Possible bugs in TES modelling for CST?

Bug 1: Energy balance correct when excluding TES, incorrect when including TES
As I understand, and as suggested in a forum post ( sam.nrel.gov/content/thermal-energy-power-block-interpretation ), the following energy balance should hold:
SF thermal energy output >= energy charged to TES + energy sent to the PB
When I checked the hourly values, I found that there are hours when the sum on the right hand side is up to 20 MWh-t greater than SF thermal energy output. For some days, this happens in every hour when the SF thermal output and TES charge are both greater than zero. For example, this happens on 19 March, 21 March, 24 March, 25 March, 28 March, 29 March, and 30 March.
When the same simulation was run with zero hours of TES, I found that the sum on the right hand side is at most 0.2 MWh-t greater than SF thermal energy output. I may attribute errors of this magnitude to rounding in calculations or the stepwise values used to estimate HTF specific heat capacity. I do not think the same explanation can be used to explain hours when the difference is two orders of magnitude larger.

Hello

I am using SAM 2014.1.14 to benchmark a linear model that simulates TES and power block (PB) operation at 1 hour resolution. The hourly solar field (SF) thermal output from SAM is used as an input to the linear model. I chose to use a model of Andasol-1 in SAM to test my linear model. The original Andasol-1 ZSAM project file can be found at the following website: ( www.austela.com.au/projects ). I made two modifications to the original file. The first was to change the PB max turbine over design operation from 1.05 to 1. The second was to disable fossil fuel backup using the method described in the following forum post: ( sam.nrel.gov/content/zero-fossil-fuel-backup ).

During my testing, I noticed some behaviour associated with the TES model that I cannot explain and I believe may be bugs. To make this easier to read, the bugs are reported in replies to this post. Attached are the ZSAM file I used and the Excel file with some SAM simulation results for Andasol-1 without TES and Andasol-1 with TES. Graphs that may help understand bug 1 and bug 2 are shown in the spreadsheet for Andasol-1 with TES.

I realise that as I am using SAM for a purpose that it is not designed for, I may be encountering behaviour which appears as a bug but is not. I would appreciate a response to indicate whether the issue is being investigated or to clarify what may be misinterpretation on my part. Thank you in advance for any time and assistance you can provide to investigate and discuss this issue.

Regards
Edward

Bug 2: TES charge is lower than expected

Continuing with the outlined SF-PB-TES energy balance, if the SF thermal energy output is not enough to operate the PB, then it may be charged to TES if TES is not full. When I examined hours when SF thermal energy output was greater than zero and PB thermal energy input was zero, I found that there were hours when the SF thermal energy output was between 19 and 33 MWh-t larger than TES charge. A check of TES charge and discharge prior to these hours showed that TES was very unlikely to be full. As TES was not full I expected that all the thermal energy from the SF should have been charged to TES instead of having a difference of at least 19 MWh-t.

Bug 3: TES discharge but no thermal energy input to PB and not enough for normal operation

I found 18 hours when the thermal energy discharged from TES but the PB did not receive any thermal energy. All 18 hours were the last hour of discharge after several consecutive hours of discharge and all 18 hours had the same value of -28.8714 MWh-t. The value of 28.8714 corresponds to the thermal energy required to start the PB using the values of 55 MW-e for PB design gross output, 0.381 for rated cycle conversion efficiency and 0.2 for PB fraction thermal energy to start, and the following equation:

PB start = (PB design gross output) / (Rated cycle conversion efficiency) * (PB fraction thermal energy to start)

As all the occurrences were final hours of TES discharge, it is unlikely that the PB was being started. In addition, the PB was operating in the previous hour so I expect that the thermal energy to start the PB was not required. There was no SF thermal output at those hours and the PB minimum turbine operation fraction is 0.25, therefore the TES discharge was not enough for normal PB operation. I cannot think of another explanation for why TES was discharged in those hours.

Dear Edward,

Sorry for my slow response on this. My colleagues and I have been busy preparing the Beta release. Would you mind attaching the weather file you used for this analysis to your original message above? (If that doesn't work, please email it to me at sam dot support at nrel dot gov.

I did some initial tests with a different weather file and could not replicate your results.

Best regards,
Paul.

Dear Paul

Thank you for responding. I have attached the weather file to the original post. I thought that the ZSAM file also saved the weather file last used which is why I did not initially include it.

Regards
Edward