CN-121977677-A - Self-calibration and drift compensation method for automatic underground water monitoring station

CN121977677ACN 121977677 ACN121977677 ACN 121977677ACN-121977677-A

Abstract

The invention discloses a self-calibration and drift compensation method of an automatic underground water monitoring station, which comprises the following steps of 1, sending a reference measurement starting instruction to an underground bubble method device by an edge computing gateway, writing reference data pairs formed by a bubble method water level reference value and a reference acquisition time into a reference data buffer zone, 2, creating a state storage structure and a covariance storage structure by the edge computing gateway, and 3, reading a drift amount estimated value from the state storage structure by the edge computing gateway to serve as a drift compensation amount, and triggering bubble method hardware reference acquisition when a drift estimated standard deviation exceeds a set threshold value. The invention solves the problems of high manual calibration cost and lack of independent reference support in pure software estimation in the prior art, and enables the monitoring station to stably operate for a long time and continuously output high-precision ground water level monitoring data under the unattended condition.

Inventors

LUO MEI
XU YUYANG
ZHANG JING
ZHANG YONGWEI
WANG XIAOWEI
LI YONGCHAO
LIANG HAO
WANG XINBING
LIU WENJING

Assignees

山东省国土空间生态修复中心(山东省地质灾害防治技术指导中心、山东省土地储备中心)

Dates

Publication Date: 20260505
Application Date: 20260203

Claims (10)

