Retrofit Digital Twin
A data-driven study on residential decarbonisation, IoT telemetry, and smart energy optimization.
Dr Lisa Haskel
Lead Researcher
Dr Jose Paredes
Co-Investigator
Bakhtiyor Djamshedovich Sohibnazarov
Research Assistant
Rationale
A data-driven project measuring household energy transformation, optimization, and real-world return on investments.
Energy transition is increasingly linked to the cost of living crisis so reducing cost – and showing cost can be reduced - is urgent.
Government policy is moving towards some homeowners making personal investment to drive transition (combined with grants), so good predictions on returns on investment are required to persuade people to invest.
Storage is expensive and the grid required updating, so using power when and where its generated is needed.
Smart Retrofit Ecosystem
Integrating IoT data streams, APIs, and micro-sensors into a unified digital twin model.
Optimisation Factors
Optimisation of consumption (cost) is a complex undertaking with multiple factors to take into account including:
Weather
Unpredictable solar generation and varying heat loads demand ambient weather intelligence.
Home Appliance Differences
Varied load cycles, non-linear wattage curves, and peak consumption signatures across devices.
Personal Needs and Goals
Balancing comfort preferences, daily routines, and financial budget thresholds.
Our Project & Components
Project Goal
To gather as much data as can easily be accessed from a retrofitted home, using mostly commodity devices, with a view to a next stage of being able to make forecasts, predictions and automations.
What We Did
- Octopus API: Used for electricity import/export monitoring
- Tapo Smart Plugs: Energy measuring plugs on key appliances
- YoLink Sensors: Temperature and humidity sensors throughout home
- Dashboard & Insights: Live monitoring and initial analytical insights
Diagram of the System Components Architecture
Things We Did Not Do
- Gather data directly from the inverter (proprietary data access)
- Interact directly with any 'smart' devices
Data & Insights – Overall Picture
Solar PV System
- Installed Since: 4 May 2022
- Operation: > 4 Years Uninterrupted
- Panels Layout: 14 panels (East/West)
- Total Capacity: 4 kWp
Heat Pump Transition
- Installed: Summer 2024
- Gas Mains: Completely Removed
- Heating Baseline: 100% Electrified
Storage & EV Status
- Battery Storage: None Installed
- Electric Vehicle: None
Seasonal Solar Generation (2025)
Annual Generation & Cost Breakdown
Annual Energy Volumes (2025)
In Money Terms... (2025)
Learnings From the Dashboard: Power Flow Scenarios
Bad Day Scenario
Net Grid: +0.47 kWhBetter Day Scenario
Thu 8 OctoberLearnings: Peaky vs Optimized Self-Generation
Sat 3 October - "Peaky" Day
Net: -0.09 kWhFri 14 August - High Self-Gen
Net: -12.06 kWhLearnings: Appliances – High Usage Items
Washing Machine
Source: Tapo plug telemetry
Dishwasher
Source: Tapo plug telemetry
Heat Pump (Hot Water)
Source: Night baseline telemetry
Washing Machine Stem Plot
Discrete Telemetry (W)Dishwasher Stem Plot
Discrete Telemetry (W)Insights: Timeshifting Scenarios
Timeshifting needed to use self-generated electricity. The most efficient timings will vary according to weather, time-of-day tariffs and patterns of individual appliances.
Scenario 1: No Scheduling
UnoptimizedHigh load spikes occur during off-peak solar production (e.g. early morning or night), forcing grid imports.
Scenario 2: Time Shift
OptimizedLoads are shifted directly into the solar generation bell-curve, maximizing self-consumption and reducing import bills.
Winter Strategy: Shifting to Low-Tariff Hours
Solar Generation Curves vs. Octopus Cosy Tariff "Triple Dip"
Comparing Summer & Winter solar yield curves with highlighted off-peak low-tariff windows (p/kWh).
Because winter solar yield drops significantly (reaching only ~500W at mid-day peak), afternoon solar timeshifting alone is no longer sufficient. Heavy loads should be shifted into the highlighted "Triple Dip" low-tariff windows (04:00–07:00, 13:00–16:00, and 22:00–00:00) to run Heat Pump hot water cycles, Dishwasher, and Washing Machine at minimal cost while avoiding the expensive 16:00–19:00 peak rate.