ESS Inverter response too slow - tripping utility meter

Question

question

Nicholas Herbst asked Jul 2, '20

ESS Inverter response too slow - tripping utility meter

Hello Community,

Our company is struggling with a number of installations in the Tshwane (Pretoria) region of South Africa. The municipality (utility provider) has rolled out the installation of Conlog meters. The majority of these meters are overly sensitive to reverse energy detection. One of the meter models is the Conlog BEC42X (Single Phase).

Many of our installations consist of an ESS setup like the following: MultiPlus II 5kVA, BlueSolar MPPT Charge Controller, Color Control GX, Carlo Gavazzi Energy Meter, BMV700 and a Lithium Battery.

Many of these installations, in the Tshwane area, are experiencing frequent tripping of the Conlog utility meter.

Some forums have indicated that these meters are tripping at a point as low as 20W over 15 seconds - or some 300 Joules of reverse (export) energy detected. This is incredibly sensitive and clearly the root problem. However, this is very difficult to resolve with the municipality (utility provider) and they are mostly unwilling to assist.

We have had to resort to setting the grid set point significantly higher than desired, sometimes around 300W and have had to limit inverter power, in some instances to as low as 750W. With these restrictions placed on the system, we are mostly unable to process (store and discharge) all of the solar energy available from the solar array resulting in compromised economic performance.

What can we do to improve the control of export power to the utility?

It is inevitable that some export energy will be supplied to the utility in a grid-connected ESS system, but one observation that we have made is that the MultiPlus & Quattro inverters appear to react quite slowly AFTER a large load is switched off - this observation comes from the Color Control display panel where we can see the grid import/export as reported by the Carlo Gavazzi Energy meter, as well as the amount of power being discharged by the battery, as reported by the BMV700.

Once a large load is dropped, can the inverter not reduce it's power faster to avoid overshoot and subsequent export energy being supplied to the utility?

From this link, we see that the response time of ESS has been limited to 400W per second as per ESS version 162... but now we need a work around?

This thread comments on a similar aspect: https://community.victronenergy.com/questions/51697/victron-ess-et112-slow-response.html

After reading the above thread, can we not adjust the refresh rate and baud rate for the energy meter (Carlo Gavazzi) to aid the situation in swiftly recognizing the change in the grid import/export condition?

Alternatively, is the 100A:50mA Current Transformer a temporary solution for MultiPlus II ESS installations?

If this cannot be achieved with the ESS assistant, is there a combination of assistants that will allow us to operate in a similar fashion to ESS while overcoming the issues presented by the hard-coded rate limiting?

Thanks in advance for your input and suggestions.

Multiplus-II ESS Energy Meter

1

2 |3000

Attachments: Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

solarhack commented · Jul 05, 2020 at 10:34 AM

I would really like to see Victron address this problem better. I think they could do a few things :

Allow the Baud rate of the Energy Meter to be adjusted. The Carlo Gavazzi supports multiple rates. This would probably affect distance so should be defaulted to a low rate but allow you to adjust it up. I'm not sure how much better this will make the response but it would be worthwhile testing.
Allow multiple grid point settings based on time or even battery level or both. This would mean you don't have to site with a high grid point indefinitely and can set it to a higher point at critical points.
Speed up the Inverter response to changes. This is probably more complex and may not be something they can do in a Firmware fix. In the linked post the person talks about better responses from the CT, which would result in less feedback to the grid, maybe you should give it a test.
- https://community.victronenergy.com/questions/51697/victron-ess-et112-slow-response.html .
- @Daniël Boekel (Victron Energy Staff) did say they are aware of issues around this and are working on a fix.

I think combination of these would go a long way to give you better control of what gets fed back to the grid

Regards

0 ·

Modifications Space

Node Red Space

Answer 1 · 2020-07-04T19:40:46Z

Alexandra answered · Jul 04, 2020 at 07:40 PM

@Nicholas Herbst

AC ignore is the other option.

https://www.google.com/url?q=https://www.victronenergy.com/live/assistants:ignore_ac_input_using_the_generator_assistant&sa=U&ved=2ahUKEwjUtLa4qrTqAhXLiFwKHa4UDh4QFjABegQIChAB&usg=AOvVaw1ivo1FZZM85omVUt3mDZL0

Where you can set conditions and state of charge for it to reconnect to grid for use like a generator.

