Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Designing model predictive control strategies for grid-interactive water heaters for load shifting applications

Elizabeth Buechler, Aaron Goldin, Ram Rajagopal

TL;DR

This work develops and evaluates a framework for model predictive control of grid-interactive residential water heaters to achieve load shifting under dynamic prices. It systematically analyzes four design choices—control-model fidelity (one-node vs three-node), temperature-sensor configuration, water-draw estimation without flow meters, and draw forecasting methods (perfect foresight, historical mean, historical quantile). Through simulations with real draw data and a high-fidelity tank model, it shows that model fidelity and sensor count most strongly affect electricity costs, while forecasting approaches mainly govern thermal comfort and runout frequency; draw estimation can be done from temperature and power without flow meters. The results provide practical guidance for deploying MPC in home energy management systems, highlighting that a three-node model with modest sensing and quantile-based forecasting can achieve substantial load shifting without compromising comfort.

Abstract

Model predictive control (MPC) strategies allow residential water heaters to shift load in response to dynamic price signals. Crucially, the performance of such strategies is sensitive to various algorithm design choices. In this work, we develop a framework for implementing model predictive controls on residential water heaters for load shifting applications. We use this framework to analyze how four different design factors affect control performance and thermal comfort: (i) control model fidelity, (ii) temperature sensor configuration, (iii) water draw estimation methodology, and (iv) water draw forecasting methodology. We propose new methods for estimating water draw patterns without the use of a flow meter. MPC strategies are compared under two different time-varying price signals through simulations using a high-fidelity tank model and real-world draw data. Results show that control model fidelity and the number of temperature sensors have the largest impact on electricity costs, while the water draw forecasting methodology has a significant impact on thermal comfort and the frequency of runout events. Results provide practical insight into effective MPC design for water heaters in home energy management systems.

Designing model predictive control strategies for grid-interactive water heaters for load shifting applications

TL;DR

This work develops and evaluates a framework for model predictive control of grid-interactive residential water heaters to achieve load shifting under dynamic prices. It systematically analyzes four design choices—control-model fidelity (one-node vs three-node), temperature-sensor configuration, water-draw estimation without flow meters, and draw forecasting methods (perfect foresight, historical mean, historical quantile). Through simulations with real draw data and a high-fidelity tank model, it shows that model fidelity and sensor count most strongly affect electricity costs, while forecasting approaches mainly govern thermal comfort and runout frequency; draw estimation can be done from temperature and power without flow meters. The results provide practical guidance for deploying MPC in home energy management systems, highlighting that a three-node model with modest sensing and quantile-based forecasting can achieve substantial load shifting without compromising comfort.

Abstract

Model predictive control (MPC) strategies allow residential water heaters to shift load in response to dynamic price signals. Crucially, the performance of such strategies is sensitive to various algorithm design choices. In this work, we develop a framework for implementing model predictive controls on residential water heaters for load shifting applications. We use this framework to analyze how four different design factors affect control performance and thermal comfort: (i) control model fidelity, (ii) temperature sensor configuration, (iii) water draw estimation methodology, and (iv) water draw forecasting methodology. We propose new methods for estimating water draw patterns without the use of a flow meter. MPC strategies are compared under two different time-varying price signals through simulations using a high-fidelity tank model and real-world draw data. Results show that control model fidelity and the number of temperature sensors have the largest impact on electricity costs, while the water draw forecasting methodology has a significant impact on thermal comfort and the frequency of runout events. Results provide practical insight into effective MPC design for water heaters in home energy management systems.
Paper Structure (32 sections, 10 equations, 8 figures, 4 tables)

This paper contains 32 sections, 10 equations, 8 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Literature Review
  3. Control model fidelity
  4. Temperature Sensor Configuration
  5. Estimating historical draw patterns
  6. Forecasting future draw patterns
  7. Supervisory vs non-supervisory MPC
  8. Water Heater MPC
  9. MPC optimization
  10. Model parameter identification
  11. Duty cycle conversion
  12. Estimating and forecasting water draws
  13. One-node MPC
  14. MPC optimization
  15. Water draw estimation
  16. ...and 17 more sections

Figures (8)

  • Figure 1: Diagram showing the water heater temperature sensor locations, heating element locations, and volumes associated with the one-node and three-node control models.
  • Figure 2: Control architecture for water heater MPC strategies, showing all computation blocks and data streams.
  • Figure 3: Water draw patterns across all homes and days of simulation. Hourly draw profiles are shown on top, and a histogram of daily water consumption across all homes and days is shown on the bottom.
  • Figure 4: Daily electricity price profiles that were used in simulations. The CalFlexHub Dynamic Rate is the spring HDP profile, available from the MIDAS server midasCEC.
  • Figure 5: Hourly estimated and actual water draw profiles for different control models and sensor configurations for one home over a two day period.
  • ...and 3 more figures