System and method for predictive volumetric and structural evaluation of storage tanks

A system for predictive volumetric estimation of storage tanks, the system comprising:

• a calibration device configured to measure calibration data relating to volumetric calibration of each tank, wherein the calibration data includes a dimension of the tank walls and wherein the calibration device is selected from the group consisting of:
• a 3D laser range finder mounted to a robotic vehicle and configured to measure the dimension of the tank walls by measuring deviations of the tank walls in an inward or outward direction relative to an interior of the tank, and
• an acoustic sensor configured to acoustically measure the dimension of the tank walls;
• and
• a computing device comprising a processor in operative communication with a communication interface, an input/output interface and a non-transitory computer readable storage medium having instructions stored therein that that are executable by the processor, wherein the instructions configure the processor to:
• receive sets of historical data concerning respective tanks among the tanks, wherein the historical data sets include multiple sets of calibration data previously measured using the calibration device for each of the respective tanks and further include values for a plurality of parameters for each of the respective tanks,
• generate a predictive machine learning (ML) model for estimating volumetric changes in a tank, wherein the volumetric changes in the tank represent a change in the volume of the tank, and wherein the predictive ML model is generated by the processor training a recurrent neural network algorithm on the historical data sets to correlate data points in the historical data sets with parameters representing a change in volume of the tank including one or more of, tank volume, tank structure, tank roundness, tank tilt, and expansion rate,
• receive and store, by the processor, current operational data concerning a given tank among the tanks,
• apply, by the processor, the predictive ML model to one or more historical data sets for the given tank among the tanks to estimate one or more current parameters for the given tank relating to a volumetric change in the tank,
• wherein the processor is further configured to estimate the one or more current parameters for the given tank by applying the predictive ML model to the current operational data along with relevant historical data from among the sets of historical data,
• wherein the one or more current parameters are selected from the group of parameters including, tank volume, tank structure, tank roundness, tank tilt, and expansion rate, and
• wherein prior to applying the predictive ML model, the processor is configured to select the predictive ML model among a plurality of trained predictive ML models based on a type of the given tank, and
• generate, based on the estimated one or more current parameters for the given tank, an output representing the estimated one or more current parameters for the given tank,
• an implementation tool, wherein the implementation tool is in operative communication with the processor and configured to perform one or more operations as a function of an instruction received from the processor, and wherein the processor is configured to generate, based on the estimated one or more current parameters for the given tank, the instruction to perform the one or more operations for the given tank for output to the implementation tool,
• wherein the implementation tool includes the calibration device, and wherein the instruction causes the calibration device to capture current calibration data in the form of one or more volume measurements for the given tank,
• wherein the processor is configured to receive the current calibration data from the calibration device and, based on the current calibration data and the predictive ML model, determine an updated estimate for the one or more current parameters and generate a revised instruction for the calibration device.