adjusted to 80% for 1 day, then to 81% for 2 days, then to 83% for another 3 days, at the end of which period the DB was approximately 62, k and the KB more than 52 8, even though the pH of the stream coming out of the filter and of the recirculation flow had dropped from 7.1 to 6, k. Thereafter, the KB decreased 15 and the efficiency was adjusted to 80% so that within approximately 2 more weeks the KB reached approximately 33.6 and the OB decreased to approximately 41.6 while the pH had decreased to approximately 6. Due to problems with the equipment was turned off for about 3 days while the recirculation flow was maintained. The automatic controller was then recommissioned to control the feed rate, first setting an efficiency of 50% for several hours and then an efficiency of 81% and then 70% as shown in Figure 9, the KB and OB increased quickly and (after 34 hours) 25 she reached a value of approximately 33.6 resp. b6, h (kg COD / nr day) even though the pH was still low, approximately 6.

Example VI

The "overflow" of the filter with the tests described in Examples II and III was fed, without further treatment (other than those listed below), to a second anaerobic filter operating without control (except those listed below). This second filter was the same size and shape as the first. The concentration of the feed after the second filter (which was practically equal to the concentration of the overflow from the first filter) varied during this 8 0 0 6 28 9 2¼ period of a COD of approximately 1 g / 1 (at the time that the first filter operated at a fixed feed rate, illustrated in Fig. 8) to a COD of approximately 7 g / l (for example, at the time when the OB of the first filter rose to well over

O

5 b8 kg COD / m day). The volumetric feed rate and the OB of the second filter obviously depended on the feed rate and the OB of the first filter; during this period, the volumetric feed rate of the second filter ranged from about 0.33 1 / h to about 1 1 / h and the OB of the second filter

O

10 ranged from about 0.1 * 8 to about 8.32 kg COD / m day.

The total percentage of COD removal (based on the COD content of the feed for the first filter) as a result of passing the water through the two filters ranged from more than 89% to more than 98%, and was for the most part of the period approximately 95% or more,

In this example, the two filters were in different spaces and the overflow from the first filter was collected every day and then fed to the second filter the next day. During this 1 day 20 waiting period, the overflow cooled; in order to keep the second filter at a temperature of about 37 ° C, this filter operated with recirculation to make use of a recirculation heater (which was placed as shown in Fig. 1). (in a technical installation, the hot overflow of the first filter would be conducted directly at the bottom of the second filter without the need for reheating or temperature control and recirculation would not be required; the second filter would then operate under plug flow conditions and can be significantly smaller than the first filter).

During the 1 day waiting period, the surface of the collected overflow was in contact with the air (before the overflow was fed to the second filter, resulting in a loss of CO 2 and an increase in pH to about 8); as a result, the methane concentration in the gas generated in the second filter was relatively high, about 90%.

8 0 06 26 9 25

The controller can be a microprocessor (for example, an 80/10 microcomputer from Intel System) coupled with programmable digital input / output devices (two Intel Silicon Gate MOS 8255 Programmable Peripheral Interfaces) 5 and an "analog" input output card (for example, a RTI-1200

Real-Time Interface of Analog Devices). The latter card receives analog signals from the sensors, converts them into digital signals and supplies them to the input of the microprocessor; it also receives digital signals from the output 10 of the microprocessor and converts them into analog signals. Specifically, the input-output card offers the following: 32 analog single-output inputs (or 16 dual-output inputs), two analog outputs, two logic powered outputs, two Real Time clocks, a 1 K byte memory. Sr, six 15 of the analog inputs are used. The two analog outputs are used to drive a recording device and to control the feed pump 8. The logic powered controllers are used for low pH and high temperature warning. The clocks are used to determine a cycle time (βθ min) and to trigger data collection every 30 s and to start billing and control changes every six min. The one kilobytes memory is part of the total 5k memory for the controller using the microprocessor.

The microprocessor is programmed to automatically receive or collect the output signal from each sensor (the wet gas meter, the methane analyzer and the tachometer connected to the motor of the feed pump 8, as well as other sensors discussed below), 30 and so that a series of values for each of the variables (namely gas evolution rate, gas methane content and feed rate and others) is stored in the controller.

