To: dudleydevices@aol.com
Subject: FW: First
go at a plan
From: Spain, Stephen <Stephen.Spain@hdrinc.com>
Sent: Thursday, March 22, 2018 1:03 PM
To: Doug <DudleyDevices@Aol.com>
Cc: Doe, Chris <Christopher.Doe@hdrinc.com>
Subject: RE: First go at a plan
Doug, I
had some time today and looked through your website, including Dorena 3D
results below, very impressive, and will keep you and Lee Sheldon in mind for
future applications, thank you! Stephen
Stephen
D. Spain, PE, PMP, VP
D 503-423-3739 M 207-776-2309
|
23 |
|
6.1MW Kaplan Turbine |
Videos |
|
|
102_Ft_Index-Test_Cartesian_Display 44.1
Meg 6 min 29 sec.wmv |
||
|
Index Test Box Setup and Diagnostic Tools
For Kaplan Turbines.html |
||
TurbineOutputvsDischargeChart.pdf
TurbineCivilWorkLongitudinalSection.pdf
../SteadyStateScanIconVideo/ATECo_Winter_Kennedy_Pressure_Transducer_Cart.html
Figure 2Kaplan turbine representative
efficiency profile.
The Index Test Box (ITB) was
developed at Woodward Governor Company in 1984 as a dual-purpose test
instrument; the primary function is Kaplan turbine optimization
to tune-up the 3-D cam surface profile. Secondary, but equal in import is
the diagnostic analysis of the turbine control system behavior.
A third function for Condition
Monitoring
has recently been added to enhance the ITB’s utility value.
From: Doug [mailto:DudleyDevices@Aol.com]
Sent: Tuesday, March 20, 2018 9:30 PM
To: Spain, Stephen <Stephen.Spain@hdrinc.com>
Subject: Re: First go at a plan
Hi Steven,
I’m getting an Index Test Box (ITB) ready to send out to the Dorena Dam,
located about 2-hours south of Portland Oregon so the guys there can run it to
analyze their turbine's operation on a daily basis. Lee and I ran 4 index
tests for them in 2016 and 2017, but I never had to go to go to the dam. The man
that setup the Governor, 3-D Cam, SCADA system and data logger set his
equipment up to record continuous data, streaming to the data logger memory for
our index testing and then they email the 24-hour data files to me. They let
the recorder run 24/7 so any incidents will get captured, which turned out to
be very fortuitous for me.
I'm always asked by people such as yourself (that are familiar with Index
Testing Kaplan Turbines) when they see this demo, "Have you ever run this
in conjunction with a normal index test run by experts?" I've always had
to say “Nothing recent or direct correlated.” Ten year old data is
uninteresting and testing at differing heads on different dates is
inconclusive. The 4th Dorena index-test at 102 Ft gross head last
year is a perfect comparative test because the two analyses are working on
exactly the same data set. Hatch snatched the test away to run it themselves; I
was watching the Dorena lake water level on the USGS website and when it got
close to Max Pool I started calling out there for them to run the test data for
it. They said that my services would not be needed because Hatch had already
run the 4th of 4 tests intended to redefine the entire cam surface –
but they didn’t reduce the data to a gate to blade cam profile nor did they
have sufficient data for me to do so. We weren't going to be able to finish the
cam surface definition and update the cam with the new surface. At first this
was troubling, but then it was remembered that the data logger had been running
all the while, so I called the dam and asked the folks there to pull the
24-hour data file from the datalogger for the day Hatch index-tested the unit
and send it to me. They also sent Hatch’s final test report, and then
Northbrook Power Management gave me the Excel data sheet that Hatch used to
reduce the data.
It would have cost up to $50k to get a parallel test to validate
the ITB run by an outfit with the credibility of Hatch, and here I got it for
free because they slithered in and snatched the job away from me. Sweet!
I also got the data for the day before the
index test when they ran load-reject tests on the unit as well.
The index-test data can be analyzed by the
ITB the same no matter who runs the machine and it will be a simple extension
of what the ITB does already to add the load reject functions.
The boon of having the exact, same data set
analyzed by both Hatch and the ITB and getting precisely the same answers from
direct side-by-side analysis is proof positive that the ITB is getting the
right answer at a fraction of the cost of conventional index testing.
Lee has verified these conclusions.
This webpage has Hatch’s test report, their spreadsheet and
several videos showing the ITB analysis of the same exact data that they ran
and analyzed for their index test.
http://www.actuationtestequipment.com/Dorena/5MW_Kaplan.html
Please call me at (815) 335-1143 to get a brief demo via Skype
showing the ITB desktop display and an introduction to its tools and features
and go over these test results.
