Memorandum submitted by Paul Davies
As an electronic engineer and pub licensee with
many years of experience in the design and use of measurement
and data recording systems, I feel I must express my real concern
over the accuracy of the Brulines cellar monitoring equipment
that is extensively used by all of the major Pub Companies.
In this paper I will show that:
1. Brulines equipment is significantly inaccurate
because it cannot compensate for the variable amount of gas contained
in "real ale".
2. Brulines equipment consistently over-estimates
the quantity of "real ale" flowing through it.
3. Because of these inaccuracies, pub licensees
are being wrongfully accused of buying beer out of tie. This is
both unfair and potentially ruinous for many licensees.
The equipment in question is used as a deterrent,
and as evidence in the enforcement of the beer tie. Brulines operate,
on behalf of pub companies, a Volume Recovery Service (VRS) which
in theory recovers profit that has been lost by the pubcos through
licensees buying out of tie. This data is used to identify buying
out, and to calculate the fine imposed in pursuit of the VRS.
I also believe that this data is used by the pub companies as
the basis for rent calculations as well as an extra source of
revenue via the VRS.
My thoughts on the system are as follows:
The Brulines system is a flow monitoring system
designed to measure, via flow meters, the volume of liquid dispensed.
Whilst this type of system, when properly calibrated, will have
a reasonable accuracy (1.5%Titan 800 series Flow meter
manufacturers data http://www.flowmeters.co.uk/datasheet-flow-turbine-800.htm
) for keg dispense systems, it will be far from accurate for hand
pull cask beers"real ales"with the cask
vented to atmosphere.
The problem arises when the beer degasses in
the line. This produces a combination of beer and gaseous CO2.
This is referred to as Two-Phase flow and is a most complex subject
that has defied accurate measurement by simple methods. Degassed
beer in the line exhibits three forms of two-phase flow; slug
flow, bubble flow and churn flow, all of which require different
techniques for accurate measurement. Whilst calibration will produce
a close result, the variable nature of cask conditioned beers
and the variable amounts of CO2 in the line will produce inconsistencies
that a single calibration, which Brulines use, cannot cope with.
Frequent calibration would be required to achieve an approximation
of volume dispensed.
Accurate flow measurement of two phase flow
with a simple turbine flow meter is not possible. Any deviation
from the physical condition of the fluid at the time of calibration,
will make that calibration invalid. The Brulines system uses a
simple flow meter which registers gas flow as well as liquid flow.
No further method is employed to determine gas, water, beer or
any other fluid flowing in a cask ale line. Even water passing
through the pipes when the barrel is being changed is counted!
On Their Website, Brulines state: "Dispense
Monitoring, records the exact volume of liquid that passes to
each fount at any minute, of any hour, on any day". This
statement is simply wrong. No measurement system is exact; there
is always an uncertainty of measurement and an associated error.
For example, the type of flow meter used has an error of 1.5%
of flow, according to Titan, the manufacturers. This error is
measured using a stable fluid such as water under laboratory conditions,
it may be more when used in the field.
Simple turbine flow meter systems, such as Brulines,
do not take into account variables that can affect volume measurement.
Therefore there will be errors. Temperature and pressure are two
such variables, human error is another. The system cannot differentiate
between gas, beer, water or any other liquid, making the results
far from exact.
The design of this system is poor and has a
most fundamental flaw: it cannot determine the direction of
the flow. The flow meter gives the same pulsed output in both
directions. This can be particularly relevant when the hand pull
equipment is old, and the non return valve is not working effectively,
allowing beer to flow back down the pipe towards the barrel turning
the flow meter.
With these flaws in the measurement technique
and design, the results of the Brulines system will always give
a higher dispense than is actually the case. To substantiate this
assessment, a comparison was made with Brulines data available
from their website and data available from our own Electronic
Point Of Sale (EPOS) system which counts pints sold. The EPOS
data is secure and cannot be tampered with; the data may be required
for VAT, Inland Revenue and accountancy audits.