This should not be used with was however as it is intentional islanding.

My other thought was when is the feedback in these systems worse? Is it when batteries are full by any chance? We have in some cases found that not having full batteries helps as the excess pv and feedback from loads dropping off has somewhere to go.

2 |3000

Attachments: Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Answer 2 · 2020-07-06T09:46:14Z

Daniël Boekel (Victron Energy Staff) answered · Jul 06, 2020 at 09:46 AM

Hi @Nicholas Herbst @SolarHack

Even with improvements in speed we are working on, these meters trip too fast.

One solution that Alexandra also gave: do not use ESS but use the grid like you'd use a generator:

https://www.victronenergy.com/live/assistants:ignore_ac_input_using_the_generator_assistant

note that I'd now use the generator start/stop function of the GX device, connected to the inverters to ignore AC input.

The other solutions is to ask the utility company for an exception so the meter doesn't trip.

setting a higher baudrate for the energy meter doesn't help, because it only sends an update once every 750 ms

2 |3000

Attachments: Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Answer 3 · 2020-07-10T08:11:51Z

Nicholas Herbst answered · Jul 10, 2020 at 08:11 AM

We have tried to implement the Current Transformer as a replacement for the ET112 in installations involving the MultiPlus II.

The observed performance of the CT is far better than the ET112, and we essentially do not get any further tripping of the Utility meters.

Unfortunately, we discovered a "fatal" error that effects this configuration (Using the CT) where the inverter enters an Error 24 mode - System protection initiated. I am going to upload a video to YouTube and create a thread specifically related to this new problem.

3

2 |3000

Attachments: Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Nicholas Herbst commented · Jun 14, 2021 at 04:14 AM

Update: Error 24 was resolved by a firmware update to the MultiPlus II.

0 ·

Nicholas Herbst commented · Jun 14, 2021 at 04:15 AM

The CT for the MultiPlus II remains a valid alternative to the Grid Meter - it results in far faster response to changing load profile at the site.

0 ·

Jay Nicholas Herbst commented · Dec 05, 2022 at 02:23 AM

@Daniël Boekel (Victron Energy Staff) when will this issue finally be resolved to permit faster responses off the ET112 (which is £100 piece of kit that I've just ordered for this purpose)?

0 ·

Answer 4 · 2020-08-30T09:19:21Z

I am running into this same issue (I think) as originally posted at the top. I get it mainly when powered by battery (e.g. at night). I can 100% reproduce it every time, and it's extremely annoying: Turn on the oven (which is on non-essential AC load circuit), while it's coming up to temperature all is fine (drawing power from the battery primarily, as seen on Victron remote console). As soon as it comes to temperature, it turns off the element, which results in power being accidentally fed back into the grid (I've seen as much as -1800W shown on the Victron remote console), causing the Conlog to switch off the grid supply, which in turn causes the oven to lose power, and then a few seconds later the grid power comes back on, and the oven resets it's time and timer and cooking status, therefore remains off. It does this with other non-essential AC load things too, like the geyser or stove, but most of those are stateless so it doesn't matter as much (that is, when power is restored, they carry on doing what they were doing). The oven is very vexing though, since it means currently we cannot use it at all to cook food, because it never remains on. It also beeps quite loudly when power is restored, so we hear it regularly during the day when it is reset by other things like the geyser too.

I've set the grid set point to as high as 300W, and it still doesn't solve the issue (but it occurs somewhat less frequently than it did on the 50W as commissioned by installer). I don't really want to set the grid point this high (or higher) of course, as I'm increasingly wasting the potential of the ESS then.

Before I go down the arduous and expensive route of convincing the municipality to install a different type of prepaid electricity meter, I'd like to know whether this is perhaps something Victron can and will fix in firmware in the very near future @Daniël Boekel (Victron Energy Staff)? If not, I guess I'll have to try to get a Landis&Gyr prepaid meter (which is apparently less sensitive), but it feels to me like I'm having to pay a bunch of money for a workaround to a Victron shortcoming basically...

--

Victron Multiplus II 5000, 4x 3.5kWh pylontech, 150/100 Victron MPPT, 12x 395W PV panels (3s4p)

question