---

EWH Optimization and Flexibility

This tool focuses on optimizing the functioning calendar of thermoelectric water heaters (EWH) through the analysis of a limited dataset. The primary goal is to identify and estimate periods during which flexibility can be introduced by enhancing the operation of the EWH. By utilizing only the electrical consumption data of the appliance as input, the tool calculates optimal periods for initiating EWH operation. This ensures the delivery of consumer-defined comfort levels, subject to certain constraints, while simultaneously minimizing operational costs or total energy consumption. The tool offers a practical solution for enhancing the efficiency and cost-effectiveness of thermoelectric water heaters based on consumption data analysis.

---

Documentation

Preliminary documentation available at the project docs/ directory. It includes the optimization documentation including all mathematical formulation.

---

Initial setup

Note: The commands below assume that you are running them from the root directory of the project.

With Local Python Interpreter:

With a local python interpreter, you'll need to manually install dependencies:

1. Install the python dependencies

   pip install -r requirements.txt

2. HiGHS Solver Integration

   When opting for the HiGHS solver in the optimization settings, please note that external binary files are required for seamless functionality. This library already includes the HiGHS binaries from v1.7.0. To update to the latest binaries, visit the HiGHS project and releases repository and specify their location in the designated path for optimal performance.

3. Carry the included tests to verify the installation (optional)

   pytest

---

How to run

The tool is designed to effortlessly process two primary inputs: the EWH load diagram and a set of specified parameters. Upon providing these inputs, the tool seamlessly channels them through a dedicated optimization pipeline. The results, refined through the optimization process, can then be plotted and exported. The end-user can run the application directly using a Graphical User Interface (GUI), or manually via an IDE, or a Python terminal.

Using the GUI:

• After setting up the library, just start the Streamlit GUI via:

  streamlit run main_gui.py

• The GUI should open automatically in your browser. The URL is also shown in the console, defaulted as http://localhost:8501

Using an IDE / Python Terminal:

• Find comprehensive guidance in the /examples folder, which not only contains sample input data but also showcases exemplar run code (via /examples/ewh_opt_example.py) for a seamless start to the project.
• Navigate to the /ewh_flex folder within the project directory, where all essential functions for execution are centralized.

---

Inputs

This tool is prepared to work with a single input containing the EWH consumption diagram with 1 min. resolution. However, there is a set of optional but recommended inputs that should be addressed in the input_parameters.json data file.

Mandatory inputs

The current version allows two alternatives:

1. To utilize this tool effectively, users are required to provide an Electric Water Heater (EWH) dataset in JSON format, comprising only timestamps and electrical consumption data (in watts) with minute resolution. From this standpoint, the water usage calendar is detected using the built-in load-to-usage converter. An illustrative example file can be found in the 'data' folder for reference (data_example_7_days.json).

2. As an alternative, the user can provide only some water usage calendar (e.g. baths), including the start period and respective duration. From this standpoint, the estimated read load diagram is created by a built-in usage-to-load converter. An illustrative example file can be found in the 'data' folder for reference (input_data.json).

Optional (but recommended) inputs

For enhanced customization and precision, the tool allows users to provide optional inputs such as:

• ewh_capacity: EWH capacity (l)
• ewh_power: EWH heating power (W)
• ewh_max_temp: EWH maximum allowed water temperature (°C)
• ewh_std_temp: EWH standard non-optimized functioning water temperature (°C)
• user_comf_temp: Hot-Water Usage Comfort Temperature (minimum user-defined temperature - °C)
• tariff: Tariff selection between simple (1) or dual (2)
• price_simple: Simple pricing value per kWh (Euro)
• price_dual_day: Dual day pricing value per kWh (Euro)
• price_dual_night: Dual night pricing value per kWh (Euro)
• tariff_simple: Fixed daily simple tariff pricing (Euro)
• tariff_dual: Fixed daily dual tariff pricing (Euro)

If the requester does not provide optional inputs, the default values, as shown in the example JSON below, will be used.

{
  "user": "sample_user",
  "datetime_start": "2022-12-07T00:00:00.000Z",
  "datetime_end": "2022-12-13T23:59:00.000Z",
  "load_diagram_exists": 1,
  "ewh_specs": {
    "ewh_capacity": 100,
    "ewh_power": 1800,
    "ewh_max_temp": 80,
    "ewh_std_temp": 60,
    "user_comf_temp": 40,
	"tariff": 1,
	"price_simple": 0.119585,
	"price_dual_day": 0.149118,
	"price_dual_night": 0.070511,
	"tariff_simple": 0.3604,
	"tariff_dual": 0.4285
  }
}

Note: The inputs user, datetime_start, and datetime_end should be ignored if the data is provided directly.

---

Outputs

The EWH Optimization Tool outputs a comprehensive set of variables in JSON format, including key information. This structured output provides users with a detailed overview of the optimized Electric Water Heater operation, facilitating easy analysis and further utilization of the results. This tool provides the following outputs:

• simulation_period: Simulated/Optimized Period including start and end datetimes (minute resolution)
• original_energy: Total accumulated energy for the real measurement profile (kWh)
• optimized_energy: Total accumulated energy for the optimized measurement profile (kWh)
• original_price: Total cost for the real measurement profile (€)
• optimized_price: Total cost for the optimized measurement profile (€)
• avg_daily_energy: Average daily energy Consumption for the optimized scenario (kWh)
• total_flexibility: Flexibility availability for the optimized period (minutes)
• perc_flexibility: Percentage of the total simulated period that can provide flexibility (%)
• avg_daily_flexibility: Average flexibility availability for the optimized scenario (minutes)
• savings_cost: Pricing savings between real and optimized scenarios (€)
• savings_energy: Energy savings between real and optimized scenarios (kWh)
• original_usage_profile: Estimated hot-water usage profile based on EWH's real load diagram
• optimized_calendar: Optimized EWH functioning calendarization