The microprocessor program is such that every 6 min the controller (a) calculates the average of each of these variables (over the previous 6 min.); (b) the mean KB for the preceding 8006289 26 over 6 min. (from the gas evolution rate and methane contents); (c) the mean OB calculated over the previous 6 min (from the feed rate values and from the feed concentration value, the latter value being manually entered into the controller via the keypad); (d) the average KB calculated for the previous hour; (e) the relationship calculated between the values of (c) and (d) (e.g. the percentage efficiency); (f) calculated the OB (and the corresponding feed pump speed) 10 necessary to maintain the preset relationship between the values of (c) and (d) and (g) an electrical signal to the motor controller of the feed pump to set a new value of OB.

The reason for applying an average ^ 5 for KB over 1h is to avoid or smooth out the input of very short-lived variations that would result in the incorrect KB not being declared. For example, vibrations or shaking of the anaerobic filter can cause an extra amount of gas to be released immediately; at the same time, every now and then some gas which has formed in the filter may collect locally in an area with packing (forming a "gas bag") and then suddenly release, whereby the installation becomes "burping", as it were. Individual measurements taken in such a period will of course not reflect the actual gas development rate.

It will be understood that the term "average" is used in a broad sense herein, as it will be appreciated by those skilled in the art that one can use any value proportional to the average, provided that the same proportionality factor is also applied to other significant variables or is built into the responses. For example, because the instrument collects 120 values for the methane content and 120 values for the gas development rate per hour, (if the calculated "average" KB when performing the above operation (e)) can be the sum of the 120 individually 80 06 28 9 27 measured methane development rates (obtained in the previous hour) or the equivalent of that sum expressed as KB and the controller can compare that sum to ten times the sum of the twelve individual OBs (obtained during the 5 previous 6 min).

In the foregoing description, it has been stated that the concentration of the feed is manually entered into the controller, that concentration being based on an analysis (e.g. daily) of the COD of the feed.

10 It will be clear that better control can be obtained by performing the COD analysis of the power supply more frequently and entering the COD value at corresponding shorter intervals in the controller and that this can be done by means of an automated COD analyzer (or a commercially available COD analyzer) that empties samples from the feeder.

Among the other signals applied to the microprocessor intermittently is a signal indicating the value of the pH. Preferably, this signal is supplied by a pH meter (eg, a pH meter model PöOL-2-1 20 from Great Lakes Instruments Inc.) with two matching glass electrodes. (One electrode is used as a reference electrode and is placed in a buffer with a pH 7). The electrical circuit is such that an output signal of zero is obtained if the solution to be measured (which is in contact with the glass electrode not serving as a reference electrode) has a pH 7 and a positive or negative signal is obtained if the pH resp. . is above or below 7. The pH cell is subjected to a cyclic comprising (a) rinsing with pressurized water (eg, an overpressure of 210 kPa) every hour for about 2 min to remove loose solids and gas bubbles from those portions that have come into contact with the to drain (b) expose recycled biomass-containing liquid for 38 min to a cleaning solution which is weakly acid for the removal of carbonates and which contains detergents (eg Alconox) 8006289 28 to remove greasy layers and contains a bactericide for example hydrogen peroxide to kill living organisms that may still be present at significant points (the hydrogen peroxide-containing cleaning solution 5 preferably contains an agent such as EDTA to complex bind metals that could cause decomposition of the hydrogen peroxide); (c) subject to a rinse with water for an additional 2 min and then (d) expose to the recirculation stream for 18 min. These consecutive 10 treatments are accomplished through a system of valves V1, V2, V3, VU and V5 (fig. Ό controlled by a timer 31. During the first b2 min of the cycle, V3 is open, during the next 18 min (the above step d)) this valve is closed. During rinsing with water (steps a and c), V5 is open and V1 and V2 are in a position that allows flow of the water supply through the pH cell to the drain 32. At the beginning of the cleaning step (step b), V1 »is open and V1 and V2 ii are in a position that allows flow from the detergent container 33 through the pH cell to the drain 20; then the valves are closed and the pH cell remains filled with a stationary amount of cleaning buffer solution. The buffer solution contains NaOH and PO 2 in amounts such that the pH is 6.2; This serves as an aid for checking the setting of the instrument, because the pH measured with this instrument is continuously visible in the usual way and the person operating the installation can check whether the displayed value for the pH in stage B is indeed 6.2. During the last 18 min (step D) the circulating recirculating solution is passed through the cleaned and flushed pH cell (by closing V3 and setting V1 and V2 in such a way that the solution can flow through V1, the pH cell and V2 and back to the point where she meets the nutrient solution). The arrangement is such that the output of the nH meter is only collected (and passed to the controller) during the last part 8006289 29 of stage d); this can be done by means of the time switch which controls the setting of the pin pin, ie the time switch can close the electrical connection between the last 10 min of stage d) and this connection for the rest of the 5 keep time open. During those 10 minutes, the controller can collect the pH value once every 30 s, as can the signals from other sensors.

