Using machine learning to predict fleet moves in hydraulic fracturing operations

A move orchestration system comprising:

• one or more processors;
• and
• memory operatively coupled to the one or more processors and storing a neural network, wherein the memory comprises instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
• presenting a first user interface configured to receive, at the move orchestration system and from devices of crew members, activity data associated with one or more events of a multi-event frac fleet move play prescribed for a fleet move for moving a frac fleet from a first well site to a second well site, wherein the multi-event frac fleet move play includes a plurality of events including the one or more events, the plurality of events including at least a rig down event, a rig transit event, and a rig up event;
• training the neural network using training data of a plurality of historical fleet moves, the training data including:
• (i) for each of the plurality of historical fleet moves, a label indicating whether the historical fleet move was completed within a corresponding allotted timeframe or failed to complete within the corresponding allotted timeframe, and (ii) for each of a plurality of historical events of each of the plurality of historical fleet moves, an allotted time duration and an actual time duration taken for completing the historical event;
• inputting features of the activity data and the prescribed multi-event frac fleet move play into the trained neural network, the activity data including an actual time duration taken for completing each of the one or more events and the prescribed multi-event frac fleet move play including an allotted time duration for each of the plurality of events;
• receiving, as output from the neural network, a probability of the fleet move having the prescribed multi-event frac fleet move play failing to complete within a predetermined allotted timeframe;
• in response to the probability being higher than a threshold, generating one or more recommendations configured to improve a likelihood of the fleet move being completed within the predetermined allotted timeframe or reduce overtime, the recommendations including adjustments to at least one of (a) the allotted time duration of one or more remaining events of the multi-event frac fleet move play and (b) an order in which the remaining events are to be performed;
• updating the prescribed multi-event frac fleet move play based on at least one of the recommendations by modifying the allotted time duration or an ordering of at least one remaining event of the multi-event frac fleet move play to produce an updated move play;
• presenting a second user interface configured to transmit, from the move orchestration system, (i) the one or more recommendations to a device of a crew member and (ii) the updated move play together with updated activity instructions to devices of crew members assigned to the at least one remaining event;
• and
• controlling, based on the updated move play and by the move orchestration system, an operation of a component of one or more of a power generation transport, a fracturing pump transport, and a blender transport of the frac fleet.

A non-transitory computer-readable storage medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:

• presenting a first user interface configured to receive, at a move orchestration system and from devices of crew members, activity data associated with one or more events of a multi-event frac fleet move play prescribed for a fleet move for moving a frac fleet from a first well site to a second well site, wherein the multi-event frac fleet move play includes a plurality of events including the one or more events, the plurality of events including at least a rig down event, a rig transit event, and a rig up event;
• training a neural network of the move orchestration system using training data of a plurality of historical fleet moves, the training data including:
• (i) for each of the plurality of historical fleet moves, a label indicating whether the historical fleet move was completed within a corresponding allotted timeframe or failed to complete within the corresponding allotted timeframe, and (ii) for each of a plurality of historical events of each of the plurality of historical fleet moves, an allotted time duration and an actual time duration taken for completing the historical event;
• inputting features of the activity data and the prescribed multi-event frac fleet move play into the trained neural network, the activity data including an actual time duration taken for completing each of the one or more events and the prescribed multi-event frac fleet move play including an allotted time duration for each of the plurality of events;
• receiving, as output from the neural network, a probability of the fleet move having the prescribed multi-event frac fleet move play failing to complete within a predetermined allotted timeframe;
• in response to the probability being higher than a threshold, generating one or more recommendations configured to improve a likelihood of the fleet move being completed within the predetermined allotted timeframe or reduce overtime, the recommendations including adjustments to at least one of (a) the allotted time duration of one or more remaining events of the multi-event frac fleet move play and (b) an order in which the remaining events are to be performed;
• updating the prescribed multi-event frac fleet move play based on at least one of the recommendations by modifying the allotted time duration or an ordering of at least one remaining event of the multi-event frac fleet move play to produce an updated move play;
• presenting a second user interface configured to transmit, from the move orchestration system, (i) the one or more recommendations to a device of a crew member and (ii) the updated move play together with updated activity instructions to devices of crew members assigned to the at least one remaining event;
• and
• controlling, based on the updated move play and by the move orchestration system, an operation of a component of one or more of a power generation transport, a fracturing pump transport, and a blender transport of the frac fleet.