66. Peak Tracking¶

The goal of this recipe is to track the peak (maximum) value of a property within a moving bounded time period, such as every 15 minutes. We want to monitor the property at a higher frequency than uploading to SolarNetwork, however, taking a reading every second but reporting the data only every 5 minutes. Some meter devices can do this sort of tracking for you, but this recipe can work with any device supported by SolarNode.

To illustrate the goal, imagine SolarNode is reading a watts property from a meter, and we'd like to track its maximum value in 15-minute periods on a wattsMax_p15 property. Here is a simplified table of example datum we would like to see uploaded to SolarNetwork over the course of an hour. The Timestamp values in bold denote the first reading in each 15-minute period. The wattsMax_p15 values in bold denote the last reading in each 15-minute period, and correspond to the peak value within that period:

66.1 Overview¶

To satisfy our goal, SolarNode must be able to "remember" the maximum value seen on the watts property. It must also know when a datum falls within a new 15-minute period, as compared to the 15-minute period associated with the last "seen" datum, so it can reset the "remembered" peak value and start the tracking cycle over again.

With the help of expresions making use of Local State entities, SolarNode can track when a datum is captured on a new 15-minute period, to facilitate resetting the "remembered" peak value. The peak value is itself stored as a Local State entity, so it is "remembered" and can be compared with watts property values over time to track its peak value.

For each datum that is evaluated, SolarNode will execute the following logic:

1. if the datum is starting a new 15-minute period, then the maximum value Local State should be set to this datum's watts value
2. else the maximum value Local State should be updated to the larger of this datum's watts value and the Local State's value

The final wattsMax_p15 property value then becomes the Local State's final value.

The overall steps of this recipe are:

• add a Parameter Expression filter PEX - 15min Period to track 15-minute periods as a parameter value.
• add an Expression filter EXP - 15min Peak to calculate the wattsMax_p15 property
• add a Downsample tiler DWS - Meter to downsample the 1-second resolution meter data down to 5-minute resolution
• add all three of these added filters to the Main filter chain

66.2 Parameter Expression¶

Parameters created by the Parameter Expression filter are variables available to subsequent filters within the filter processing pipeline. They do not become datum properties, and thus act like temporary variables. We will use a Parameter Expression to create a p15min_changed parameter that will be true when a new 15-minute period has started or false otherwise.

In order to know if the current 15-minute period is new, we need to be able to compare it to the 15-minute period the previous datum fell within. To do this, we'll calculate the start of the current 15-minute period as a Unix epoch seconds number, and save that on a p15min-timestamp Local State entity. We can use the result of the changeLocalState() function to tell us when the p15min-timestamp has changed, thus signaling that a new 15-minute period has begun.

Create a Parameter Expression filter component:

1. Go to the Settings > Datum Filters page.
2. Scroll down to the User Datum Filters section.
3. Click the Manage button next to the Parameter Expression Datum Filter item.
4. Click the + Add new Parameter Expression Datum Filter button.

Configure the component with the following settings:

Add a new Expression by clicking on the + button next to Expressions, then configure these expression settings:

The expression is this:

changeLocalState('p15min-timestamp', epochSecs(dateFloor(timestamp(), 'PT15M')))

The component settings should look like this:

66.3 Expression¶

The next step is to configure an Expression filter to calculate our desired wattsMax_p15 datum property. It will make use of the p15min_changed parameter provided by the PEX - 15min Period Parameter Expression filter configured in the previous section, as well a p15min-mtr1-watts-max Local State entity to track the maximum watts value.

Create an Expression filter component:

1. Go to the Settings > Datum Filters page.
2. Scroll down to the User Datum Filters section.
3. Click the Manage button next to the Expression Datum Filter item.
4. Click the + Add new Expression Datum Filter button.

Configure the component with the following settings:

Add a new Expression by clicking on the + button next to Expressions, then configure these expression settings:

The expression is this:

1
2
3
4
5

has('watts') && has('p15min_changed') && p15min_changed
? saveLocalState('p15min-mtr1-watts-max', watts)
: has('watts') && has('p15min_changed')
? saveLocalState('p15min-mtr1-watts-max', max(watts, localState('p15min-mtr1-watts-max', watts)))
: null

The component settings should look like this:

66.4 Downsample¶

The next step is to configure an Downsample filter to aggregate our 1-second resolution meter readings down to 5-minute readings. This allows us to track the maximum watts values at a much higher frequency that we post up to SolarNetwork.

Create a Downsample filter component:

1. Go to the Settings > Datum Filters page.
2. Scroll down to the User Datum Filters section.
3. Click the Manage button next to the Downsample Datum Filter item.
4. Click the + Add new Downsample Datum Filter button.

Configure the component with the following settings:

The component settings should look like this:

66.5 Configure Main filter chain¶

Now all that remains is to add our 3 newly configured filters to the Main filter chain.

1. Go to the Settings > Datum Filters page.
2. Scroll down to the User Datum Filters section.
3. Click the Manage button next to the Datum Filter Chain item.
4. Click the + button next to Datum Filters 3 times to add 3 new references.
5. Click the Choose button in each of the 3 reference settings to select the appropiate filters.

The order we configure the filters is important here:

1. First PEX - 15min Period because EXP - 15min Peak depends on its output
2. Next EXP - 15min Peak because it must execute on the 1-second resolution data, before the DWS - Meter filter aggregates it down to 5-minute resolution
3. Finally the DWS - Meter to downsample the 1-second resolution data down to 5-minute resolution

The component settings should look like this:

Congratulations! That completes this recipe.

66.6 Monitoring the Local State¶

Once you have this recipe deployed, you can use the Local State management page to monitor their values. You should see entities like the following:

The values will change over time, as datum are processed and the expressions configured for this recipe perform their work.

66.7 Validating the results¶

If you export this datum stream as CSV from SolarNetwork, you can open it in a spreadsheet application like Excel to vaildate the calculations are working as expected. In the following screen shot, the MAX(watts) column is a MAX() calculation on the watts values collected during 15-minute time periods.