The pH indicates the equilibrium between acid-forming and acid-using (methanogenic) bacteria and can be used for control in conjunction with the KB. Then the controller is programmed to turn on a light (or other alarm signal) indicating "low pH" when the pH drops to a predetermined level (eg pH 6.8) and so that the feed pump motor shuts down when the pH drops 15 to a predetermined lower value (e.g. pH 6.2).

If the feed pump is turned off in this way, the recirculation pump will continue to run; It has been found that during this period, the methane-forming bacteria will gradually consume the excess of volatile fatty acids and will develop methane (because the fatty acid-forming bacteria have no new food sources from which to form new amounts of volatile fatty acids) and the pHzal will rise. The controller is programmed to have a dead zone (which ranges, for example, from pH 6.2 to 6.8); if the pH has risen to a value above the upper dead zone level, the controller will turn the feed motor back on to feed feed at a rate determined by the KB of the filter at the time the pH rises above the level of the dead zone. A more elaborated pH strategy that can be used includes maintaining the pH at an optimal value for the methanogenic bacteria. This may include (1) presetting the regulator so that a 0B is maintained such that there is an excess supply of volatile fatty acids used as "feed" by the methanogenic bacteria, eg preset the efficiency to 80 06 28 § 30 approx. 80% and (2) presetting the controller so that if the pH is above an optimal minimum, for example above pH 7.05, the controller will operate in the usual way, but if the pH drops to a minimum of 7.05 (or other minimum) or up to a lower value, the controller maintains a lower 0B: KB ratio (or lower Δ B). For example, the controller can be programmed to increase the preset efficiency by 1% (for example, to 81% from the previous 80%) for each new pH reading 10, if the pH is 7.05 or less at or below the minimum of 7.05 (so that, in the scheme described above, the new pH measurement is taken once an hour, the efficiency will be increased by 1% every hour / “eg to 81% in the first hour , up to 82% in the second, etc. 7 as long as the pH is at or below the minimum It will be understood that such an increase in efficiency means that, for a given KB, the 0B (for example, the feed rate) is lower will be than before.

Another signal applied to the micro-20 processor every 30 s is the signal indicating the temperature of the interior of the filter. The controller is programmed to turn on a larnn (or other alarm signal), indicating "high temperature", when the temperature reaches a value of b2 ° C, to avoid damaging the methanol bacteria. The heating device in the recirculation cycle is controlled (for example, manually but preferably thermostatically) in response to the temperature in the filter so that the interior of the filter is kept at the optimum temperature, a second thermo-static heating controller based on the temperature of the flow leaving the heating device prevents the temperature of the recirculation flow from rising above U2 ° C and thus damages the bacteria in that flow.

A particularly preferred heating technique consists in that a heat exchange fluid (eg steam 8006289 31) is passed through heating coils in the filter itself.

Like infLg. 6, it has been found that the rate at which methane is evolved is very sensitive to changes in temperature. The heater 5 was turned off at time Λ and the temperature was allowed to drop (from the predetermined value of 37 ° C) to the ambient temperature. Then, at time B, the heater was turned on again. At time C, the heater was turned off again. During this entire time, the controller was set to obtain a feed rate such that the efficiency was 85%.

The wet gas meter that measures the volume velocity of the gas development is sensitive to temperature changes. In one set-up, that meter was placed in a place where the temperature was kept constant within narrow limits; that was on top of the infrared methane analyzer that radiated heat from the internal heater. In this ------------ -......... setup, the wet gas meter temperature generally dropped from a daytime temperature of 29 ° C to a temperature of 20 during the night of approximately 26 ° C. When this happens, the meter (for exactly the same mass flow rate of gas of the same methane content) indicates that there is a drop in the volume rate of about 1%. The controller therefore decreases the 0B (feed rate) by about the same percentage.

