Pumping efficiency apparatus and method

A computer system for creating a neural network model for real-world mechanical system control, the computer system comprising:

• a processor;
• and
• a memory with computer code instructions stored thereon, the processor and the memory, with the computer code instructions, being configured to cause the system to:
• extract (i) an indication of efficiency and (ii) values of operational characteristics of a plurality of devices from a plurality of characteristic curves, wherein each characteristic curve corresponds to a respective device of the plurality of devices, in the mechanical system, functioning at a given speed, the plurality of devices including a plurality of device types;
• create a training dataset by determining efficiency and values of the operational characteristics for the mechanical system functioning with multiple combinations of the plurality of devices operating at each speed from a set of a plurality of speeds using the extracted indication of efficiency, the extracted values of the operational characteristics, and one or more physics-based laws that indicate changes to the operational characteristics as a function of speed of the plurality of devices, wherein determining the values of the operational characteristics using the one or more physics based laws comprises:
• (a) calculating coefficients for a plurality of polynomial equations by fitting the extracted values of the operational characteristics to the plurality of polynomial equations, wherein (i) the polynomial equations are indicated by the one or more physics based laws and (ii) each of the plurality of polynomial equations indicates values of the operational characteristics for the mechanical system as a function of speed of the plurality of devices and (b) using the calculated coefficients in at least a subset of the plurality of polynomial equations to determine the values of the operational characteristics for the mechanical system functioning with multiple combinations of the plurality of devices operating at each speed from the set of the plurality of speeds;
• train a neural network model with the created training dataset, wherein the training configures the neural network model to predict efficiency of the mechanical system based on operating data;
• and
• deploy the trained neural network model to control operation of the mechanical system by:
• processing real-time operating data for the mechanical system with the trained neural network model to predict efficiency of the mechanical system operating in accordance with each of a plurality of operating scenarios;
• based on the predicted efficiency of each of the plurality of operating scenarios, selecting a given operating scenario of the plurality of operating scenarios, the selected operating scenario including a number of devices of the plurality of devices and an indication of which device types to operate of the plurality of device types;
• and
• controlling the mechanical system in accordance with the selected operating scenario.

A computer program product for creating a neural network model for real-world mechanical system control, the computer program product comprising:

• one or more non-transitory computer-readable storage devices and program instructions stored on at least one of the one or more storage devices, the program instructions, when loaded and executed by a processor, cause an apparatus associated with the processor to:
• extract (i) an indication of efficiency and (ii) values of operational characteristics of a plurality of devices from a plurality of characteristic curves, wherein each characteristic curve corresponds to a respective device of the plurality of devices, in the mechanical system, functioning at a given speed, the plurality of devices including a plurality of device types;
• create a training dataset by determining efficiency and values of the operational characteristics for the mechanical system functioning with multiple combinations of the plurality of devices operating at each speed from a set of a plurality of speeds using the extracted indication of efficiency, the extracted values of the operational characteristics, and one or more physics-based laws that indicate changes to the operational characteristics as a function of speed of the plurality of devices, wherein determining the values of the operational characteristics using the one or more physics based laws comprises:
• (a) calculating coefficients for a plurality of polynomial equations by fitting the extracted values of the operational characteristics to the plurality of polynomial equations, wherein (i) the polynomial equations are indicated by the one or more physics based laws and (ii) each of the plurality of polynomial equations indicates values of the operational characteristics for the mechanical system as a function of speed of the plurality of devices and (b) using the calculated coefficients in at least a subset of the plurality of polynomial equations to determine the values of the operational characteristics for the mechanical system functioning with multiple combinations of the plurality of devices operating at each speed from the set of the plurality of speeds;
• train a neural network model with the created training dataset, wherein the training configures the neural network model to predict efficiency of the mechanical system based on operating data;
• and
• deploy the trained neural network model to control operation of the mechanical system by:
• processing real-time operating data for the mechanical system with the trained neural network model to predict efficiency of the mechanical system operating in accordance with each of a plurality of operating scenarios;
• based on the predicted efficiency of each of the plurality of operating scenarios, selecting a given operating scenario of the plurality of operating scenarios, the selected operating scenario including a number of devices of the plurality of devices and an indication of which device types to operate of the plurality of device types;
• and
• controlling the mechanical system in accordance with the selected operating scenario.

A method of controlling a mechanical system, the method comprising:

• obtaining real-time operating data for the mechanical system;
• processing the obtained real-time operating data with a trained neural network model to predict efficiency of the mechanical system operating in accordance with each of a plurality of operating scenarios, wherein the neural network model is:
• (i) configured to predict efficiency of the mechanical system based on operating data and (ii) trained using a dataset created by determining efficiency and values of operational characteristics for the mechanical system functioning with multiple combinations of a plurality of devices operating at each speed from a set of a plurality of speeds using an indication of efficiency and values of the operational characteristics extracted from a plurality of characteristic curves of the plurality of devices, and one or more physics-based laws that indicate changes to the operational characteristics as a function of speed of the plurality of devices, wherein determining the values of the operational characteristics using the one or more physics based laws comprises:
• (a) calculating coefficients for a plurality of polynomial equations by fitting the extracted values of the operational characteristics to the plurality of polynomial equations, wherein (i) the polynomial equations are indicated by the one or more physics based laws and (ii) each of the plurality of polynomial equations indicates values of the operational characteristics for the mechanical system as a function of speed of the plurality of devices and (b) using the calculated coefficients in at least a subset of the plurality of polynomial equations to determine the values of the operational characteristics for the mechanical system functioning with multiple combinations of the plurality of devices operating at each speed from the set of the plurality of speeds and where each characteristic curve corresponds to a respective device of the plurality of devices functioning at a given speed, the plurality of devices including a plurality of device types;
• selecting a given operating scenario of the plurality based on the predicted efficiency of each of the plurality of operating scenarios, the selected operating scenario including a number of devices of the plurality of devices and an indication of which device types to operate of the plurality of device types;
• and
• controlling the mechanical system in accordance with the selected operating scenario.