Operational field data is often stored in systems that make accessing, analyzing, and sharing data outside of that system quite difficult. Operators need a data historian in place that allows operational field data to be easily accessed and analyzed by anyone who needs the data and easily integrated with other software solutions. Hence, data can feed formal analytics and machine learning initiatives and gain greater operational insights.
The eLynx Historian provides operators with a tool set that is designed for modern data initiatives. Operators can easily inbound and outbound data with other cloud solutions and software applications, including the following solutions: Spotfire, ARIES, Excel, Informatica, Snowflake, and P2. With the eLynx Historian solution, operators can trend data over time at a well, pad, and gathering system level, and query data based on metadata associated with each asset type.
Many operators are seeking to leverage machine learning to obtain increasingly advanced warning of production issues. eLynx has discovered there is a new and critical tool set required for operators to obtain the data they need to feed machine learning and to incorporate this type of data into their operations.
To utilize machine learning with granular time series data sets obtained from SCADA, you must label that data set at a time series granularity, rather than a daily granularity. Most operators have, at best, daily well failure data. If you want time series SCADA data fed into machine learning, you must note the near exact start-time and near exact end-time of the events you want to predict.
The eLynx system allows users to label time series data manually or by creating rules which specify the start-time criteria and end-time criteria. As data comes into the eLynx Historian, the Rules Engine evaluates the data. If any criteria are met, an event is started or ended. The Rules Engine allows engineers to specify criteria beyond traditional SCADA alarms. The rules can include multiple variables with different durations and mathematical functions like slope changes, independent or in relation to another variable (deviating/converging), sharp increase/decrease, gradual increase/decrease, etc. The rules can also scale across different wells, as you can modify the thresholds on a per well basis.
The eLynx system allows engineers to run rules across historical data sets to review the results prior to deploying the rule to operations. The ability to do this helps engineers verify the expected efficacy of a rule.
The eLynx system allows operators to validate results from each rule as valid or invalid. This provides a necessary feedback loop between the operators and the engineers so that rules can be modified as needed to improve the efficacy of the results. If the performance of a rule on a specific well is degrading or not providing meaningful results, engineers can adjust the rule thresholds or remove the rule altogether.
Rules can ultimately replace alarm setpoints by providing a diagnostic rather than an alarm for each individual symptom. Current SCADA alarms place a heavy burden on operators by sending individual symptoms over time and requiring the operator to aggregate the symptoms together to determine the diagnosis. Operators are being overwhelmed with alarm notifications, resulting in alarms being missed. The ability to reduce notifications by any amount has a huge impact on operational efficiency.
The eLynx system has built-in data quality functions that can be incorporated into rules, so if there is an issue with data quality, the system is not evaluating a rule against invalid data and thus sending out invalid results. The data quality functions include stuck sensors, out of range readings, and communications losses.
The eLynx system has a built-in production dashboard to view any sub-section of your assets—from a single pumper route up to an entire operating area. The production dashboard quickly shows which wells are producing, which wells are down, and which wells are at risk of going down, based on the rules you have defined. In addition, the associated target production for each well is specified. Meaning, operators can immediately identify the greatest production impact and prioritize accordingly. This view is also available in the mobile app.
The eLynx mobile app makes it easy for operators to capture much needed contextual data that is not captured via automation, such as downtime codes and field notes. Operators can easily capture field notes by speaking into their phone using the voice to text option. The field notes are instantly captured in the eLynx system so engineers can see the operator’s notes alongside the graphs they are viewing.
The eLynx Historian service provides operators with a modern operational data platform that enables your team to automate the detection, classification, and notification of well conditions that your engineers and operators look for manually today. In addition to the immediate operational efficiency gains, over time you are creating a validated labeled time series data set that can be leveraged with machine learning.