Mitsubishi Ecodan &...
 
Notifications
Clear all

Mitsubishi Ecodan & Raspberry Pi automations

91 Posts
11 Users
34 Reactions
10.2 K Views
(@kev-m)
Famed Member Moderator
5561 kWhs
Joined: 3 years ago
Posts: 1299
 

Derek,

Mitsubishi claim there is some sort of 'learning' built in to Auto Adaptation.  I assume this means that if it does overshoot once or more then in future similar conditions it will adjust its response to stop this happening.  I've no idea if this happens or not but have just started to use Auto Adaptation in anger and will do some analysis to try and see how it works. 

 


   
ReplyQuote
(@derek-m)
Illustrious Member Moderator
13948 kWhs
Veteran Expert
Joined: 3 years ago
Posts: 4196
 

@kev-m

It will be interesting to see what actually happens, and how well the controller actually 'learns'.


   
ReplyQuote
(@sunandair)
Prominent Member Member
2538 kWhs
Joined: 1 year ago
Posts: 345
 

@kev-m 

watching with interest 

what have you done so far? What parameters are you setting? 
also are you moving dip switches yourself eg sw5 no2 for activation of auto adaptation and sw1 no8 to power the wireless thermostat (if you’re using it) 

how do you adjust the various other other settings?

Wishing you every success. 👍🏼


   
ReplyQuote
(@kev-m)
Famed Member Moderator
5561 kWhs
Joined: 3 years ago
Posts: 1299
 

Posted by: @sunandair

@kev-m 

watching with interest 

what have you done so far? What parameters are you setting? 
also are you moving dip switches yourself eg sw5 no2 for activation of auto adaptation and sw1 no8 to power the wireless thermostat (if you’re using it) 

how do you adjust the various other other settings?

Wishing you every success. 👍🏼

@sunandair,

yes I moved the dip switches (well one actually, the auto adapt one was on by default).  I have 2 zones; one with the wireless thermostat and one with a remote thermistor.  They are both currently physically in the same room so I'm effectively running the house as one zone based on the temperature in my living room. It's set at 21 deg daytime and 18 at night.  

I've left the auto adapt settings default.

And thanks, I'll be sure to report back; I'll start a new thread.  If you want a teaser, so far there are some interesting positive changes in the way things are working.  

 


   
Derek M and SUNandAIR reacted
ReplyQuote
(@scrchngwsl)
Estimable Member Member
1469 kWhs
Joined: 2 years ago
Posts: 90
 

Posted by: @kev-m

Derek,

Mitsubishi claim there is some sort of 'learning' built in to Auto Adaptation.  I assume this means that if it does overshoot once or more then in future similar conditions it will adjust its response to stop this happening.  I've no idea if this happens or not but have just started to use Auto Adaptation in anger and will do some analysis to try and see how it works. 

I wonder if this means that, in addition to the PID-style feedback system, it also has a "model" for heat loss, which it populates with the data that it collects? These devices surely collect better data and over a longer period of time than anything you could do on a heat-loss study or even a controlled experiment on a single day. It wouldn't be difficult to have an model-based analytical solution in addition to a feedback-based numerical solution.

 

ASHP: Mitsubishi Ecodan 8.5kW
PV: 5.2kWp
Battery: 8.2kWh


   
ReplyQuote
(@ajdunlop)
Eminent Member Member
255 kWhs
Joined: 1 year ago
Posts: 17
 

Posted by: @kev-m

And thanks, I'll be sure to report back; I'll start a new thread.  If you want a teaser, so far there are some interesting positive changes in the way things are working.

The waiting is killing me, any chance of some findings so far?

 

I went a bit mad with the settings last week (should have changed one at a time to be able to know what is changing what).

I limited the max flow temperature in auto-adaption to 40C, put the interval back down to 20mins (I might put this up again next week to see what happens) and also lowered the pump speed which has widened the delta T to around 5 now.

 

This is then what I was getting last week:

Screenshot 2023 01 18 100102

There was less of a huge spike in flow temperature and power consumption at the start of each cycle and each one carried on a bit longer.

