Smart Water - Turbidity sensor

Post Reply
swe
Posts: 24
Joined: Tue Dec 11, 2018 12:29 am
Company: -

Smart Water - Turbidity sensor

Post by swe » Mon Jul 01, 2019 1:15 am

Hi,

We have observed some spikes (most of the time just 1 or 2 spikes) of Turbidity sensor readings in Smart Water (Plug&Sense) from time to time. The following displays two examples. Note that the spike readings could vary from 100 to 500, which should be noises. What could be the reasons and corresponding solutions?

2019-06-30 06:48:05 10.90
2019-06-30 06:33:05 11.40
2019-06-30 06:17:05 12.60
2019-06-30 06:02:05 164.90
2019-06-30 05:47:05 160.80
2019-06-30 05:32:05 9.70
2019-06-30 05:17:05 10.00


2019-06-30 04:17:20 10.20
2019-06-30 04:02:20 10.90
2019-06-30 03:47:20 12.10
2019-06-30 03:32:20 311.10
2019-06-30 03:17:20 10.20
2019-06-30 03:02:20 8.00

Thanks.

libelium-dev
Posts: 27967
Joined: Mon Sep 28, 2009 1:06 pm

Re: Smart Water - Turbidity sensor

Post by libelium-dev » Mon Jul 01, 2019 11:06 am

Hi,

Where is the sensor installed?

The interferences could appear due to the distance with the walls or if there are water pumps close to the sensor. Or maybe there are solid particles at that moment.

You could try to perform some measurements and perform the average to avoid those measurements.

Regards

swe
Posts: 24
Joined: Tue Dec 11, 2018 12:29 am
Company: -

Re: Smart Water - Turbidity sensor

Post by swe » Mon Jul 01, 2019 11:43 pm

The sensor is installed in the field.

There are no water pumps close to the sensor. There should be no solid particles at that moment either.

What's the exact interference due to the distance with the walls? Do you have some data from labs or field trials? What's the best practice to deploy the sensor to avoid such interferences (if any)?

Could you please elaborate a bit more about "perform some measurements and perform the average to avoid those measurements"? How to perform the average with those spike readings?

Thanks.

libelium-dev
Posts: 27967
Joined: Mon Sep 28, 2009 1:06 pm

Re: Smart Water - Turbidity sensor

Post by libelium-dev » Tue Jul 02, 2019 8:50 am

Hi.

The interferences with walls can occur during calibrations procedure since it is done in small bottles but if the sensor is deployed in the field without any object close it shouldn't occur.

You can simply perform ten measurements and perform the average

Code: Select all

  float turVal = 0.0;
  for(uint8_t i=0; i<10; i++){
    turVal = turVal + turbidity.getTurbidity();
    delay(200);
  }
  turVal = turVal/10;

  USB.print(F("Turbidity Value: "));
  USB.print(turVal);    
  USB.print(F(" NTU | "));  
Or you can discard values when they are 10 times higher than the previous value.

Code: Select all

  float turVal, previousVal;
  
  do{
    turVal = turbidity.getTurbidity();
  }
  while(turVal/10 > previousVal);

  USB.print(F("Turbidity Value: "));
  USB.print(turVal);    
  USB.print(F(" NTU | "));    

  previousVal = turVal;

Regards

swe
Posts: 24
Joined: Tue Dec 11, 2018 12:29 am
Company: -

Re: Smart Water - Turbidity sensor

Post by swe » Tue Jul 02, 2019 11:38 pm

There are multiple sensors deployed at the same time in the field and they are quite close. What's the minimum distance to avoid the inference? What's your best practice to deploy multiple sensors simultaneously in the field?

It may be better to use mean value instead of average since the spike values are most likely false readings. Right?

It's not safe to discard values when they are 10 times higher than the previous value because the environment may change during the time interval.

Thanks.

libelium-dev
Posts: 27967
Joined: Mon Sep 28, 2009 1:06 pm

Re: Smart Water - Turbidity sensor

Post by libelium-dev » Wed Jul 03, 2019 8:49 am

Hi,

We recommend that there be a distance of about 20 centimetres between the sensors to avoid interference.

We agree. It will depend on the final code. If the time between measures is low you could discard values very different but when the time between measurements is long the environmental conditions may change. Maybe the best option is mean and discards noise values.

Regards

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest