Dave--
What do you predict with the pump turned off and then back on after a fixed
time? The internal t/c should be properly predicted with a period of say 4
hours. The change in the water temperature from the pump stopping to a short
period of time after the pump is restarted and water can mix with the reservoir
and come to a equilibrium temperature should also be predictable. This may
allow a confirmation of the 41000 joules per degree C that is assumed for the
heat capacity of the reactor and other components. I will bet that number is
accurate to only +- 10% if it is calculated based on materials and their mass.
With your modeling which will allow a better determination of this important
parameter, a more accurate determination of any excess energy associated with
energy pulses should be possible.
Bob Cook
----- Original Message -----
From: David Roberson
To: vortex-l@eskimo.com
Sent: Friday, January 30, 2015 5:29 PM
Subject: Re: [Vo]:Alternate Calculation and Calibration Method for Mizuno
Report
Sorry about the late response to this important posting. I was diverted by
many interruptions and lost track of where I was before they arose. You have
made so many interesting points that I may fail to address them all properly,
and if that happens please send them back for another run. :-)
I agree that the pump power and any trims applied by my model for calibration
purposes will be balanced out over the long term if it remains constant. This
assumes that the coolant temperature is measured both before and then after the
input power pulses have completed their effects upon the coolant. Under this
condition, there remains a constant temperature difference between the coolant
and the ambient that is briefly modified by the power pulses but returns to the
original difference after several external time constants have passed.
The signal begins to leak from the thermal capacitance immediately upon being
deposited by each pulse. This is a significant amount of signal loss,
especially for the first of the series of pulses. If for example 3 pulses are
generated, the first one at a time of 0 seconds and the value of the pulse
magnitude generates a 1.0 degree change within the thermal capacity the
remaining amount of signal after 6.2 hours is only 1.0 * e^(-6.2*3600/27470).
This calculates out to be .4437 degrees. The residual for the second pulse
that occurs at 2 hours is less effected but only .576 degrees of delta remains
for a 1.0 degree signal. The thermal leakage has eaten up nearly half of the
original signal and that is one of the major reasons that the time constant
must be very large for this type of calorimeter to be applied to this type of
device if accuracy is required.
The actual amount of pump power that is leaking into the system can be
determined by the time domain response of the system or by a static and
accurate measurement of the difference between the ambient temperature and the
coolant temperature. The data submitted within the November report contained
an excellent period of time after the last pulse was generated until late in
that evening. The curve of coolant temperature displayed a slope of zero at
what is referred to as the peak level on October 22 at approximately 6.8 hours.
It so happens that the peak occurs at exactly the temperature at which the
leakage power of the pump times the thermal resistance equals the difference in
temperature between the coolant and the ambient. This is because the ambient
is slowly dropping throughout that time period. When the ambient plus the
contribution of the pump is greater than the coolant, it rises. When they
match, the system is static for a short period of time. And, once the ambient
plus pump effect is below the coolant temperature, the coolant begins to cool
off. This allows me to get an estimate of the pump leakage power.
My technique has a form of self calibration included in its operation. The
input power is accurately known and thus the number of joules deposited per
pulse can be calculated. This energy due to the input alone can not escape by
any means except the thermal resistance of the device. Since we know the
thermal capacitance by your calculation, we know exactly how much temperature
delta should be contained due to each pulse. My model subtracts off the
effects due to ambient variation and therefore each pulse is capable of direct
measurement. The amount of signal leakage due to the modest external time
constant is relatively small just after a pulse completes. At the moment I
am seeing about a 25 % increase in each pulse amplitude compared to what is
expected and I am assuming that is due to excess power. It would be great to
have a known dummy system so that I can verify that the 25% is real and not due
to an error in the assumptions.
I do not get anywhere near to the 3 to 1 excess power out over input power
that is reported. I am seeing a 1.25 ratio instead of 3.0 or so. Perhaps I
misread the report, but it seemed to be clear within it that the number of
joules you determined at the output was in the vicinity of 100,000 while the
signal only contained 30,000 joules. I am seeing more like 30,000 * 1.25
joules output. And, that is including both of the parts of the system with a
heat capacity of 41000 joules per degree C.
I do not neglect to include the HVAC system in any manner. The effect of
these systems is shown in the room ambient conditions. When I take the ambient
data supplied and the initial water temperature as inputs and predict the time
domain temperature of the water to within .1 degrees C for many hours into the
future, something must be working right. I am applying reasonable assumptions
supported by careful measurements and calibration techniques. It is not easy,
but it is good practice. My model is relatively simple and effective and the
fact that the operating temperatures are low makes the system reasonably linear
and not prone to being influenced by radiation complications.
If I missed some issues that need to be considered please bring them up
again. I want the model to accurately reflect what is seen under real life
conditions.
Dave
-----Original Message-----
From: Jed Rothwell <jedrothw...@gmail.com>
To: vortex-l <vortex-l@eskimo.com>
Sent: Fri, Jan 30, 2015 10:33 am
Subject: Re: [Vo]:Alternate Calculation and Calibration Method for Mizuno
Report
David Roberson <dlrober...@aol.com> wrote:
Finally, the leakage power entering the water from the pump has to be
determined since it has a significant effect upon the total calculation. Even
though it is on continuously, it is impossible to achieve an accurate
calculation without its influence being considered.