 

This last week days this is what I've been getting:

Screenshot 2023 01 18 100126

I assume because of the colder temperatures the HP has been staying on a bit more constantly (note that last night I upped the set back temperature as yesterday it took a long time to get up to temperature). 

 

Looking at all this compared to when I was using pure WC the system does seem to be much more aggressive.

I might change to WC today to see what happens as the outdoor temperature is to be similar to yesterday to see how it behaves.


   
ReplyQuote



(@derek-m)
Illustrious Member Moderator
13948 kWhs
Veteran Expert
Joined: 3 years ago
Posts: 4196
 

Posted by: @ajdunlop

Posted by: @kev-m

And thanks, I'll be sure to report back; I'll start a new thread.  If you want a teaser, so far there are some interesting positive changes in the way things are working.

The waiting is killing me, any chance of some findings so far?

 

I went a bit mad with the settings last week (should have changed one at a time to be able to know what is changing what).

I limited the max flow temperature in auto-adaption to 40C, put the interval back down to 20mins (I might put this up again next week to see what happens) and also lowered the pump speed which has widened the delta T to around 5 now.

 

This is then what I was getting last week:

Screenshot 2023 01 18 100102

There was less of a huge spike in flow temperature and power consumption at the start of each cycle and each one carried on a bit longer.

 

This last week days this is what I've been getting:

Screenshot 2023 01 18 100126

I assume because of the colder temperatures the HP has been staying on a bit more constantly (note that last night I upped the set back temperature as yesterday it took a long time to get up to temperature). 

 

Looking at all this compared to when I was using pure WC the system does seem to be much more aggressive.

I might change to WC today to see what happens as the outdoor temperature is to be similar to yesterday to see how it behaves.

I would suggest setting the Interval Timer to 60 minutes to help reduce the cycling.

 


   
ReplyQuote
(@kev-m)
Famed Member Moderator
5561 kWhs
Joined: 3 years ago
Posts: 1299
 

Posted by: @ajdunlop

Posted by: @kev-m

And thanks, I'll be sure to report back; I'll start a new thread.  If you want a teaser, so far there are some interesting positive changes in the way things are working.

The waiting is killing me, any chance of some findings so far?

 

I went a bit mad with the settings last week (should have changed one at a time to be able to know what is changing what).

I limited the max flow temperature in auto-adaption to 40C, put the interval back down to 20mins (I might put this up again next week to see what happens) and also lowered the pump speed which has widened the delta T to around 5 now.

 

This is then what I was getting last week:

Screenshot 2023 01 18 100102

There was less of a huge spike in flow temperature and power consumption at the start of each cycle and each one carried on a bit longer.

 

This last week days this is what I've been getting:

Screenshot 2023 01 18 100126

I assume because of the colder temperatures the HP has been staying on a bit more constantly (note that last night I upped the set back temperature as yesterday it took a long time to get up to temperature). 

 

Looking at all this compared to when I was using pure WC the system does seem to be much more aggressive.

I might change to WC today to see what happens as the outdoor temperature is to be similar to yesterday to see how it behaves.

It's better in warmer weather.  Less cycling and a smoother, less spiky start up in the morning.  

My tool for analysing the data is playing up but I'll post more when I get it working as it's a pain to do manually.  

However, like you, I am finding it very aggressive in the cold weather.

 

 


   
ReplyQuote
(@derek-m)
Illustrious Member Moderator
13948 kWhs
Veteran Expert
Joined: 3 years ago
Posts: 4196
 

@kev-m

Hi Kev,

What is the setting of the Interval Timer?


   
ReplyQuote
(@kev-m)
Famed Member Moderator
5561 kWhs
Joined: 3 years ago
Posts: 1299
 

Posted by: @derek-m

@kev-m

Hi Kev,

What is the setting of the Interval Timer?

10 minutes.  I'd never looked before. It prompted me to limit the ASHP's LWT to 48 degrees. It was 55 and it has been that high a few times (auto adaptation).

  

This is what I'm talking about.  The smooth one is AA, the bumpy one is WC.  Similar ambient and room temps.  Once I fix my reporting tool I'll do some analysis on the colder weather.  The peaks at 9 and 3 are DHW.