In the examples described in this application, variations in the day and night temperature of the order of magnitude occurred and thereby the rate at which the OB increased; that is, it can be expected that by keeping the temperature of the wet gas meter more constant, the capacity of the anaerobic filter could be increased even faster than already appears in the examples. Preferably, the wet gas meter is kept at a practically constant temperature or the system is set up to compensate for variations in the wet gas meter temperature (for example, by measuring the temperature of the wet gas meter, the value of that 8006289 32 temperature each 30 s to the controller and program the computer to use temperature signal and multiply the volumetric speed signal by an appropriate temperature factor (based on ideal gas laws) or use a wet gas meter equipped with a suitable temperature compensator.

It will be clear that when using the control system, the transmission and processing of data and control signals need not be fully electric. For example, systems operating with compressed air or other analog (or digital) systems can be used for at least part of the control system.

Another aspect of the invention is the use of the disclosed system for maintaining an anaerobic filter at a substantially constant OB and at such a high

possible efficiency. In that case, if the desired OB

O

has always been reached (for example an OB of 12.8 kg COD / m day achieved by working with an efficiency of, for example, approximately 80%), the controller has been programmed in such a way that the OB does not exceed

O

12.8 kg COD / m day is kept and that the OB is lowered to a value below 12.8 only if the KB falls below some predetermined value. That predetermined value may be, for example, (a) some fixed KB (for example 10.2b which means an efficiency of 80%) or (b) some 25% of the KB achieved during the same certain period with constant OB (for example 90% of that height KB). The end result of this control is an increase in the efficiency of the system to a level that can be clearly above the set efficiency.

30 Only if the KB falls below the predetermined value will the normal program take over the control and lower the OB in order to follow the KB value. In some situations, for example, if the relative amount of inert compounds (in the feed) (for example, pentaerythritol) increases and the 0B 35 is calculated from measurements of the total COD of the 8006289 33 feed, the efficiency may appear to drop (even if the anaerobic biomass is not adversely affected) and the control scheme will then cause an unnecessary reduction of the OB. To overcome this, the control scheme may include means for measuring the rate at which KB decreases and the control may be programmed (eg using a general purpose computer) that OB will not fall below a predetermined lower limit unless the measured the rate of decrease in KB is relatively high, which is an indication for an adversely affected biomass.

As previously indicated, the invention is very suitable for achieving a rapid increase in the load of a new filter. Example I of the aforementioned German patent application 027 ^ 831.32 and US patent application 57.5 ^ 5 of July 13, 1979 describes a start-up method using a start-up mixture containing relatively non-toxic and non-inhibiting compounds (methanol, acetic acid and formic acid) . Aqueous methanol alone or mixed with acetic acid can also be used as a feed-up. The feed for the start-up is preferably supplied until the KB has reached a value of at least 3 about 1.6 kg COD / m day or more (e.g. 3.2) and the methane efficiency is above 50% (e.g. 6o to 70%) .

Then the bacterial population is often large and varied enough to be multiplied quickly by means of the automatic control. A strict strategy can be applied during the initial regulatory period. In one strategy, the controller is used, still with the power-up for start-up purposes, to increase the capacity to a desired level of 30 (e.g. to a 0B of 12.8) and then the power-up is gradually replaced by the effluent before start-up to be treated in the filter (for example, a mixture of 95% feed before start-up with 5% regular feed the first day, the next day a mixture of 90% feed before start-up with 10% regular feed, etc., etc. .). In another 8006289s strategy, the gradual replacement of power supply before start-up by the regular regular power supply begins at the end of the above initial period, that is, approximately at the same time as the filter is applied to the filter.

If the filter is operating properly under anaerobic conditions, the redox potential of the liquid in the filter (as measured by the liquid being recycled) is generally more negative than -400 mV, for example approximately 10 -1 * 30 to -U60 mV. The redox potential can be measured with a standard redox electrode, which is placed in series with a pH electrode between valves V1 and V2.