I also have 2 days of running time from HDC’s 2007 field testing
of their Gate Blade Optimizer (GBO) on Unit 9 at McNary that was acquired by
FOIA. The GBO is HDC’s knock-off of my ITB that they built in-house instead of
buying the ITB from me in 2006. I’m sure you’ll find interesting…
I had hoped to come out to Portland to do some dog-n-pony shows
but it didn’t work out in the short term, but I’d still like for you to see
this.
Best regards,
Doug Albright
Actuation Test Equipment Company
On 4/21/2017 1:35
PM, Spain, Stephen wrote:
Doug, thank you. I will forward and discuss with Ken. Stephen
Stephen D. Spain, PE, PMP, VP
D 503-423-3739 M 207-776-2309
From: DudleyDevices@aol.com [mailto:DudleyDevices@aol.com]
Sent: Friday, April 21, 2017 10:57 AM
To: Spain, Stephen <Stephen.Spain@hdrinc.com>
Subject: Re: First go at a plan
Hi Stephen,
After discussing this plan with Lee I have some questions about how the AQFlow ASMF flowmeter will be employed and the nature of its output data stream.
I'm familiar with the ASFM's capabilities and favor how it can be easily moved from one unit to another to facilitate the machine-to-machine-cross-calibration necessary for a proper Type-2 implementation. Using the exact same absolute-value flowmeter to calibrate every unit's Winter Kennedy taps will minimize the tolerance stack-up between the unit's flowmeters that will result from using a different absolute-value flowmeter for every unit. To get the best possible results from a Type-2 optimization the calibrations of the individual machines must be as close as possible to each other, which would be assured by using the same ASMF to calibrate every set of WK taps. Maintaining sensor calibrations traceable to NIST will maintain long-term calibration stability to make trending turbine efficiency performance possible.
AQFlow's Jan Buermans said the ASFM provides a single flow-data value every 7-minutes for a 32 path measurement instead of a steady stream of data values every 1/2 second or so like typical "time-of-flight" type flowmeters such as they have at Dorena Dam.
Can the ASFM output stream be input to the SCADA system and datalogger and recorded along with the rest of the 2-Hz data in its own data column? Getting these flow values merged into the turbine data file will be a big help in evaluating what it all means after the testing is completed.
It would also help to get timing marks in the recording at the beginning and end of the ASFM's 7-minute measurement interval so the ITB's analysis window's timing can be synchronized to the ASFM's measurement timing window. To make a meaningful comparison of the two flow measurements they must be measuring the exact same data points so we'll be comparing "apples to apples."
Getting timing marks could be accomplished by the ASFM sending a dummy value, such as "9999" at the beginning of the measurement interval, and perhaps "5555" at the end of the measurement interval, and then sending the computed flow value result when it becomes available and recording them in the data logger.
Along with the recorded data, can we get the 3-D Cam profiles for all of (or a few of) the machines? Either data tables or pictures of the 3-D Cam chart will work. The ITB has a handy feature to compare the actual 3-D Cam profile being executed by a machine with what it's supposed to be. The desired profile can be entered by either loading a picture of the 3-D Cam profile or the data table entered into the machine, displaying this data as the background and then plotting the new data points on top of it. The This feature works with not only the "canned data" we've been discussing, but will also work with a live data stream coming from the SCADA system.
Regards,
Doug Albright
(815) 335-1143
In a message dated 3/20/2017 12:24:04 P.M. Central Daylight Time, Stephen.Spain@hdrinc.com writes:
Doug, thank you!
I will discuss with Ken Pflueger at Douglas PUD and get back to you. We are interested in doing this with you and Lee, and once we refine scope, schedule budget, would issue a PO.
Stephen
Stephen D. Spain, PE, PMP
Vice-President, Hydropower
Senior Professional Associate
HDR
1001 SW 5th Ave, Suite 1800
Portland, OR, 97204
/D 503-423-3739 M 207-776-2309*+
tephen.Spain@hdrinc.com
From: DudleyDevices@aol.com [mailto:DudleyDevices@aol.com]
Sent: Saturday, March 18, 2017 12:50 PM
To: Spain, Stephen <Stephen.Spain@hdrinc.com>
Subject: First go at a plan
Hi Steve,
Thanks for you call yesterday.
