System and method for detecting, diagnosing, and correcting trips or failures of electrical submersible pumps

A method for predicting and preventing electrical submersible pump (ESP) trip and failure events, the method comprising:

• collecting real-time data from a well and an ESP positioned in a well via a plurality of sensors;
• manipulating historical data to determine and evaluate a plurality of decision variables;
• using mathematical models to derive trends and patterns based on a relationship between the plurality of decision variables and behavior of the ESP;
• associating the trends and patterns with trip and failure events;
• diagnosing causes of the trip and failure events;
• suggesting remedial action to prevent the trip and failure events;
• and
• determining a stable operating range of each of the plurality of decision variables;
• wherein the historical or real-time data includes (i) time-series of information of the plurality of decision variables across surface, wellbore, and downhole gauges and (ii) on whether a trip or the failure occurred.

A system for monitoring performance of an electrical submersible pump (ESP) in a well, the system comprising:

• a sensor configured to sense real-time data;
• and
• a controller having at least one processor used to control a software algorithm to:
• select a plurality of decision variables;
• obtain real-time data, via the sensor, related to the plurality of decision variables;
• detect patterns from the plurality of decision variables indicating impending events;
• rank the plurality of decision variables based on a contribution of each variable to the impending events;
• determine which of the plurality of decision variables are operating outside a stable operating range;
• suggest remedial action to prevent impending events;
• and
• identify a training dataset to build a Robust Principal Component Analysis (Robust PCA) model, the training dataset derived from a stable time period free from impending events where the plurality of decision variables operate at a stable value.

A system for predicting and preventing electrical submersible pump (ESP) trip and failure events, the system comprising at least one processor used to control a software algorithm to:

• collect data, in real-time, from a well and an ESP positioned in a well via a plurality of sensors;
• manipulate historical data to determine and evaluate a plurality of decision variables;
• use mathematical models to derive trends and patterns based on a relationship between the plurality of decision variables and behavior of the ESP;
• associate the trends and patterns with trip and failure events;
• diagnose causes of the trip and failure events;
• prescribe remedial action to prevent the trip and failure events;
• and
• receive a training dataset to build a Robust Principal Component Analysis (Robust PCA) model, the training dataset derived from a stable time period free from impending events where the plurality of decision variables operate at a stable value.

A method for predicting and preventing electrical submersible pump (ESP) trip and failure events, the method comprising:

• collecting real-time data from a well and an ESP positioned in a well via a sensor;
• manipulating historical data to determine and evaluate a plurality of decision variables;
• using mathematical models to derive trends and patterns based on a relationship between the plurality of decision variables and behavior of the ESP;
• associating the trends and patterns with trip and failure events;
• using a Robust Principal Component Analysis (Robust PCA) methodology to model the plurality of decision variables to identify the trends and patterns in the real-time data;
• diagnosing causes of the trip and failure events;
• and
• prescribing remedial action to prevent the trip and failure events.