Figures 7, TA, 8 and 9 indicate that at any given time, the filter appears to have a certain natural efficiency (which probably depends on the type and amounts of bacteria in the filter in relation to the tire and the concentration of COD and of potentially toxic or inhibitory substances in the diet and recirculation flows at that time). As in Fig. 8, the level of KB during the period in which 20 0B was held constant such that at a natural efficiency of about 80%. The addition of more COD (ie a higher 0B) than can be worked at this natural efficiency has been found to stimulate the growth of the methanogenic population and thus the production of more and more methane (see for example fig. 8). (OB), however, being set to obtain a higher efficiency than the natural efficiency, methane production decreases as can be seen in Fig. 7; probably the existing methanogenic population then does not have enough "feed" to continue making methane at the earlier rate. However, this reduction in feed rate (OB) does not appear to be harmful to the methanogenic population *, see for example the effect in Figures 7 and 7A where the efficiency decrease from 85 to 80% (the latter slightly below 35 natural efficiency). at that time) the KB 8006289 35 rose rapidly and see also Fig. H in which the effect of a reduced feeding rate on days 22 and 23 was largely absorbed in the following days. Another strategy (not illustrated) that can be applied if it is desired to rapidly increase the 5 KB and OB and arrive at a relatively high efficiency is that the controller is programmed to continuously calculate the natural efficiency and continuously increase the predetermined and preset efficiency increases to almost that natural level (for example, to a level that is 10 a certain amount below the natural level, for example 1 to 2 or 10 to 20% below the natural level).

Another aspect of the invention is related to the use of a second anaerobic filter which directly receives the outflow from the first recirculation filter, preferably without substantial exposure to the atmosphere (or aerobic obtained in another way circumstances). This second filter can operate without recirculation (or with very little recirculation). It can be operated without separate control of the feed rate (ie, the rate at which the first filter outflow is fed to it). It can work without a heater (or with very little heat input) because the liquid that recirculates in the first filter (and therefore the first filter's outflow) has already been heated to about 25 the optimum temperature for the methanogenic activity of the biomass. The second filter receives a feed with a significantly lower concentration (measured as COD in g / l) than the feed for the first filter and achieves a further significant reduction in the concentration even if its measurements are considerably smaller (e.g. less than half the volume) of the first filter.

It was found that if the feed of the first filter contains relatively inert organic compounds (which pass through the first filter almost without being degraded), the second filter will have an increased power to degrade this "inert" develops connections. For example, in an experiment in which the feed for the first filter consisted of a mixture with a degradable COD of 16 g / l (a mixture of acetic acid, formaldehyde, 5 methanol and butanol each of which gave a COD contribution of ^ g / l) and with a COD contribution of 9 g / l from relatively inert hydroxypropylated guar (guar gum which has been chemically modified to contain 0.53 hydroxypropyl ether groups per monosaccharide unit of the guar) and the first filter 10 (without the methane sensitive control described above) operates under conditions such that there is almost complete removal of the degradable material in the first filter (1.2 m high and 30 cm in diameter), the addition of a much smaller second filter 15 (height 1.2 m and diameter 15 cm, ie 1Λ of the size of the first filter) approximately inert removal of inert material; for example, after a period of acclimation, the removal of inert material in the first filter was about 25, while the total removal of that inert material in the two filters was close to 50%. It may be that the second filter, the feed of which contains relatively few common degradable components and has a high content of usually inert components, tends to develop a bacterial population that can degrade a larger relative amount of those ordinary inert components.

It will be appreciated that a number of such smaller filters can be used in series behind the first recirculation filter.

If the first filter in the series of two filters works with an automatic methane sensitive control, the second filter is protected against toxic overload because the presence of significant amounts of toxic material will reduce the feed rate to the first filter (and with it a reduction in the discharge) and a corresponding reduction in the load of the second filter. That is, the volume feed rate of the second filter is practically equal to the discharge rate of the first filter, which in turn is practically equal to the feed rate of the first filter.

The filters in the examples were 1.2 a high.

In large technical installations, the filters will generally be much higher (for example, they can be 6, 9 or 12m high) and their diameters can be correspondingly 10 larger (for example, diameters of 9, 15 or 30m or more) and the average hydraulic speeds may be higher.

As noted earlier, in the examples, the hydraulic speed (volumetric speed of the total liquid divided by the cross-sectional area of the filter) was in the range of 1 to 2 cm / min; for a filter that is about 9 m high, the feed speed can be such that the hydraulic speed is 5 to 10 times greater, for example about 5 to 20 cm / min. The residence time (volume of the filter divided by the volume rate of fresh feed) will usually be between about 0.5 and b days (e.g. about 1 day).

In the examples, the controller determined the 0B based on the value of the changing average KB over a 1 h. With a large commercial filter, the measured gas development rate may be less dependent on the effects of vibrations and impacts (belching). Then the instantaneous, or over a short time average (for example over 6 min average) values - .... KB can be used as control values, instead of using a changing average over 1h. This will make the control respond even faster and facilitate rapid increases of 0B (and KB) and facilitate rapid response or eventual disturbances.

