Can you make minute level simulations?

[bbartling]

This is a lot of code to view which I apologize for but I am curious to experiment on minute level simulation for demand response algorithms to see load reductions.

import requests
import matplotlib.pyplot as plt
import pandas as pd
import time

# BOPTEST API URL
url = 'http://api.boptest.net'  # Replace with correct API URL

# Step 1: Select the test case
testcase = 'multizone_office_simple_air'

# Check if the test case is already running and stop it before starting a new one
try:
    requests.put('{0}/stop/{1}'.format(url, testid))
except Exception as e:
    print(f"No active test case to stop or error occurred: {e}")

# Start a new test case and get the test ID
response = requests.post('{0}/testcases/{1}/select'.format(url, testcase))
testid = response.json().get('testid')
if testid:
    print(f"Test ID: {testid}")
else:
    print(f"Error starting test case: {response.json()}")
    exit()

# Step 2: Set the test scenario to 'peak_cool_day' with dynamic electricity pricing
scenario = requests.put('{0}/scenario/{1}'.format(url, testid), 
                        json={'time_period': 'peak_cool_day', 'electricity_price': 'dynamic'}).json()
print(f"Scenario: {scenario}")

# Step 3: Set the control step to 5 minutes (300 seconds)
requests.put('{0}/step/{1}'.format(url, testid), json={'step': 300})
print("Control step set to 5 minutes.")

# Step 4: Define the specific hours of interest (e.g., 12 PM to 6 PM)
hours_of_interest_start = 12 * 3600  # 12 PM
hours_of_interest_end = 18 * 3600  # 6 PM

total_minutes = (hours_of_interest_end - hours_of_interest_start) // 300  # Number of 5-minute intervals

# Step 5: Advance the simulation in 5-minute steps during peak hours
power_usage = []
time_points = []

for minute in range(total_minutes):
    print(f"Advancing simulation for 5 minutes during high-usage hours... (Minute {minute + 1}/{total_minutes})")
    
    # Capture the power usage data during these intervals
    measurements = requests.get('{0}/measurements/{1}'.format(url, testid)).json()['payload']
    power_dict = measurements.get('hvac_reaAhu_PFanSup_y', 0)  # Example power measurement for supply fan
    
    # Extract the actual power value (if available) from the dictionary
    if isinstance(power_dict, dict) and 'value' in power_dict:
        power = power_dict['value']
    else:
        power = 0  # If no valid value is found, use 0
    
    # Append power and time data to lists
    power_usage.append(power)
    time_points.append(minute * 5)  # Time in minutes since the start of the high-usage period
    
    # Advance the simulation by 5 minutes
    simulation_advance = requests.post('{0}/advance/{1}'.format(url, testid), json={}).json()
    
    time.sleep(1)  # Adding a small delay between requests

print("power_usage ", power_usage)
print("time_points ", time_points)

# Step 6: Store power usage data in a Pandas DataFrame and plot it
df_power = pd.DataFrame({'Time (min)': time_points, 'Power (W)': power_usage})

# Plot the power usage over the time window of interest
plt.figure(figsize=(10, 6))
plt.plot(df_power['Time (min)'], df_power['Power (W)'], marker='o', color='orange')
plt.xlabel('Time (minutes since 12 PM)')
plt.ylabel('Power Usage (W)')
plt.title('Power Usage During Peak Demand Hours (12 PM to 6 PM)')
plt.grid(True)
plt.show()

# Step 7: Print peak power usage during this period
peak_power = max(power_usage)
print(f"Peak power usage during 12 PM to 6 PM: {peak_power} W")

But the power_usage = [] will read zeros.

[dhblum]

Hey @bbartling, thanks for posting. Yes you can run boptest simulations with 1-minute control steps if you wanted. Note that the simulation data retrievable with the /results request is saved at 30 seconds no matter your control step (unless your control step is smaller than 30 seconds in which case it will save at that step you define).