Please find and example of the output JSON below:

{
  "user": "sample_user",
  "simulation_period": {
    "start": "2022-07-12 00:00:00+00:00",
    "end": "2022-12-13 23:59:00+00:00",
    "days_in_simulation": 7.0
  },
  "original_energy": {
    "value": 24.3384,
    "unit": "kWh"
  },
  "optimized_energy": {
    "value": 20.196,
    "unit": "kWh"
  },
  "original_price": {
    "value": 5.43,
    "unit": "Euro"
  },
  "optimized_price": {
    "value": 4.94,
    "unit": "Euro"
  },
  "avg_daily_energy": {
    "value": 2.8851,
    "unit": "kWh"
  },
  "total_flexibility": {
    "value": 9332.0,
    "unit": "minutes"
  },
  "perc_flexibility": {
    "value": 92.58,
    "unit": "Percent"
  },
  "avg_daily_flexibility": {
    "value": 1333.0,
    "unit": "minutes"
  },
  "savings_cost": {
    "value": 0.5,
    "unit": "Euro"
  },
  "savings_energy": {
    "value": 4.1424,
    "unit": "kWh"
  },
  "original_usage_profile": [
    {"timestamp": "2022-07-12 00:00:00+00:00", "hot_water_usage": 0},
    {"timestamp": "2022-07-12 00:01:00+00:00", "hot_water_usage": 0}
    /* ... more entries ... */
  ],
  "optimized_calendar": [
    {"timestamp": "2022-07-12 00:00:00+00:00", "ewh_on": 0.0},
    {"timestamp": "2022-07-12 00:01:00+00:00", "ewh_on": 0.0}
    /* ... more entries ... */
  ]
}

---

EWH Optimization Modelling, Methodology, and Mathematical Foundation

• Explore the /docs folder for a detailed exposition of the tool's methodology and underlying mathematical principles. The modeling approach is anchored in fundamental principles, ensuring transparency and clarity in the application of mathematical concepts for robust optimization:

1. Input Handling:

   • Provide the EWH load diagram as the initial input to the service.

2. Conversion to Hot Water Usage:

   • Utilize the built-in converter to transform the load diagram into nuanced patterns of hot water usage, laying the foundation for subsequent optimization.

3. Optimization Model:

   • Temperature Assurance:
     • Guarantee that the water temperature within the EWH never falls below the user-defined comfort threshold during and after hot water usage intervals.
   • Energy Estimates:
     • Calculate the minimum accumulated energy required for comfort and the total equivalent energy accumulated in the EWH after each hot water usage, considering fluid mixing dynamics.
   • Comfort Guarantee:
     • Ensure that the final energy accumulated after use is equal to or greater than the minimum equivalent comfort energy value.
   • Penalty Mechanism:
     • Incorporate a penalty variable to handle challenges in guaranteeing constant hot water availability, penalizing excessive temperature reductions.
   • Energy Considerations:
     • Account for energy losses by usage, gains by heating, and losses by convection through the EWH's surface area.
   • Cost Minimization:
     • Minimize the total operating cost by optimizing the proportion of EWH heating usage and the respective pricing.

4. Non-linear Constraint Handling:

   • The optimization model includes a non-linear constraint, skillfully linearized via linear regression. This constraint calculates the equivalent internal energy after hot water usage, ensuring precision and efficiency in diverse EWH operating conditions.

5. Assumptions/Defaults

   • The service assumes several default variables such as:
     • Specific heat capacity of water (4.186 J/g°C)
     • Ambient temperature (20°C)
     • Inlet water temperature (20°C)
     • EWH initial water temperature (60°C)
     • EWH overall heat transfer coefficient (0.00125 kW/(m2*K))
     • EWH height (100 cm)
     • Total flow rate (8.5 l/min)

The current version also allows a built-in optimization resolution resampling from 1-min to 15-min or 1-hour, that should be addressed in the resample parameter from ewh_optimization function. Please do note that, while it resamples the data to 15-min/1-hour resolution, the provided data should always respect the 1-min resolution, in order to guarantee a proper load-to-usage conversion.

---

Ranking Mechanism

To grant the users a KPI, the service includes a built-in points/ranking mechanism that addresses the overall optimization results. For this service, the KPI weights both the simulated period and the energy savings. There are, however, some restrictions and other notes to the formulation:

• Since the overall service architecture works on the assumption that the water is already hot in the beginning of the simulation, the savings for the first 24-hours are biased by this standard. This is, since the water is already hot (defaulted as 60°C) the first water usages would not require the EWH to function accordingly. Hence, to remove this bias, the ranking system only accepts simulation results larger than 24 hours. For a simulation with a period smaller than 24 hours, the ranking system will assign 0 points and 0% of savings.

• Since it would be more challenging to have better savings for larger periods of time, the period weight includes a linear scaling with diminishing returns using square roots. This grows more proportionally across the range but still rewards longer periods with diminishing returns.

• Apart from the numerical results shown in the ranking table, no other personal information is stored.

• For the mathematical foundation, please check the \docs folder

---

Contacts

If you have any questions regarding this project, please contact the following people:

Developers (SW source code / methodology questions):

• José Paulos ([email protected])