In the examples, the top of the filter was at practically atmospheric pressure, while the contents of the filter at the bottom were under a pressure of approximately 1.2 m water column. In a large technical installation, the filter 80 06 28 9 38 will be under considerable pressure at the bottom (for example, a pressure of approximately 9 or more meters of water column) and the filter at the top may also be under some pressure (for example, a back pressure of the order from 10 to 25 cm water column, due to the way in which the developed gas is pumped out). The higher pressure will cause more CO2 to be dissolved in the liquid in the lower part of the filter, which can lead to a small drop in pH (for example, a drop of about 0.2 pH units).

In the examples, the liquid-gas separation zone was located in a vessel placed above the filter.

That zone can also be present in the same vessel as the filter. For example, above the branch there may be a zone with overflows where the liquid (and suspended biomass) flows over 15 to a collection space from which part is discharged as effluent while the residue is recycled; above the liquid level defined by the overflow ends, there is a free space from which the gas is collected.

The invention has been explained above with reference to mainly waste water from the petrochemical industry with an acidic pH of less than 6, for example a pH in the range of about 3 to 5. The invention can also be applied with neutral or alkaline flows (in in which case the acids are formed anaerobically in the fitter by the action of acid-forming bacteria). Such streams may contain sugars or higher carbohydrates or protein (for example, in waste water streams from the food industry). It has now been found that after a relatively short acclimation period, the recirculating anaerobic filter will decompose even insoluble carbohydrates such as dispersed starch granules.

As disclosed in the aforementioned German Patent Application p2T ^ 831.32 and U.S. Patent Application 57,5 ^ 5 of July 13, 1979, the effluent of the process is preferably fed to an aerobic poisoner.

The aerobic sludge formed can then be partially recycled 8006289 39 to the anaerobic filter controlled in the manner described. The aerobic sludge can be conveniently recirculated at a constant predetermined rate at which the COD is measured at regular intervals, so that the OB value supplied by the aerobic sludge is known and can be fed to the controller programmed to the rate at which fresh feed is supplied is controlled to maintain the preset efficiency (or the preset Δ B) based on the total OB (i.e. OB of sLib plus OB of 10 fresh feed). Instead of feeding the aerobic sludge directly to the anaerobic fiber, the sludge can be subjected to a treatment (e.g. at high temperature under pressure as in the known Porteous process; see British Patent 653-9ÖU and Water Waste Treatment Journal 15 I 51 + 3-5 / "19607) to convert much of the soluble material which is then fed to the anaerobic filter which is controlled as described.

The method described in this application can be used for the treatment of a wide variety of waste water streams, for example waste water as described in the aforementioned literature references. Unless otherwise indicated, the operating conditions, filter structures, etc., which are described in German Patent Application No. ^27 ^ δ31.32 and U.S. Patent Application No. 57,5 ^ 5 of July 13, 1979, are suitable.

More generally, the control technique described here can be used not only for anaerobic filters, but also for other anaerobic methanogenic reactors in which, as is the case with anaerobic filters, a bed of methanogenic bacteria is maintained in the reactor vessel while waste water 30 is in an upward direction flows through that bed, the effluent from that vessel having a much lower bacterial concentration than is present in the bed of methanogenic bacteria. One such reactor is, for example, the "Upflow Sludge Blanket (USB) Reactor" described in the article by G. Lettinga et al, entitled "Use of the 35 Upflow Sludge Blanket (USB) Reactor Concept for Biological 8006289 \ ko

Waste-Water Treatment, especially for Anaerobic Treatment "in Biotechnology and Bioengineering Vol. XII No. It April 19Ö0 biz. 699 — T31 *; this reactor and its operation are also described in the article by G. Lettinga entitled" Direct Anaerobic 5 Treatment Bandies Wastes Effectively "in the magazine

Industrial Wastes, January / February 19T99 pp. 18-2U, Uo and Ui * German Offenlegungsschrift 29209TÖ and 29 21070, Lettings article entitled "Anaerobic Treatment of Methanolic Wastes" in the journal "Water Research" Vol. 13 (1979) pp. 725-737 10 and the article entitled "Feasibility of Anaerobic Digestion for the Purification of Industrial Waste Waters" written by Lettinga in Documentation - Europe Sewage & Refuse Symposium EAS,

