www.heatmyhome.ninja is a web tool to allow consumers to simulate their home's heating for a range of heating systems.
- Technologies Simulated
- Performance Metrics
- Input Form
- Core Simulation Stages
- Computational Complexity
- The Code
- Credits
The following heating systems are simulated:
1. Air Source Heat Pumps (ASHP)
2. Ground Source Heat Pumps (GSHP)
3. Electric Resistance Heating (ERH)
*(also known as Electric Boilers or Direct Electrical heating (DEH))*
4. Hydrogen Boilers
5. Hydrogen Fuel Cells
6. Biomass Boilers
7. Natural Gas Boilers
ERH is also known as Electric Boilers or Direct Electrical heating (DEH).
The electrified heating technologies (ASHPs, GSHPs and ERHs) are simulated with solar ancillaries and thermal energy storage, along with different energy tariffs. These technologies are sized based on the peak hourly energy demand experience over the simulated year.
The following solar ancillaries (a mix of solar photovoltaic and solar thermal technologies) are simulated:
1. Photovoltaic Panels (PV)
2. Evacuated Tube Solar Thermal Collectors (ET)
3. Flat Plate Solar Thermal Collectors (FP)
4. Photovoltaic-Thermal Hydrid Collectors
PV and ET, and PV and FP are also simulated in combination. The solar ancillaries are sized from 2m2 upto 1/4 the inputted floor area of the property (assume half the roof area), in 2m2 increments.
Water tanks are the only thermal energy storage (TES) technology currently simulated. The minimum water tank volume is 0.1m3 with a maximum of 3m3, and is simulated in 0.1m3 increments.
The heating systems are simulated with five different tariff options to find the cheapest option. The five tariffs are:
1. Flat-rate
2. Night off-peak
3. Evening-peak
4. EV off-peak
5. Variable.
The tariff are based off 2019 prior to the economic impact on the UK of Covid-19. Therefore expect current prices to vary from those estimated by the model.
For the hydrogen technologies (boiler and fuel cell) they are simulated using hydrogen from three different sources:
Grey: Hydrogen produced from steam methane reforming.
Blue: Hydrogen produced from steam methane reforming, but with the addition of Carbon Capture and Storage (CCS).
Green: Hydrogen produced from the electrolysis of water, with the electroylsis powered using grid electricity.
The simulator calaculates the following metrics for each of the systems modeled:
1. Capital Expenditure (Upfront Cost)
2. Operational Expenditure (Yearly Cost)
3. Net Present Cost (Lifetime Cost / Cost over 20 years)
Additionally for the electrified heating technologies the solar ancillery type and size, along with TES size that achieves the lowest lifetime cost is outputted.
The model requires six inputs from the user, being the:
1. Dwelling’s postcode
2. Number of occupants
3. Thermostat temperature
4. Floor area
5. Yearly space heating energy consumption
6. Maximum TES volume.
From the postcode, the longitude and latitude of the home can be determined, which are rounded to the nearest half a degree. The rounded position is used to load the hourly outside temperature and solar radiation datasets for the location. While the user may not know their home’s floor area and yearly space heating energy consumption, this information can usually be found on their Energy Performance Certificate (EPC), a legal requirement for any building sold, let or constructed in the UK since 2007. As the simulator only outputs the optimal TES size for a given system an upper TES size limit is required as input from the user to ensure the model does not optimise for a TES size that could not feasibly fit into the user’s home.
In addition to the inputs required by the model, users can also select:
- Where the simualtion is computed, either server-side using Rust (default), client-side using Rust or clinet-side using C++.
- If the simulation is run client-side you can choose to turn off optimisation (a global optmisation algorithm)
You can choose to save a completed, valid input form as a URL for sharing or later use. You can also download results and reupload the file to reload the results as well as the inputs for those results.
The first step of the model is to estimate the thermal transmittance (U value) of the property. This is done by simulating the home’s yearly space heating demand at an hourly resolution using a range of U values from 0.5 to 3.0 in 0.01 W/m2K step increments and selecting the U value which results in a yearly space heating demand closest to that specified by the user.
The hourly space heating demand is taken as the energy required to raise the inside temperature to the desired temperature. The inside temperature is a function of heat loss to the environment, and gain from body heat and solar radiation.
With the U value calculated, the space heating demand is recalculated, in addition to hot water demand, using the inputted thermostat temperature. Hot water energy consumption is a function of inlet cold water temperature and hot water consumption per occupant, accounting for daily and monthly variation.
The total demand is then used to calculate the operational costs and emissions of the hydrogen, biomass and gas heating systems, by multiplying the energy demand with operating costs and emissions per unit energy respectively, whilst adjusting for system efficiencies.