Let me clear up a few areas. First my relationship with Lee. Lee and I have a gentleman's agreement where he is a casual employee of my company. No contract or any other paperwork, except the checks to pay him when we're done. Depending on my participation in his work, we share proportionally.
The index-testing work we are doing now is 50%-50%. A few years ago Lee consulted for DOE to evaluate grant applications and needed a corporate entity with a DUNNS number to get paid. As a professional courtesy my company got him paid without taking a pinch of his poke. My first wife made me get a corporation to protect her assets. It's been useful to me from time to time.
A bit about myself. I don't usually work in hydro. My primary business area is aircraft test equipment.
Hydro has always been something I hoped to get into with the Index Test Box (ITB), but I've had some problems. This cartoon best tells my feelings about it:
Your suggested $10k price for analyzing field data from an index test is acceptable to us.
We'd like $50% to start, and 50% net-30 upon completion.
A trip to the dam would be most appreciated, but is not really necessary. However you may want me to bring my laptop to show you and your posse how the ITB software works. There are many beneficial features to it:
- A Condition Monitor will provide high and low limits for unit efficiency under varying conditions of head and load.
- An Index Test function will automatically index-test a unit at a single power level using the Constant Power method.
- A SteadyState detector function will monitor a unit and light a light when the unit is running "steady-state".
- Canned Data playback
My part of this is massaging the numbers to glean the SteadyState data points and prepare them for 3-D Cam analysis by Lee Sheldon. Lee checks my work and passes it along to students from recent-classes, one of whom wrote a program for 3-D Cam data surface analysis that works well.
Lee prepares our final answer in the form of a spreadsheet and sends it back to me.
I write a cover letter explaining what the report says in simpler terms that my wife (no engineering training or field experience) understands, and publish it on my website - and then I prepare an invoice to get us paid.
When the check comes I send Lee his cut.
What I'd like to sell to you
Individual index test analyses at $10k each
Index Test Boxes to use for Unit Monitors and index testing.
The software itself so you can make your own ITBs.
Lee would like to sell you a 3rd generation Type-2 optimizer that his students created using their new book-learnin'. They really do teach useful stuff in college - if you get to them before they forget it...
We can haggle out the exact terms later, for now let's just take a look at the available data streams and prepare a better proposal after we see what we're getting into.
How it will work
Your guys setup the SCADA System to collect the data needed to optimize the turbine in continuous scans at rates of up to 100 Hz.
The recorder will run in background during your conventional index testing, and when the testing is done you'll send the data files to me for analysis.
Long term data collection of 8-months on the three turbines at Clergue was slightly over 7GB. It fit nicely on thumb drives to go through the mail.
For initial diagnostics, I'd like some 100 Hz scans during startup and ramping up to full load and then back to shutdown. The purpose is to observe for noise blooms that would indicate hydraulic instabilities, or rough zones to stay away from during normal operation.
The data channels for each unit will consist of::
- Date
- Time
- Scan#
- Forebay,
- Trashrack drop
- Tailwater
- Gate SetPoint
- Gate
- Blade SetPoint
- Blade
- Winter Kennedy pressure
- Any other available flow measurements
- Power SetPoint
- Power
- Power Factor
- Frequency
- Line Voltage
- ASFM Data Stream
Scans will be of 5-units abreast, the unit under test and two adjacent units on both sides. If close to the end of the dam, record the outermost 5 units on that side.
During the index-testing, 2 Hz will be sufficient.
The first step of the analysis is to setup the ITB data input format to read the files you send.
Next the gate and blade motions will be plotted in Cartesian coordinates to evaluate the behavior of the gates, 3-D Cam and blades.
Evaluate power tracking of the SetPoint and the effects of grid voltage and frequency fluctuations during startup and shutdown and how the unit behaves when the 3-D Cam mounts and dismounts the data table.
It is preferable to run an index test in the Constant Power mode, where the governor moves the gates to hold power constant, and the blades are positioned independently by the ITB. This is the method used at Clarence Cannon in 1984, Bull Run in 1987 but due to bad governor dynamics and contractual restrictions against operating "off-cam" I've never been able to use it again.
This magazine article draft explains Constant Power Testing.
Most likely the method you will be using is the classical fixed-blades, swept-gates which is how we ran the Dorena index-tests. This will work very well.
Attached as an example is the data from the Clergue powerplant in Sault Ste. Marie.
If you made something like this with 5 or more units abreast, that'd be sweet.
Thanks again for your call,
Cheers,
Doug Albright
Actuation Test Equipment Company
(815) 335-1143