Uth Munich 1978 (published by Abwassertechnische Vereinigung St. Augustin, Federal Republic of Germany 5205) pp. 226-256, the 15th lecture on "Elimination of Organic Wastes from Surface Water" by Th. M. Van Bellegem at the 13th International TNO Conference in Rotterdam March 27-28, 1980 (text available from TNO) and the article "A Pilot Scale Anaerobic Upflow Reactor Treating Distillery Wastewaters" by Pipyn et al in Biotechnology 20 Letters 1, page U95 -500 (1979). Another reactor for this purpose is a reactor in which the bacteria are attached to carrier particles, as described in the M.S. Switzenbaum et al, entitled "Anaerobic Attached - Film Expanded-Bed Reactor Treatment" in Journal WPCF Vol. 52 No. 7 (July 1980) pp. 1953-25 1965j the articles of B. Atkinson et al entitled "Process

Intensification Using Cell Support Systems "(in Process Biochemistry, May 1980 p. 2U-32) and" Biological Particles of Given Size, Shape and Density for Use in Biological Reactors "(in Biotechnology and Bioengineering Vol. XXI p. 193-200 ( 1979) 30 and in published GB-A-2006 181, published May 2, 1979.

In both these types of reactors, the bacteria are present on relatively large particles or floating or floating aggregates of such size and density that they give a high rate of sedimentation in standing water so that even under high load conditions (for example, with a hydraulic 8006289 * hl residence time of less than 2 days, e.g. 1 day or less) the concentration of suspended solid bacterial material in the effluent at the top of the reactor is significantly low, e.g. less than 0.05 g (e.g. 0.01 g) suspended- Dehydrated solid bacterial material per g of COD in the feed for the reactor, and the bacteria are kept in the reactor for a long time (for example, their average residence time in the reactor is more than 10 days, for example, about 30 to 100 days or more).

8006289

Claims (13)

A method for treating an organic matter-containing wastewater stream while simultaneously mixing methane gas in an anaerobic reactor comprising a vessel with a back-mixed bed of methanogenic bacteria which have been combined into a mass that the bed remains practically entirely in the vessel while the waste water flows upward through that bed, characterized in that the rate at which methane is generated is continuously measured and the feed rate of COD is continuously varied so that a predetermined time relationship exists between the feed rate of COD which is varied and the measured methane production rate, such that the 0B corresponding to the feed rate of COD (feed rate) is greater than the KB corresponding to the methane development rate. 2. Process according to claim 1, characterized in that the reactor is an anaerobic filter with recirculation. 3. Method according to claim 2, characterized in that the measuring and varying take place at least once every 4 hours. Method according to claim 2 or 3, characterized in that the measuring and varying take place at least once an hour. Method according to claims 2-4, 25, characterized in that the measuring and varying take place at least once every 15 min. Method according to any one of the preceding claims, characterized in that the filter operates at a 0B of more than 8 kg COD / m day. A method according to any one of claims 2-6, characterized in that the KB is increased, by continuously maintaining the predetermined time-dependent relationship, to a value of at least 25 kg COD / m day. A method according to any one of claims 2-6, 35, characterized in that, by maintaining the predetermined relationship determined by time 8006289 ", the amount of biomass is increased to such a level that the amount of free liquid in the filter less than 70% of the open volume of the filter Method according to any one of claims 2-6, characterized in that, by continuing to maintain the predetermined time-dependent relationship, the amount of biomass is increased to such that the amount of free liquid in the filter is less than 50% of the open volume of the fillet. 10. Anaerobic methane-developing filter with recirculation, which is fed with an organic substance-containing waste water stream, characterized in that the filter contains an amount of biomass such that the amount of free liquid in the filter is less than 50% of the open volume of the filter, the biomass being formed in the filter by feeding said waste water stream at such a rate that the 0B is at least about 32 kg COD / m day. A method according to claims 2-6, 20, characterized in that the db feed rate of COD is varied such that the ratio of KB to 0B is kept at a predetermined numerical value of less than 1. Method according to claims 10 and 11, characterized in that the numerical value is less than 0.9. Method according to claim 10. characterized in that the feed rate of COD is varied such that 0B minus KB is kept at a certain numerical value, 1U. Process according to claims 2-6, characterized in that the outflow from the anaerobic filter with recirculation is fed directly to a second anaerobic filter. 15. A method according to claim 1, characterized in that the waste water stream in the anaerobic filter is heated with recirculation and that the outgoing stream in the thus heated state is led directly to the second filter. 8006289