1. The self-calibration and drift compensation method of the automatic underground water monitoring station is characterized by comprising the following steps of: The method comprises the steps that 1, an edge computing gateway sends a reference measurement starting instruction to an underground bubble method device, compressed gas is injected into a gas guide pipe after a precise gas pump of the underground bubble method device is started, a wellhead gas pressure sensor collects instantaneous gas pressure readings in the gas guide pipe and sends the instantaneous gas pressure readings to the edge computing gateway, the edge computing gateway carries out balance state judgment on the instantaneous gas pressure readings, after judging that gas bubbles escape to reach a balance state, a gas bubble method water level reference value is obtained through calculation according to a balance gas pressure value and an ambient atmosphere gas pressure value, and reference data pairs are formed by the gas bubble method water level reference value and a reference acquisition time and written into a reference data buffer area; the edge computing gateway establishes a state storage structure and a covariance storage structure, wherein the state storage structure comprises a storage unit for storing a real water level estimated value, a drift amount estimated value and a drift change rate estimated value; And 3, the edge calculation gateway reads the estimated drift amount value from the state storage structure as a drift compensation amount, subtracts the drift compensation amount from the converted water level value output by the underground pressure sensor to obtain a compensated water level value and outputs the compensated water level value, calculates a drift estimation standard deviation according to the covariance storage structure, and triggers the acquisition of a bubble method hardware reference when the drift estimation standard deviation exceeds a set threshold value.
2. The method according to claim 1, wherein in step 1, the balance state determining process includes the steps that the edge computing gateway sets a temporary air pressure reading queue, writes a new instantaneous air pressure reading into the tail of the temporary air pressure reading queue and removes the earliest reading of the head of the queue each time the new instantaneous air pressure reading is received, the edge computing gateway reads all air pressure readings in the temporary air pressure reading queue, calculates the difference between the maximum value and the minimum value as the air pressure fluctuation amplitude, and determines that air bubbles escape to reach the balance state when the air pressure fluctuation amplitude is smaller than a preset pressure stability threshold value.
3. The method according to claim 2, wherein in the step 1, the process of calculating the bubble method water level reference value includes that an edge calculation gateway reads an arithmetic average value of all air pressure readings in an air pressure reading temporary storage queue as an equilibrium air pressure value, synchronously reads a current output value of an wellhead air pressure sensor as an environment air pressure value, subtracts the environment air pressure value from the equilibrium air pressure value to obtain a water column pressure value, and divides the water column pressure value by the weight of water to obtain the bubble method water level reference value.
4. The method of claim 1, wherein in step 2, the state storage structure comprises 3 storage units, the 1 st storage unit is used for storing a real water level estimated value and initializing to a first converted water level reading of the downhole pressure sensor, the 2 nd storage unit is used for storing a drift amount estimated value and initializing to 0, the 3rd storage unit is used for storing a drift change rate estimated value and initializing to 0, the covariance storage structure is a numerical matrix of 3 rows and 3 columns, diagonal elements are initialized to a preset initial variance value, and non-diagonal elements are initialized to 0.
5. The method of claim 4 wherein in step 2, the process of state prediction updating includes the edge computation gateway reading the current value of the 1 st memory location in the state storage structure and denoted as W, reading the current value of the 2 nd memory location and denoted as D, reading the current value of the 3 rd memory location and denoted as R, the edge computation gateway writing the value of W back to the 1 st memory location, writing the result of D plus R times the number of seconds of the sampling period to the 2 nd memory location, writing the value of R back to the 3 rd memory location, and the edge computation gateway adding the corresponding process noise increment to each of the 3 elements on the diagonal in the covariance storage structure.
6. The method according to claim 5, wherein in the step 2, the pressure observation correction process includes the steps that an edge calculation gateway reads a converted water level value currently output by the downhole pressure sensor and marks the converted water level value as P, a predicted observed value is obtained by adding the value of a1 st storage unit and the value of a 2 nd storage unit from a state storage structure, the predicted observed value is subtracted from the P to obtain an observation residual, the edge calculation gateway reads a1 st row and 1 st column element value from a covariance storage structure and marks the 1 st row and 2 nd column element value as C11, reads a 2 nd row and 2 nd column element value as C12, reads a result of C11 plus C12 plus C22 plus an observation noise variance as a residual variance, the edge calculation gateway calculates correction coefficients corresponding to each state component according to the element value and the residual variance of the covariance storage structure, corrects and updates the values of each storage unit in the state storage structure by using products of the correction coefficients and the observation residual, and synchronously updates the covariance storage structure.
7. The method according to claim 6, wherein in the step 2, the bubble method reference fusion correction process includes the steps that an edge calculation gateway checks whether an unprocessed reference data pair exists in a reference data buffer area, reads a bubble method water level reference value in the reference data pair and marks B when the unprocessed reference data pair exists, and reads corresponding reference acquisition time, the edge calculation gateway calculates a time difference between the current time and the reference acquisition time and marks T seconds, performs reverse backtracking on a state storage structure to obtain a backtracking drift amount and a backtracking water level value, the edge calculation gateway subtracts the B from the backtracking water level value to obtain a reference observation residual, calculates a reference correction coefficient according to a 1 st column element value of a covariance storage structure and bubble method observation noise variance, corrects and updates a value of a 1 st storage unit in the state storage structure by using a product of the reference correction coefficient and the reference observation residual, and marks the reference data pair as processed.
8. The method of claim 7, wherein the reverse traceback is performed on the state storage structure by subtracting the value of the 3 rd storage unit from the value of the 2 nd storage unit to multiply the value of the 3 rd storage unit by T to obtain a traceback drift amount, and maintaining the value of the 1 st storage unit unchanged as a traceback water level value.
9. The method according to claim 6, wherein in the step 2, the drift rate adaptive adjustment process includes the steps that an edge computing gateway maintains a residual record queue, writes current observation residual into the residual record queue, counts the number of residual values which are positive and recorded as positive residual counts when the residual record queue is full of a preset number of residual values, counts the number of residual values which are negative and recorded as negative residual counts, determines that a unidirectional drift trend exists when the positive residual count or the negative residual count exceeds a unidirectional drift determination threshold, multiplies the element value of the 3 rd row and the 3 rd column in a covariance storage structure by a preset amplification factor, and writes back when the positive residual count and the negative residual count are both in a stable determination range, determines that drift tends to be stable, and multiplies the element value of the 3 rd row and the 3 rd column in the covariance storage structure by a preset reduction factor, and writes back.
10. The method according to claim 1, wherein the step 3 further comprises the steps that the edge calculation gateway triggers the bubble method hardware reference acquisition according to the regular reference measurement period, the edge calculation gateway counts the number of times of the unplanned reference acquisition triggered in a preset time period, shortens the regular reference measurement period when the number of times of the unplanned reference acquisition exceeds a frequent trigger threshold, and delays the regular reference measurement period when the number of times of the unplanned reference acquisition is lower than a sparse trigger threshold.

