Skip to content

WP1.2 Coordination Meeting December 10, 2019

Javier edited this page Dec 11, 2019 · 1 revision

Meeting Report WP1.2 ‘Modelica library for MPC’

1. MEETING SUBJECT, DATE

Subject: WP1.2

Date: 10-12-2019

Location: Skype for Business

Minutes taken by: Lieve Helsen (KU Leuven)

2. PARTICIPANTS

Company/Organisation - Participant

KU Leuven - Lieve Helsen (WP Leader)

KU Leuven - Filip Jorissen

KU Leuven - Iago Cupeiro (now @ Université Politechnique de Montréal)

KU Leuven - Javier Arroyo

LBNL - David Blum

IK4 Tekniker - aura Zabala

SINTEF - Harald Taxt Walnum

PNNL - Chen Yan

PNNL - Huang Sen

SDU - Toa yang

Politecnico de Milano - Ettore Zanetti

Oak Ridge national Lab – ORNL (US) - Piljae Im

Oak Ridge national Lab – ORNL (US) - Yeonjin Bae

Argonne National Laboratory – ANL (US) - Qi Li

Viessmann - Moritz Lauster (as observer)

3. AGENDA and REPORT

3.1. Welcome new members

a. Oak Ridge National Laboratory (ORNL)

Is joining the US BOPTEST team, with funding to develop a test case based on their light office flexible research platform, a highly instrumented and controllable test office building. The PI of that work, Piljae Im, has registered for IBPSA Project 1 as an individual and started attending our coordination calls. He is also developing/testing MPC algorithms in other project work. Yeonjin Bae from ORNL also joined.

b. A staff researcher at Argonne National Laboratory (ANL - also in U.S.) named Qi Li

Reached out in Rome after our BOPTEST presentation with interest in contributing to the development of functionality related to uncertainty - in forecasts and measurements - and their effect on MPC controllers. His expertise is in accounting for uncertainty in building performance modeling, so he can be a big help. Qi Li participated in the working group on this topic. He also registered for IBPSA Project 1 as an individual participant and started attending our coordination calls (and any uncertainty calls).

3.2. Approval/ Modification Agenda

Approved

3.3. Status & further discussion where needed

a. Emulators:

4 models selected to start MPC testing

i. BESTEST Air (LBNL/KU Leuven):

Finalized and to be sent to Filip for review (action Dave). Question about how much detail we want to see. Do we include a detailed FCU model (fan, valves)? It is sufficient to supply a flow at a particular temperature. This is the simplest emulator, we gradually increase the complexity in the other emulators.

ii. BESTEST Hydronic (KU Leuven /LBNL):

Control input is the water supply temperature, no update on this emulatoe (action Filip).

iii. 8-Zone Residential Hydronic (Engie/SDU):

Review has been finalized (Toa) and sent to Valentin. Confirmation needed by Valentin (action Valentin).

iv. Multi-Zone Office Hybrid Simple (KU Leuven/Engie):

Sent to Valentin for review. Valentin is not able to open the C-file needed for the borefield model. Iago checks the reason of the problem (action Iago and Valentin).

However further progress has been made in other models too!

v. Single-Zone Commercial Building (SDU):

BOPTEST ready, review by Valentin has been finalized.

vi. Multi-Zone Air-Based Prototype Complex (PNNL):

model is BOPTEST ready, documentation ready, control sequence still needs to be included, when ready a meeting will be scheduled with Iago for review (action Chen Yan).

Emulators on BOP-TEST repository:

Emulators are put on BOPTEST repository, a new directory (named Testcases) has been created by Dave, maybe later a new repository will be created for the emulators.

b. KPIs - scenarios

Due to funding gaps not much progress, but now funding is available again, so work can progress further. Choice between TMYx (1946-2018) and TMY3 (1973-2005) weather files. We have a data reader for TMY3 in IBPSA. Both will be compared in the representative days/weeks selection. Representative days/weeks (PNNL): approach will be further improved by looking into the energy end use such that not all is based on weather only, also on type of the day.

i. Purely based on weather data:

PCA, clustering and sampling to generate regression.

ii. Based on day of the week:

Process differently, in progress.

c. BOP-TEST:

i. Feedback on updated draft of the test case peer review document (Dave) that we have been using to develop BOPTEST test cases:

The update is based on feedback on the initial draft from the Aachen meeting, work presented in the Rome meeting, and continued discussions thereafter. It is primarily focused on clarifications for making the test cases BOPTEST-ready.

Comments: the CSV files for multi-zone buildings should mention the proper zone identifier and BOPTEST should be able to read this. Javier has sent a pull request to resolve this issue. Action Dave: update the document.

The documentation should describe the building (e.g. construction year …) since this influences the energy use. The documentation should give a full description of the test case.

ii. Web interface for BOPTEST

Development of a web interface for BOPTEST to store and present test results and show other information about test cases, etc (proposal by Kyle Benne – NREL). Access prototype here. This work needs to be coordinated.

Comments: Willing to share can have different categories based on the amount/detail of information shared. List of controller features should be provided to Kyle, such that it can be included in the test case information (optionally to be provided by the tester). Action Javier: google doc to start composing this list. Kyle can continue working in that direction, we’re very positive about this initiative and the preliminary result.

iii. Development needs.

Not treated

iv. How to parameterize and implement forecast errors (led by Tekniker):

More people joint the clarification-of-purpose-for-implementing-forecast-errors meeting. Literature review has been made (Laura). Next week another meeting is planned to present the first ideas and results. Action Lieve: agenda point for next call.

v. How to parameterize and implement measurement noise (led by Sintef):

Harald did further tests with noise library. Again computation time increased a lot. Advanced noise library does not show any improvement, the main cause is the sampling. Action Lieve: agenda point for next call.

vi. How to define workflow for training data for grey-black-box models different than testing data.

Proposed to next Skype meeting.

d. MPC Library (IbpsaMpc):

Proposed to next Skype meeting.

4. Reporting & Dissemination

See initiative by Kyle Benne (point 3c-ii)

5. Initiatives joint papers

a. MPC Review paper:

Lieve is performing the final review.

6. Miscellaneous

Doodle to be made for meetings in March and April: action Lieve.

Agenda next call:

  • BOP-TEST

o Implementation of forecast errors (Jesus/Laura)

o Implementation of measurement noise (Harald)

o Use of TMYx data to to define workflow for training data for grey-black-box models different than testing data (Dave)

  • MPC Library:

o merge type 3 works (Filip)

o pull requests merged (Filip)

o Discussion: how to proceed further, which models to be merged? (Filip)

o Unit testing framework in progress

  • Reporting & Dissemination: Call for internal testing

o Internal reporting within IBPSA Project 1: what needs to be reported? Spreadsheet

o Database (public): limited information

o End user agreement: permission to share the information publicly

  • OpenAIGym: new initiatives that are being taken, should we take initiative too?
Clone this wiki locally