Screenshot 2023 01 13 17.01.05
Screenshot 2023 01 13 16.59.54

 

 

 


   
Derek M reacted
ReplyQuote
(@ajdunlop)
Eminent Member Member
255 kWhs
Joined: 1 year ago
Posts: 17
 

Carrying on the conversation about Auto Adaption mode.

 

Our office Ecodan has been running in AA mode for a while.

It is set to ramp up from a 19 degree setback temperature at night to 21 degrees. I do this by increasing by 0.5 every 45mins from 05:45 reaching target temperature of 21 at 08:00. I do this using the MelClould timers. This appears to help to flatten the initial spike in power use when I just go straight from 19 to 21.

 

What I have been noticing recently (although not a new problem) is that it appears to be operating too high and overshooting the 21. Looking at today you can see that we reach 21 at just after 09:00 but the ecodan keeps putting heat in until just before 11:00.

Screenshot 2023 02 21 at 18.16.46

Interval rate is 40 mins but I'm not sure if this effects deciding to stop inputting heat or just how quickly to ramp up the flow temperature when getting up to target temperature.

Also I know that the HP can go lower than the consumed power shown so it isn't that it is going as slow as it can.

 

Any ideas?


   
ReplyQuote
(@derek-m)
Illustrious Member Moderator
13948 kWhs
Veteran Expert
Joined: 3 years ago
Posts: 4196
 

Posted by: @ajdunlop

Carrying on the conversation about Auto Adaption mode.

 

Our office Ecodan has been running in AA mode for a while.

It is set to ramp up from a 19 degree setback temperature at night to 21 degrees. I do this by increasing by 0.5 every 45mins from 05:45 reaching target temperature of 21 at 08:00. I do this using the MelClould timers. This appears to help to flatten the initial spike in power use when I just go straight from 19 to 21.

 

What I have been noticing recently (although not a new problem) is that it appears to be operating too high and overshooting the 21. Looking at today you can see that we reach 21 at just after 09:00 but the ecodan keeps putting heat in until just before 11:00.

Screenshot 2023 02 21 at 18.16.46

Interval rate is 40 mins but I'm not sure if this effects deciding to stop inputting heat or just how quickly to ramp up the flow temperature when getting up to target temperature.

Also I know that the HP can go lower than the consumed power shown so it isn't that it is going as slow as it can.

 

Any ideas?

This could be one of the problems with AA, or any other type of modulating control system, as the weather gets milder.

The problem is that the operating parameters are changing, which can affect how the system responds.

In colder weather conditions, the heat demand is higher, so if the heat source is removed (setback), the indoor temperature will reduce faster than if it was milder outside. When the heat pump restarts, it will undoubtedly have more work to do since the building has probably lost more heat energy during the setback period. Unfortunately, I don't know how the control algorithm inside the controller works, but I assume it takes the outside air temperature measurement along with the indoor air temperature measurement, and along with the settings of the WC curve, and then decides upon the required LWT to bring the indoor temperature back to the desired setting, within a reasonable period of time. If, at the present outside temperature, a LWT of 35C would be required to maintain an indoor temperature of 21C, but this would take say 5 hours to bring the indoor temperature back up to 21C from a setback of 19C, the controller may decide that a LWT of 38C is preferable to 35C, and may then achieve the desired indoor temperature within only 2 hours.

In milder conditions, the controller may still be increasing the LWT to 3C above the actual required LWT during the reheat period, but the heat loss and heat demand is now reduced, so the building will heat up faster, with the likelihood that the indoor temperature will overshoot the desired value.

Industrial control system are often fine tuned to produce a limited overshoot, since this normally speeds up the process of achieving the desired setting in the shortest time, and also makes the system more responsive to changing conditions.

I would suggest that you check the settings of the WC curve, since this could be causing a higher LWT than required. You could also try different settings for the Time Interval, which I do believe has an effect upon the response time of the system.

 


   
mjr reacted
ReplyQuote



Page 5 / 8
Share:

Join Us!

Latest Posts

Heat Pump T-Shirts

Delta T Sounds Greek to Me
x  Powerful Protection for WordPress, from Shield Security
This Site Is Protected By
Shield Security