[Energy Calculation] Riemann sum integral calculates too high power usage for my fridge

[jakami (Jannic)]

Hey guys,

I’m using the Integration “Riemann sum integral” to calculate my energy usage in kWh for the energy dashboard. The sensors I’m using are smart plugs over the “local tuya integration” which update the power every 1-2 seconds. This work’s perfectly. But today I noticed that my fridge uses ~5 kWh/day which can’t be true. So i crosschecked with the tuya app “smart Life” and it reports ~0.5 kWh/day. The energy readings for all other kWh-calculations (server, pc,…) report pretty plausible readings. The differnce between my measured devices and the fridge is, that the fridge consumes energy in spikes and then drops to 0 which is common for a fridge. Is the drop to 0 maybe an issue for the “Riemann sum integral”-calculation?

I attached the screenshots and the part of the configuration which is responsible for the calculation. Just ignore the sensor “*_all”, since this is a calculated one with the template integration.
The sensor for W is sensor.power_4
The calculated Sensor for kWh is power_4_KUHLFRANK_spent

Thanks in advance <3

### POWER SPENT kWh ###
#riemann_sum_integral for kWh

#ALLE
  - platform: integration
    source: sensor.power_all
    name: power_all_spent
    unit_prefix: k
    round: 3
#Server
  - platform: integration
    source: sensor.power_1
    name: power_1_SERVER_spent
    unit_prefix: k
    round: 3
#TV
  - platform: integration
    source: sensor.power_2
    name: power_2_TV_spent
    unit_prefix: k
    round: 3    
#PC    
  - platform: integration
    source: sensor.power_3
    name: power_3_PC_spent
    unit_prefix: k
    round: 3    
#Kühlfrank
  - platform: integration
    source: sensor.power_4
    name: power_4_KUHLFRANK_spent
    unit_prefix: k
    round: 3    

######################################################
##############  SENSOR ENDE ##########################
######################################################

The spikes of the frige

The Energy dashboard showing ~1 kWh used today at 5AM

The tuya app “smart life” showing 0.10 kWh used today at 5AM

[tom_l]

The answer is in the docs:

[jakami (Jannic)]

I googled for it, but didn’t found anything but next time I RTFM ^^
Thanks and sorry, lesson learned

[chadbaldwin (Chad)]

Did this actually fix your issue? I tried this for my fridge and it didn’t seem to help at all.

[jakami (Jannic)]

The calculations where still a bit inaccurate but it was definetly better

[zillion42]

Hi,

I have a very similar scenario. I am also using the local Tuya Integration to read my smartplugs power reading in Watts and then using the Riemann sum integral to get the energy spent using the default trapezoid setting. The smartplug is connected to my laundry dryer (condenser dryer without heat pump). Depending on the dryer load, it should draw between 2.27 kW/h and 4.13 kW/h, according to the datasheet. In fact, depending on the day and maybe luck, the Riemann integral actually integrates values in that range, on other days and maybe with less luck, the result of the Riemann integral is more than 10 times as high, and I have no idea what could be causing this.

I will attach the screenshots of two workloads, one resulting in 3.64 kW/h and another where it results in 74.02 kW/h, which obviously can’t be right. Maybe someone has some idea what could be going wrong here?

Normal Range:

Wrong Result:

This is my config:

#-------------------------------------------------------------------------------
    #LOCAL TUYA RIEMAN INTEGRATIONS
  - platform: integration
    source: sensor.laundry_dryer_power
    name: laundry_dryer_energy_spent
    unique_id: laundry_dryer_energy_spent_id
    unit_prefix: k
    round: 2

Thx a lot in advance.

EDIT: I tried to export the data from the SQLite Database but my query didn’t work, so I used alexarch21/history-explorer-card to export the data for the 19th of January as a .csv File. Next Step, integrate the Riemann sum myself.

Here is the data:

Time stamp,State
sensor.laundry_dryer_power
2023-01-18 16:02:07,0.0
2023-01-19 13:50:05,2220.8
2023-01-19 13:51:13,2141.4
2023-01-19 13:51:15,878.4
2023-01-19 13:51:17,26.3
2023-01-19 13:51:19,294.6
2023-01-19 13:51:49,2116.2
2023-01-19 13:52:19,7.2
2023-01-19 13:52:21,2170.1
2023-01-19 13:53:02,2840.2
2023-01-19 14:15:45,2138.6
2023-01-19 14:16:21,2615.4
2023-01-19 14:17:50,2405.5
2023-01-19 14:17:52,2116.2
2023-01-19 14:18:33,2273.8
2023-01-19 14:18:35,2840.2
2023-01-19 14:19:33,2767.4
2023-01-19 14:20:03,2116.2
2023-01-19 14:20:33,2835.3
2023-01-19 14:21:19,2840.2
2023-01-19 14:21:21,2502.2
2023-01-19 14:21:23,2121.8
2023-01-19 14:22:15,2251.6
2023-01-19 14:22:17,2840.2
2023-01-19 14:23:09,2099.8
2023-01-19 14:24:26,2835.3
2023-01-19 14:25:11,2121.8
2023-01-19 14:26:09,2144.3
2023-01-19 14:26:11,2449.2
2023-01-19 14:26:13,2870.5
2023-01-19 14:26:51,2791.3
2023-01-19 14:26:53,2113.5
2023-01-19 14:28:08,2119.0
2023-01-19 14:28:10,2387.8
2023-01-19 14:28:12,2840.2
2023-01-19 14:28:48,2734.7
2023-01-19 14:29:18,2133.0
2023-01-19 14:30:13,2434.5
2023-01-19 14:30:15,2830.3
2023-01-19 14:30:57,2835.3
2023-01-19 14:30:59,2586.2
2023-01-19 14:31:01,2116.2
2023-01-19 14:32:22,2573.8
2023-01-19 14:32:24,2840.2
2023-01-19 14:32:59,2791.3
2023-01-19 14:33:01,2135.8
2023-01-19 14:34:25,2105.2
2023-01-19 14:34:27,2791.3
2023-01-19 14:35:04,2744.0
2023-01-19 14:35:34,2102.5
2023-01-19 14:36:41,2119.0
2023-01-19 14:36:43,2423.5
2023-01-19 14:36:45,2825.4
2023-01-19 14:37:21,2187.7
2023-01-19 14:39:12,2734.7
2023-01-19 14:39:49,2573.8
2023-01-19 14:39:51,2127.4
2023-01-19 14:42:21,2094.3
2023-01-19 14:42:24,2767.4
2023-01-19 14:43:01,2452.9
2023-01-19 14:43:31,2121.8
2023-01-19 14:45:22,2510.0
2023-01-19 14:45:24,2815.5
2023-01-19 14:45:54,2387.8
2023-01-19 14:45:56,2088.9
2023-01-19 14:48:04,2475.4
2023-01-19 14:48:06,2796.1
2023-01-19 14:48:38,2083.6
2023-01-19 14:51:35,2152.8
2023-01-19 14:51:37,2835.3
2023-01-19 14:52:08,2097.1
2023-01-19 14:55:28,2102.5
2023-01-19 14:55:30,2791.3
2023-01-19 14:56:00,2108.0
2023-01-19 14:56:30,955.7
2023-01-19 14:57:00,187.3
2023-01-19 14:57:30,979.4
2023-01-19 14:58:00,2119.0
2023-01-19 15:01:59,2105.2
2023-01-19 15:02:01,1162.2
2023-01-19 15:02:31,954.0
2023-01-19 15:03:01,2105.2
2023-01-19 15:06:31,949.5
2023-01-19 15:07:07,1109.6
2023-01-19 15:07:09,2138.6
2023-01-19 15:10:58,977.6
2023-01-19 15:11:34,954.0
2023-01-19 15:11:36,2044.1
2023-01-19 15:16:56,2051.9
2023-01-19 15:16:58,981.7
2023-01-19 15:17:36,1889.1
2023-01-19 15:22:13,1318.3
2023-01-19 15:22:15,926.1
2023-01-19 15:22:45,173.7
2023-01-19 15:24:48,174.3
2023-01-19 15:24:50,195.9
2023-01-19 15:25:22,174.6
2023-01-19 15:28:19,198.0
2023-01-19 15:28:49,183.3
2023-01-19 15:30:38,92.3
2023-01-19 15:30:40,0.0
2023-01-19 15:31:10,199.8
2023-01-19 15:31:40,0.0
2023-01-20 00:44:01,unavailable
2023-01-20 00:44:03,0.0

EDIT2:
It looks like the faulty integrations are interpolating from the beginning of the day (midnight) until the first spike. It looks like this:

whereas a normal integration doesn’t do this:

[tom_l]

Use method: left just like the answer to this topic advises.

[hapklaar]

Great info, thanks. Now why did they remove that from the doc page…

[tom_l]

Not only is it still there, it has been expanded upon too: https://www.home-assistant.io/integrations/integration/#integration-method

[hapklaar]

There is nothing in there that mentions spikey consumption anymore. They might mean something like that with the current wording, but it is nowhere as clear as what is in your screenshot.

[tom_l]

Spikey was never the correct term. “Rectangular” is much better.