As for the electrified heating technologies, the model is more complex, being simulated for a range of solar thermal and photovoltaic ancillary technology sizes and combinations, TES sizes and tariffs. For each hour over a year, the following steps occur.
- Calculate the change in inside temperature due to heat loss to the environment and heat gain from body heat, solar radiation and heat loss from the TES into the living space.
- Update the TES state of charge due to heat loss, solar thermal generation, space and hot water usage, and heat produced from the electrified heating technology, powered either from grid import or solar photovoltaics.
- Calculate operational costs from electricity import for the hour using the selected tariff and operational emissions associated with the solar ancillaries and grid import.
The optimal solar ancillary size, TES size and tariff is then outputted to the user for each electrified heat technology - solar ancillary combination, where the optimal system is considered that with the lowest lifetime cost.
Due to each of the 3 electrified heating systems being simulated at an hourly resolution over a year, for each combination and size of solar ancillary, TES sizes and tariff options, del has a considerable computational cost.
The number of solar ancillary combinations tested per heating technology, O_solar, given by Eq. 1, is dependent on the dwelling’s floor area, A. The number of iterations required to go through all TES sizes is given by Eq. 2, and the total number of hours simulated for all electrified heating system combinations is given by Eq. 3.
The number of solar ancillary combinations tested per heating technology if given by:
solar_combinations = 6 * floor( floor_area / 8) - 1
tes_combinations = tes_max / 10
total_hours_simulated = 8670*3*5*solar_combinations*tes_combinations
total_hours_simulated = 130,050*solar_combinations*tes_combinations
The number of tes combinations tested per solar ancillary if given by:
tes_combinations = tes_max / 10
The total number of hours simulated for each heating technology, solar ancillary, TES and tariff option.
total_hours_simulated = 8670 * 3 * 5 * solar_combinations * tes_combinations
total_hours_simulated = 130,050 * solar_combinations * tes_combinations
The simulator was originally developed in Python, but was migrated first to C++ and then Rust. This was to achieve improved performance and utilisation of WASM. Please contact Jack Kirby via GitHub if you wish to discuss implementation aspects fo the code.
Here is a summary of the contents of the main folders:
docscontains all the files required for the HeatMyHome website (which is hosted through GitHub Pages)native-cppcontains the native C++ code for the heating simulator. This version is not as well maintained / written after the Rust implementation, although should produce the same outputs. You can choose to run this version client side on the website under experimental options.native-rustcontains the native Rust code for the heating simulator. This version is run server side, and can also be run client side on the website under experimental options.
If you are wanting to use a version of the heating model but unsure about which version to decide on I have summarised a list of pros and cons of the C++ and Rust versions below.
- The C++ version is faster than Rust, especially for large property sizes.
- The WASM version compiled from C++ is faster than that compiled from Rust, especially for large property sizes. (Although this is not a problem if you plan to run the code natively).
- The C++ version is much easier to performance profile (I used Visual Studio, for Windows). This is partially because the code is highly functionised, making it easy to see the performance of small sections of code.
- If you are unfamiliar with Rust or C++, but you have used Python or other similar language, I would recommend Rust.
- The C++ version is highly functionised, but this can make the code difficult to read as one must jump between functions, rather than being in chronological order.
- If you wanting to use the WASM compiled version, you must use Emscripten, while the Rust provides a crate (library).
- The Rust version contains an improved global optimisation algorithm (although the C++ version still performs better).
The website UI was developed with pure HTML, CSS and Javascript.
Chart.js was used to create the output graph.
Simulations are by default run on a server using Rust compiled WASM, but one can select to run the simulations client-side with Rust or C++ compiled WASM.
The structure of the website pages are outlined below:
- Home
- About Us
- Our Team
- Privacy Policy
- Motivation
- Education
- Heating Technologies
- Solar Technologies
- Why Switch?
- Documentation
- Simulator Inputs
- Simulator Outputs
- How-It-Works
- Simulate
The server is split into two major parts:
- For fetching EPC certificates.
- For computing the heating simulations.
The API is currently unavailable for public use due to limited compute hours. However the code is publicly available if you wish to view it or host the server yourself.
For further information please see the API specific repository.
Credit to Vecteezy for the following Vector Graphics: Tree, Nature and Human. We wish to thank credit Dr Wei He and Micheal Ryland for developing the heating simulation model. HeatMyHome is a University of Warwick Engineering Master's project developed by Jack Kirby, Aaron Hickman, Lina Khalil, Tudor Sucala, and Saahil Dhand. If you have any questions, please contact us at heatmyhome.ninja@gmail.com.