In your code, check a couple potential issues (only from my quick review):

• The /measurements API request gives you the names of the points you can read, but not their values. For that, you must request them using the /results API request.
• Double check the time of your simulation. You initialize the peak_cool_day scenario, which is fine, but note the scenario moves the simulation to a day that starts at midnight. You then go on to advance the simulation for 6 hours, defined by your time of interest. But I think you should make an advance call first to move to the hour of interest (12pm), and then step through your hours of interest?

[bbartling]

Hmmm. i am still having some issues. Could there ever be an example for this : )

[EttoreZ]

Hey @bbartling, I suggest you may take a look at the colab example we have on BOPTEST website example. Regarding the request to get the values for the measurements that you want find example below:

args = {'point_names':['reaTZon_y'],'start_time':-365*24*3600, 'final_time':365*24*3600}
res = requests.put('{0}/results/{1}'.format(url, testid), json=args).json()['payload']

[bbartling]

Thanks for the feedback! We were trying to run the simulation in 1-minute control steps for a demand response algorithm but encountered issues retrieving valid data during the simulation, which led to repeated errors. It seems that the /measurements API request only gives point names but not values. We would appreciate some guidance on how to correctly use the /results API to retrieve the actual data points after advancing the simulation.

measurements = requests.get('{0}/measurements/{1}'.format(url, testid)).json()['payload']
power_dict = measurements.get('hvac_reaAhu_PFanSup_y', 0)  # Example power measurement

Could you help with a snippet or an example of how to retrieve and interpret the values correctly?

[bbartling]

Thanks again for the feedback and the provided code snippets!

We tried incorporating your suggestions and made the switch to using the /results API for retrieving the actual values during the simulation. However, we are still encountering issues with getting the data. It seems that the results retrieval step is not providing any valid data even after setting the correct start_time and final_time.

Here's an updated snippet based on your example:

results_start = 0
results_final = 365*86400
point_name = 'hvac_reaAhu_PFanSup_y'
power = requests.put('{0}/results/{1}'.format(url, testid), json={'point_names':[point_name], 'start_time':results_start, 'final_time':results_final}).json()['payload']

df_power = pd.DataFrame({'Time (min)': power['time'], 'Power (W)': power[point_name]})

Even with this, we’re getting empty or invalid results. Could this be related to how we are advancing the simulation, or is there another step we're missing to get the proper data at 1-minute intervals?

Any further guidance or examples for handling minute-level simulation or specific scenarios like ours would be greatly appreciated!

[dhblum]

Here's a code snippet for using the /results API for your example (note that results_start and results_final are in seconds of simulation time starting from midnight Jan 1 and in my example will just get all the available simulation data, but they can be set to encapsulate whatever snippet of simulation time you want to retrieve for at the time of the request):

results_start = 0
results_final = 365*86400
point_name = 'hvac_reaAhu_PFanSup_y'
power = requests.put('{0}/results/{1}'.format(url, testid), json={'point_names':[point_name], 'start_time':results_start, 'final_time':results_final}).json()['payload']
df_power = pd.DataFrame({'Time (min)': power['time'], 'Power (W)': power[point_name]})

EDIT: The label in the DataFrame for time as (min) will actually be incorrect in the above. To be correct, it should be (seconds), or the result of power['time'] needs to be divided by 60.

[bbartling]

shoot i may have responded to the wrong message.

[dhblum]

Hey @bbartling, running the following code based on what you provided, I got the following plot. Note I added steps 2a and 6. Also, realizing the title of the plot isn't quite right. The data is from the start of the day until 6pm, not 12pm to 6pm. Hope that helps!

import requests
import matplotlib.pyplot as plt
import pandas as pd
import time

# BOPTEST API URL
url = 'http://api.boptest.net'  # Replace with correct API URL

# Step 1: Select the test case
testcase = 'multizone_office_simple_air'

# Check if the test case is already running and stop it before starting a new one
try:
    testid = '4c51b05b-dbe4-47a3-ab23-9170ce6d1913'
    requests.put('{0}/stop/{1}'.format(url, testid))
except Exception as e:
    print(f"No active test case to stop or error occurred: {e}")

