Protocol conversion

The collector is also designed to act as a protocol converter, that is, it can read the data in OPC-DA or OPC-UA format and deliver it on WITSML format. That action could be performed in any deployed instance.

OPC to WITSML Architecture

The OPC to WITSML conversion works by running a basic WITSML server backed by a PostgreSQL or TimescaleDB database, where the Live Rig Collector will query the OPC server and expose its data as an WITSML server.

To configure this feature, you need to configure an OPC-DA or OPC-UA Source on the LiveRig Collector. For more information, see the OPC-DA or OPC-UA pages, along with the sources.xml and store.json pages for file configurations.

The store.json file contains 4 fields exclusively for the OPC to WITSML conversion feature. These are: database, endpoint, limit and purge.

Database

The database field is required for connecting to the local database instance used in the conversion. To specify which database will be used (PostgreSQL, TimescaleDB), just configure a flag inside the parameter field of the database section.

Example:

"database": {
    "url": "jdbc:postgresql://localhost:5432/?user=root&password=rootpassword",
    "parameters": {
        "timescale": false
    }
}

The default database is PostgreSQL, which means that if configuration parameters are not provided, or if it's provided with the timescale flag set to false, PostgreSQL will be used.

By enabling TimescaleDB, we can take advantage of its compression features, capable of decreasing the total amount of used disk space. To install a TimescaleDB instance, you need to configure the extension accordingly. See Installing TimescaleDB for further details.

TimescaleDB Configuration

A valid TimescaleDB configuration looks like the following:

"database": {
    "url": "jdbc:postgresql://localhost:5432/?user=root&password=rootpassword",
    "parameters": {
        "timescale": true,
        "timescale.chunk_interval": 604800000,
        "timescale.compress_after": 3600000
    }
}

• Chunk Interval: Hypertables in TimescaleDB are automatically partitioned into smaller pieces, called chunks. Each chunk contains a specific amount of data, defined by chunk interval configuration. Behind the scenes, each chunk is the smallest portion of data that can be compressed and decompressed. timescale.chunk_interval setting is expressed in milliseconds, and defaults to 7 days (604800000 ms).

• Compress After: Represents the amount of time after which the hypertable chunks will be automatically compressed in the background. A recurrent policy is set to compress every chunk containing data older than this configuration. timescale.compress_after setting is also expressed in milliseconds, and defaults to 1 hour (3600000 ms).

Endpoint

The endpoint field is necessary for exposing an WITSML Server endpoint with the collected data. Example:

"endpoint": "http://127.0.0.1:1234/witsml/store"

Limit

The limit field is not a necessary field. Its purpose is to limit the number of values to be returned on a request to the WITSML store. The default value is 1000.

Purge

The purge field is not a necessary field. Its purpose is to set a period to purge old values from the WITSML store (to avoid the collector's disk filling up). This is calculated using the following formula: CURRENT_TIMESTAMP - PURGE_INTERVAL. This interval is used in seconds. Example: If the purge field value is 1000, that means that values older than 1000 seconds from the current time will be deleted. The default state of this feature is off.

After configuring the OPC to WITSML conversion an WITSML server endpoint will be created. With this new endpoint you can configure a new WITSML Source at a LiveRig Collector or use a WITSML Store for fetching data.

When used in pure protocol conversion mode, the collector does not need the presence of a centralized Intelie Live environment.

Limitations

1. Unit conversion is not supported.

2. The actual data will be available as WITSML date-time logs only (no depth support).

3. A straightforward WITSML well and wellbore structure will be used to represent the actual data as WITSML logs.