The power has to be measured, but it does not have to be measured with
precision. As long as you show that it remains the same at all times, then you
can be sure this heat is included in the baseline and it cannot affect the
adiabatic calorimetry.
The calibrations show that the pump heat is stable, but it is difficult to
measure exactly how much heat this is. The pump heat is low, and it can only be
measured when everything else in the system is turned off. This means total
heat is close to zero and the noise from ambient temperature changes drown the
signal. Calorimetry always works better power level above zero. When a
calorimeter measures from 0 to 10 W, it can detect the difference between 1.0
and 1.1 W with greater confidence than between 0.0 and 0.1 W.
The pump is described here:
http://www.iwakipumps.com.vn/doc_viewer.aspx?fileName=/upload/file/md.pdf
There are two sources of heat from the pump, and they are both likely to
stable:
1. Work by the pump itself, by the impeller. The pump consumes 10.8 W of
electricity. The other pumps in the MD series are 15% efficient, so the pump
cannot deliver more than 1.5 W of mechanical energy. It probably delivers much
less than this. In any case, power consumption is steady so mechanical power
will also be steady.
2. Heat from the motor. There is no direct, physical connection between the
motor and the pump. As shown in the figure on p. 4, the motor turns a magnet,
which turns another magnet inside a waterproof section at the end of the pump.
The second magnet turns the impeller. This is done to make pump waterproof and
gas tight. This also minimizes the heat conducted from the pump motor to the
impeller. Very little heat will be conducted by this path because plastic is a
poor conductor; and because the motor is designed to be self cooling; the pump
housing is cool to the touch; and Mizuno has a small fan blowing on the pump at
all times.
The conductivity of the plastic shell cannot change, so however much heat it
conducts, as long as the motor consumes the same amount of power the level of
conducted heat must be the same. Therefore it cannot affect the adiabatic
calorimetery.
It was then possible to subtract this curve from the measured coolant
temperature response to have a clear view of the true signal that is generated
by the power pulse entering the system and any excess power it originates. I
added a three pole digital filter following the subtraction to eliminate most
of the nasty noise remaining.
This is complicated because you do not know how much anomalous heat there is
in the first place. That is what you are trying to derive. We need calibrations
without any reactive Pd metal in the reactor, and no anomalous heat, so that we
know exactly how much energy is input and output. I hope we will soon have
these. This simplifies the problem greatly. It will also tell us exactly how
much heat is captured by the reactor stainless steel.
Each of the three input power pulses contained within this particular data
file (October 21, 2014 ) was easy to measure when subjected to my process. I
could determine that about 25% extra energy was generated by the Mizuno device
for each pulse.
I got 38%, based on a simpler method. (See Table 1.) I ignore heat from the
pump for the reasons explained above. If you are including a small contribution
from the pump that might explain the difference between 25% and 38%. And, if
you are including that heat, I am confident you are making a mistake.
That is about 2500 joules for each one of the three. An explanation as
to why this number is less than that reported is revealed by my latest
technique of separating out the individual contributions.
Did you also take into account heat captured by the reactor metal? My 38% is
only for the water.
The fact that the pump is on all of the time does in fact tend to hide its
contributions to the final coolant measurements as has been assumed.
No, it does not "hide" the contributions. It negates them. The contributions
are already there at the start, and they cannot increase in the "final coolant
measurements." (The plastic cannot suddenly conduct more heat; the impeller
cannot do more work.)
The heat from the pump is in exactly the same category as the heat from the
overhead lights. It is part of the unchanging baseline. The heat from the HVAC
would also be part of that unchanging baseline if Mizuno would improve the HVAC
and leave the heater turned on day and night. I think he is doing this now, at
my request. As I noted in the report I hope he can upgrade the quality of the
HVAC equipment.
The fact that the pump impeller heat is generated on the inside of the
calorimeter, whereas HVAC heat and pump motor heat comes from the outside, has
no bearing on the situation. The location where the heat originates is not the
issue. You cannot negate the HVAC contribution the way you negate the pump
because the HVAC varies so much and it turns off for hours! If the pump were
turned off for hours you could not negate it either. Or if the pump suddenly
doubled its output, you couldn't negate it.
As I said, this only applies to adiabatic calorimetry. Pump heat would
definitely be a problem with any other method of calorimetry.
The biggest problem is that during the night time hours when the ambient
drops heat is extracted from the system through the thermal resistance as the
device cools.
Yes. I hope this problem will soon be corrected.
In the early morning hours the heating system begins to operate and the
ambient rises several degrees C at a rapid rate. If you recall the step
response portion of my model above you will understand why this is so
important. The final temperature that the average ambient step wants to drive
the coolant toward becomes greater than the coolants initial value. The
coolant is not able to keep up with the rapidly rising morning ambient due to
the time constant so it lags behind.
Exactly. This is too complicated to model properly, I think, although you
seem to be doing a brave attempt. I hope that Mizuno eliminates these
fluctuations, so that we do not have to account for them.
The good news is that the calorimeter can be used as is with my calibration
system obtaining an accuracy of approximately 1000 joules per pulse.
The better news is that with proper environmental controls we will be able to
use a simpler calibration model. That's an example of fixing the problem with
hardware rather than software.
- Jed