# Start a new test case and get the test ID
response = requests.post('{0}/testcases/{1}/select'.format(url, testcase))
testid = response.json().get('testid')
if testid:
    print(f"Test ID: {testid}")
else:
    print(f"Error starting test case: {response.json()}")
    exit()

# Step 2: Set the test scenario to 'peak_cool_day' with dynamic electricity pricing
scenario = requests.put('{0}/scenario/{1}'.format(url, testid), 
                        json={'time_period': 'peak_cool_day', 'electricity_price': 'dynamic'}).json()
print(f"Scenario: {scenario}")



# Step 2a: Record current time (starting_time) and advance simulation to hour of interest (12)
starting_time = scenario['payload']['time_period']['time']
requests.put('{0}/step/{1}'.format(url, testid), json={'step': 12*3600})
print("Control step set to 12 hours.")
res = requests.post('{0}/advance/{1}'.format(url, testid), json={}).json()['payload']
print("Simulation advanced 12 hours.")
print('Start time is {0}'.format(starting_time))



# Step 3: Set the control step to 5 minutes (300 seconds)
requests.put('{0}/step/{1}'.format(url, testid), json={'step': 300})
print("Control step set to 5 minutes.")

# Step 4: Define the specific hours of interest (e.g., 12 PM to 6 PM)
hours_of_interest_start = 12 * 3600  # 12 PM
hours_of_interest_end = 18 * 3600  # 6 PM

total_minutes = (hours_of_interest_end - hours_of_interest_start) // 300  # Number of 5-minute intervals

# Step 5: Advance the simulation in 5-minute steps during peak hours
power_usage = []
time_points = []

for minute in range(total_minutes):
    print(f"Advancing simulation for 5 minutes during high-usage hours... (Minute {minute + 1}/{total_minutes})")
    
    # Capture the power usage data during these intervals
    measurements = requests.get('{0}/measurements/{1}'.format(url, testid)).json()['payload']
    power_dict = measurements.get('hvac_reaAhu_PFanSup_y', 0)  # Example power measurement for supply fan
    
    # Extract the actual power value (if available) from the dictionary
    if isinstance(power_dict, dict) and 'value' in power_dict:
        power = power_dict['value']
    else:
        power = 0  # If no valid value is found, use 0
    
    # Append power and time data to lists
    power_usage.append(power)
    time_points.append(minute * 5)  # Time in minutes since the start of the high-usage period
    
    # Advance the simulation by 5 minutes
    simulation_advance = requests.post('{0}/advance/{1}'.format(url, testid), json={}).json()
    
#    time.sleep(1)  # Adding a small delay between requests

print("power_usage ", power_usage)
print("time_points ", time_points)


# Step 6: Get power data from boptest since starting_time, store power usage data in a Pandas DataFrame, stop test case, and plot it
results_start = starting_time
results_final = 365*86400
point_name = 'hvac_reaAhu_PFanSup_y'
power = requests.put('{0}/results/{1}'.format(url, testid), json={'point_names':[point_name], 'start_time':results_start, 'final_time':results_final}).json()['payload']
df_power = pd.DataFrame({'Time (s)': power['time'], 'Power (W)': power[point_name]})
# Stop test case to release web-service resources
requests.put('{0}/stop/{1}'.format(url, testid))
print('Stopped test with testid {0}'.format(testid))



# Plot the power usage over the time window of interest
plt.figure(figsize=(10, 6))
plt.plot(df_power['Time (s)'], df_power['Power (W)'], marker='o', color='orange')
plt.xlabel('Time (s)')
plt.ylabel('Power Usage (W)')
plt.title('Power Usage During Peak Demand Hours (12 PM to 6 PM)')
plt.grid(True)
plt.show()

# Step 7: Print peak power usage during this period
peak_power = max(power_usage)
print(f"Peak power usage during 12 PM to 6 PM: {peak_power} W")