Description

Self-calibration and drift compensation method for automatic underground water monitoring station Technical Field The invention relates to the technical field of data processing and analysis, in particular to the technical field of intelligent sensing systems, and specifically relates to a self-calibration and drift compensation method of an automatic underground water monitoring station. Background At present, a drop-in pressure sensor is generally adopted as a core measuring element for an automatic underground water monitoring station. The throw-in pressure sensor is directly immersed in water in the well, and the water level depth is converted by measuring the static pressure of the water column on the sensitive membrane. The sensor has the advantages of simple structure, convenient installation, large measuring range, high response speed and the like, and becomes a mainstream technical scheme for automatically monitoring the underground water level. However, drop-in pressure sensors face the common technical problem of sensor drift during long-term operation. Sensor drift refers to the phenomenon that the output value of a sensor gradually deviates from a real measured value along with time, and the reasons for the sensor drift include creep and fatigue of a sensitive diaphragm, ageing of an electronic element, thermal stress accumulation caused by temperature circulation, corrosion of chemical components of well water to sensor packaging materials, structural micro-deformation under the long-term action of water pressure and the like. The rate of accumulation of drift varies with sensor quality, downhole environment and operating conditions, typically up to several centimeters per year or even higher, which if uncorrected would severely impact the accuracy and availability of the monitored data. Aiming at the sensor drift problem, the prior art mainly adopts a periodic manual calibration mode for processing. And (3) a technician goes to a monitoring station according to a fixed period, uses tools such as a portable depth finder or a steel ruler and the like to independently measure the water level in the well, compares the measurement result with the reading of the pressure sensor, and adjusts the zero point and the range parameter of the sensor according to the deviation value. Such manual calibration methods, although capable of correcting the sensor bias to a small extent at the time of calibration, have significant limitations. First, manual calibration requires a professional to carry the equipment to the site, and for monitoring sites distributed in remote areas, inconvenient traffic or numerous, the calibration operation consumes a lot of manpower, material resources and time costs. Second, manual calibration can only be performed at discrete points in time, with sensor drift continuously accumulating between calibrations, resulting in a systematic deviation of the monitored data over the period of time that is difficult to quantify. Again, the calibration period is usually fixed in advance, and cannot be dynamically adjusted according to the actual drift condition of the sensor, and the situations that the calibration is too frequent to cause resource waste or the calibration interval is too long to cause data quality degradation may occur. Disclosure of Invention The invention aims to provide a self-calibration and drift compensation method for an automatic underground water monitoring station, which is characterized in that a water level reference value which is not influenced by sensor drift is obtained through a bubble method device, a state space model which comprises a real water level estimated value, a drift amount estimated value and a drift change rate estimated value is established, state prediction updating, pressure observation correction, bubble method reference fusion correction and drift rate self-adaptive adjustment are sequentially carried out in each pressure sensor sampling period, continuous tracking and real-time compensation of sensor drift are realized, the acquisition of the bubble method reference is automatically triggered according to the uncertainty degree of drift estimation, a soft-hard cooperative closed-loop self-calibration mechanism is formed, the problems that the manual calibration cost is high and the pure software estimation lacks independent reference support in the prior art are solved, and the monitoring station can stably operate for a long time and continuously output high-precision underground water level monitoring data under the unattended condition. In order to solve the technical problems, the invention provides a self-calibration and drift compensation method of an automatic underground water monitoring station, which comprises the following steps: The method comprises the steps that 1, an edge computing gateway sends a reference measurement starting instruction to an underground bubble method device, compressed gas is injected into a gas guide