Search

Asset Publisher

Use Case: KIT/Using heat storage to minimize heat use and provide electrical flexibility

Description Of The Use Case

Name of use case

*Use case identification*
ID	*System configuration(s)*	*Name of use case*
UC11	SC Office Buildings 445 and 449 at KIT	Minimize heating, cooling and electrical energy consumption via Model Predictive Control

Version management

*Version management*
*Version No.*	*Date*	*Author(s)*	*Changes*	*Approval status*
	19/06/2017	A.Engelmann	Added MPC building Use Case	Draft
	05/07/2017	A.Engelmann	Added electrical storage	Draft

Scope and objectives of use case

*Scope and objectives of use case*
*Scope*	[Why? Motivation and high-level objectives] To meet the goal of reducing energy consumption of buildings (and thereby CO2 emissions), advanced control schemes can be used to lower the energy consumption of buildings. Thereby the incorporation of weather forecasts, advanced building models and the active use of heat/cold storages should lower the consumption of limited resources. Another aim is to reduce the needed amount of grid expansion at campus and the volatility in the grid. We try to achieve this by a combination of load shifting and the use of an electrical storage (peak shaving). [What? System under discussion, main functions, main actors] The considered building is equipped with a weather station and partial measurements of incoming heat of a distinct heating system. Whether or not we get active control access to the heating system heating system has to be clarified. Control variables could be the temperature of a heat storage, temperatures of the concrete core activation system, control of windows and the room temperature itself within certain bounds. An electrical storage is currently in planning.
*Objective(s)*	[identify specific objectives] O1: Minimize energy consumption satisfying temperature bounds in the building O2: Minimize fluctuation in building energy consumption O3: Minimize fluctuation in electrical energy consumption by means of the electrical storage (peak shaving)
*Belongs to use case group (if applicable)*	[Specify an arbitrary group name here in order to link multiple related UCs together] Building Energy Management

Narrative of use case

Narrative of use case

Short description

[How? Free text description of main function of controller and how it interacts with most relevant actors (Controllers & Control signals]

The occupants of the rooms have to agree on temperature limits of the building, such that they can be used as heat storages. One possible control scheme would be Model Predictive Control (MPC). The MPC controller incorporates a state space model of the building and calculates based on the available measurements an optimal control signal. An extension would be, to incorporate weather predictions in the decision. Based on that, the controller calculates optimal control inputs for the heating system, the cooling system and the concrete core activation minimizing the consumed energy of the building. The coupling between electrical consumption and heat consumption can be optimized simultaneously by using the heating/cooling system and the electrical storage.

Complete description

[More verbose description; include for example details about control domain, requirements towards input signals or applicable system operating modes (normal, emergency, …) ]

If devices with binary decision variables (e.g. window openings) should be included, mixed-integer MPC has to be used what can be challenging.
The measurements are based on a building automation system, which collects the available measurements centrally. Furthermore, aggregated measurements for the electrical energy consumption are available. They can also be used to estimate the heat emission of electronic devices like personal computers and the lightning system.
Aggregated heat flows into the radiators, into the warm water storage and air flows from the concrete core activation can be used as control variables.
Actually, there are no control-adequate models of the considered buildings available yet. Since the building model is crucial for a good control performance, an adequate modelling of the building is needed. There are different possibilities to obtain a state space model for the MPC controller. One way is to model the building components individually and aggregate the model components to a big model. Since this is in general costly we want to investigate a different approach. In that approach we assume a model structure and identify the free parameters by available measurement time series. The needed model order is here of particular interest.
An electrical storage with a capacity of around 20kWh is currently being planned. It should be used to minimize the fluctuation in the electrical energy consumption.

Optimality Criteria

(Directly associated with objectives. E.g. by what metric to 'minimise' something)

*Optimality Criteria*
ID	*Name*	*Description*	*Reference to mentioned use case objectives*
C1	Maximize efficiency	Minimize the energy consumption of the building in terms of the heating/cooling system.	O1
C2	Minimize fluctuation	Minimize the fluctuation in energy consumption to relieve the distinct heating system.	O2
C3	Minimize electrical fluctuation	Minimize the fluctuation in electrical energy consumption to relieve the campus grid.	O3

Use case conditions

Use case conditions

Assumptions

[Assumption; assumed relation to other systems: e.g. higher level controller sends a signal]

The required measurement values are available in real-time.
Control access to the heating/cooling system.
Accurate weather predictions are available.
The parameters of the building are available.
The electrical storage will be installed right in time.

Prerequisites

[Triggering Event (update of control signal or disturbance ...)]

The required sampling period depends on the time constants of the system dynamics and the required control performance. A rough estimate of one recalculation per minute is reasonable from our point of view. In case of the electrical storage the time constants are much smaller, i.e. peak shaving can be performed in time scales up to one second.

General remarks

General remarks

[everything which doesn't fit in any of the other categories].

Whether or not we get control access to the devices must be clarified.

Graphical RepresentationS Of Use Case

Graphical representation(s) of use case

a) Model Predictive Control

(Reference: Hysteresis modeling and displacement control of piezoelectric actuators with the frequency-dependent behavior using a generalized Bouc–Wen model Wei Zhu and Xiao-Ting Rui 2016 Precision Engineering 43 299)

Technical Details

Actors

*Actors*
*Grouping*		*Group description*



*Actor name*	*Actor type*	*Actor description*	*Further information specific to this use case*
Occupants	Human	Occupants of the temperature controlled rooms	The occupants specify the maximum/minimum temperature of the building.
Predictive controller	System	The system controlling the temperatures of the building	There is a centralized controller which controls all temperatures and heat flows simultaneously.

Step By Step Analysis Of Use Case Optional

Overview of use case scenarios

Identify all relevant use case scenarios; rel. e.g. to Sequence Diagram or Use Case diagram

*Scenario conditions*
*No.*	*Scenario name*	*Scenario description*	*Primary actor*	*Triggering event*	*Pre-condition*	*Post-condition*
1	Minimize energy consumption		Heat flows, air flows, window positions	Available measurements regarding the heating/cooling, weather forecast		Controller has recalculated controller response based on new input
2	Minimize fluctuation in energy consumption		Heat flows, air flows, window positions	Available measurements regarding the heating/cooling, weather forecast		Controller has recalculated controller response based on new input
3	Minimize fluctuation in electrical energy consumption		Electrical storage	Available measurements regarding the electricity consumption		Controller has recalculated controller response based on new input

Steps – Scenarios

Alternative / complementary to sequence diagrams.

*Scenario*
*Scenario name :*
*Step No.*	*Event*	*Name of process/ activity*	*Description of process/ activity*	*Service*	*Information producer (actor)*	*Information receiver (actor)*	*Information exchanged (IDs)*	*Requirements R-ID*

Common Terms And Definitions

*Common terms and definitions*
*Term*	*Definition*
Model Predictive Control (MPC)	An MPC controller controls the system by repeatedly applying an optimal control scheme to the system. By doing so, one get optimal control inputs in the sense of a specified criterion. Predictions can be incorporated in the control scheme.