Method and system of risk prevention based on energy of natural gas in a full cycle

Patent metadata

Field	Value
Publication number	US-12117328-B2
Application number	US-202217649336-A
Country	US
Kind code	B2
Filing date	Jan 28, 2022
Priority date	Feb 4, 2021
Publication date	Oct 15, 2024
Grant date	Oct 15, 2024

Abstract

The present disclosure discloses a risk prevention method based on energy of natural gas in a full cycle, including: obtaining loss data of natural gas based on metering equipment, wherein the loss data of the natural gas includes metering data of the natural gas consumed in a measured area during a plurality of time periods; obtaining output data of natural gas, wherein the output data of the natural gas includes metering data of the natural gas transmitted by a gas supplier in the measured area during the plurality of time periods; based on the processing of loss data of natural gas and output data of natural gas, determining whether natural gas transmission is abnormal.

First claim

What is claimed is:

1. A method for risk prevention based on energy of natural gas in a full cycle, implemented on a computing device comprising at least one storage device and at least one processor, wherein the method is executed by the at least one processor, the method comprising:
obtaining loss data of the natural gas based on metering equipment, wherein the loss data of the natural gas includes metering data of the natural gas consumed in a measured area, wherein the metering of the consumed natural gas is performed during a plurality of time periods;
wherein an interval between the plurality of time periods is determined by processing a time type of a current time and second use data of the natural gas in the measured area using a second model; wherein the second model is a combined model of a graph neural network (GNN) model and a deep neural network (DNN) model; wherein
an input of the GNN model includes a length value of public pipe and a natural gas use feature; and an output of the GNN model is a transmission feature vector of the natural gas for each natural gas user; wherein each natural gas user is taken as a node of a map, and a public pipe connected between natural gas users is taken as an edge of the map, node feature is the natural gas use feature corresponding to the natural gas user, and edge feature is the length value of the public pipe connected between two natural gas users;
the DNN model is configured to determine a time interval; an input of the DNN model is the output of the GNN model and the time type of the current time; and an output of the DNN model is the time interval between a next data acquisition time and the current time;
the GNN model and the DNN model are obtained through joint training based on training samples and labels; wherein the training samples include a map of transmission feature of the natural gas and a corresponding historical time type; nodes included in the map of transmission feature of the natural gas are users who use the natural gas in the historical time, and a node feature represents natural gas use feature in the corresponding historical time; an edge is a connecting line between two users, and an edge feature represents a length value of the public pipe connected between two users; and the labels are time interval between multiple historical time periods; and
the joint training includes: inputting the map of transmission feature of the natural gas into an initial GNN model of the second model; inputting the time type of the current time and an output of the initial GNN model into an initial DNN model of the second model; constructing a loss function based on an output of the initial GNN mode and the labels; updating parameters of the initial GNN model and the initial DNN model iteratively based on the loss function until preset conditions are met; and obtaining the GNN model and the DNN model, wherein the preset conditions include a convergence of the loss function and a number of iterations reaching a threshold;

obtaining output data of the natural gas, wherein the output data of the natural gas includes metering data of the natural gas transmitted by a gas supplier in the measured area, wherein the metering of the transmitted natural gas is performed during the plurality of time periods; and
determining whether natural gas transmission abnormality occurs by processing the loss data of the natural gas and the output data of the natural gas;
in response to determining that the natural gas transmission abnormality has occurred;
obtaining measured data of each transmission equipment of the natural gas; wherein the measured data include pressure value detected by a pressure sensor and flow value detected by a flow sensor; and
determining location of the abnormality by processing the measured data using a third model; wherein the third model is a GNN model, wherein an input of the third model includes a graph composed of multiple candidate points, and an output of the third model is a leakage probability at each candidate point; wherein node of the graph is the candidate points, edge of the graph is a connecting line between the two candidate points, node feature is measured data corresponding to the candidate point, edge feature is a distance between the two candidate points; wherein
the third model is obtained by training based on third training samples and third labels, wherein the third training samples include a sample graph composed of multiple sample candidate points, in which node of the sample graph is the sample candidate points, edge of the sample graph is a connecting line between the two sample candidate points, node feature is historically measured data corresponding to the sample candidate point, edge feature is a distance between the two sample candidate points, the third labels represent whether there is gas leakage at the sample candidate point corresponding to the historically measured data;
the training includes: inputting the third training samples into an initial third model; constructing a loss function based on the third labels and an output of the initial third model; updating parameters of the initial third model iteratively based on the loss function; and obtaining the third model when the loss function meets a preset condition, wherein the preset condition include a convergence of the loss function and a number of iterations reaching a threshold; and
a candidate point with a largest gas leakage probability among all candidate points or candidate points with the gas leakage probability exceeding a threshold is determined as the location of the abnormality;

sending the location of the abnormality to a transmission station to make the transmission station adjust natural gas transmission parameters in response to a result of abnormal natural gas transmission.

2. The method of claim 1 , wherein the determining whether natural gas transmission abnormality occurs by processing the loss data of the natural gas and the output data of the natural gas includes:
determining whether a difference between the output data of the natural gas and the loss data of the natural gas is less than a preset threshold; and
if the difference between the output data of the natural gas and the loss data of the natural gas is less than the preset threshold, determining that the natural gas transmission abnormality has not occurred; or
if the difference between the output data of the natural gas and the loss data of the natural gas is not less than the preset threshold, determining that natural gas transmission abnormality has occurred.

3. The method of claim 2 , wherein a unit of the metering data includes an energy unit and/or a volume unit.

4. The method of claim 3 , wherein the determining whether natural gas transmission abnormality occurs by processing the loss data of the natural gas and the output data of the natural gas includes:
calculating an energy difference between the output data of the natural gas and the loss data of the natural gas and judging whether the energy difference is less than a first threshold,
if the energy difference is less than the first threshold, calculating a volume difference between the output data of the natural gas and the loss data of the natural gas, or
if the energy difference is not less than a first threshold, determining that the natural gas transmission abnormality has occurred; and
judging whether the volume difference is less than a second threshold,
if the volume difference is less than the second threshold, determining that the natural gas transmission abnormality has not occurred, or
if the volume difference is not less than the second threshold, determining that the natural gas transmission abnormality has occurred.

5. The method of claim 1 , wherein the determining whether natural gas transmis

Assignees

• Chengdu Qinchuan Iot Tech Co Ltd

Inventors

• Shao Zehua
• Xiang Haitang
• Liu Bin

Classifications

• G01F25/15

  specially adapted for gas meters (G01F25/11 - G01F25/14, G01F25/17 take precedence) · CPC title

• G01N27/4163

  checking the operation of, or calibrating, the measuring apparatus (G01N27/3274, G01N27/4175 and G01N33/0006 take precedence) · CPC title

• G01F23/16 Primary

  Indicating, recording, or alarm devices being actuated by mechanical or fluid means, e.g. using gas, mercury, or a diaphragm as transmitting element, or by a column of liquid · CPC title

• G01F15/063 Primary

  using electrical means · CPC title