Five random weeks in my own pub were chosen,
and the data for those weeks compared. This is what was found:
|Week beginning||EPOS Data|
Although the EPOS system does not account for drip tray contents,
this is insignificant in comparison to the size of the difference
between Brulines and our EPOS data: a maximum of two gallons per
week. We do not suffer pilfering, our stock taking figures and
EPOS system would show if pilfering happens and we have good trusted
As can be seen, Brulines consistently overstates the amount
dispensed. This data is used to determine infringements of the
Beer Tie, and from this data, it can wrongly be assumed buying
out has taken place. What is more, if I am correct, this data
is also used as a basis for rent calculations, the assumption
being that the Brulines data is correct and is the real volume
of beer sold. This could result in significantly higher rent in
every Public House that has a Brulines system. The CEO of Brulines
has been quoted in a Publican article:
"Some pubs might understate the level of their business
in order to secure a lower rent and a system like Brulines can
give data that tells people, owners, what's actually going on,"
This statement supports my assumption on the use of this
data for rent calculations.
Further data integrity and Brulines process concerns:
The data is collected automatically and is relayed to head
office automatically. It appears on the website some two weeks
after it is collected and made available for use by the operator.
I have tried to use this data and it is of little or no value.
The data is two weeks old and does not tally with other data available
from our EPOS system and stock taking.
The two week delay, I find is very suspicious. What has been
done to the data in that time? How good is the audit trail to
ensure the collected data has not been corrupted? From my experience
of such remote systems, it is possible that the data may be corrupted
at any point in the data trail. It is even possible to inject
false figures into the data logger remotely unless stringent security
measures are in place to ensure data integrity and recovery should
there be system failures or errors.
Are Brulines an ISO (International Organization for Standards)
registered and approved Company? They do not advertise that they
are. ISO approval to the relevant standard would ensure the quality
of their systems through independent audit. What approvals do
Brulines have? I suspect none. In which case, the only verification
of the system is Brulines own with an obvious reluctance to admit
to any failings. It is suggested by this passage in the same Publican
article, that Brulines have no approvals.
"Brulines, which also supplies petrol forecourt data
and vending machine analysis, was still seeking a form of national
recognition for its systems through talks with trading standards
officials, Dickson said."
Trading Standards have denied any such talks.
Without approvals, what recourse is there for licensees to
question the accuracy of the data?
The collection of data and the operation of VRS by the same
company is incompatible with good practice. The temptation to
tamper with the data is always a possibility, especially when
Brulines profit from monies recovered by the VRS and there is
an absence of independent validation. There is no protection for
the licensee from collusion between the two departments.
In conclusion, the Brulines system is a far too simple measurement
system that has serious design flawsso serious, in my opinion,
that any data from it should be deemed invalid. For a system to
be used with such certainty, much greater accuracy is required.
A +1.5% error in measurement over twelve months would result in
a possible accusation of 12 x 9 gallon containers being
bought out of tie and an approximate £5,000 fine. (This
is based upon our use of 7200 gallons of beer annually and
known fine values.) From the specifications available to me, an
error of 1.5% would be the very best the system would be able
to claim, it would be significantly higher for hand pull cask
In this paper, I have just been dealing with errors in the
flow meter used to produce a signal proportional to volume. I
have not taken into account the errors that will be present in
the calibration, signal conditioning, recording and processing
equipment that is used to produce the final figures. Errors that
are cumulative and add to the overall uncertainty of measurement
which should be quoted as a total percentage error for the whole
Many people have been accused, fined and even lost their
livelihood because of this system, the data it generates and the
belief that Brulines are infallible.
I and many others have contacted Trading Standards with our
concerns but due to a bad interpretation of the law they have
chosen to do nothing. This leaves all licensees with this system
installed in their pub unprotected and vulnerable to exploitation.
One thing is certain: it is impossible, at present, to find
a measuring system that gets anywhere near the required accuracy
for the measurement of real ale. The devices used by Brulines
do not even have the required accuracy for keg beers.
Would it be too much to ask the pub companies to return to
the system that existed for many, many yearsthe mutual
trust between Landlord and licensee respecting each other to adhere
18 November 2009