Water conservancy data acquisition and transmission system based on narrowband internet of things

Abstract

The application discloses a water conservancy data acquisition and transmission system based on narrowband Internet of Things, and relates to the technical field of narrowband Internet of Things; the system comprises a data acquisition and event construction module, a non-radio frequency contrast construction module, a pressure recovery module and a sending control and calibration module; the system performs timing correction on pressure measurement sequences and communication module sending operation data, constructs real transmission events; controllable equivalent load branches of common power supply paths and grounding paths are controlled to reproduce incremental current trajectory, form non-radio frequency load events, separate power supply coupling constraints and radio frequency associated coupling constraints, establish a bounded coupling model, the system recovers a real pressure feasible interval based on the model, controls alarm sending according to the position relationship between the interval and an event reporting threshold, and remote receiving equipment returns a model to distinguish a target, and a terminal performs non-radio frequency calibration and updates the bounded coupling model.

Description

Technical Field

[0001] This invention relates to the field of narrowband Internet of Things (IoT) technology, specifically to a water conservancy data acquisition and transmission system based on narrowband IoT. Background Technology

[0002] In water conservancy projects, telemetry terminals are typically installed at seepage pressure monitoring wells of dams, water level monitoring points upstream of sluice gates, or pressure measurement points of water conveyance pipelines. Pressure sensors convert the water pressure at the monitored location into an electrical signal, which is then conditioned and converted from analog to digital to form a pressure measurement value. This value is then transmitted to a remote receiving device via a narrowband IoT wireless link. Taking dam seepage pressure monitoring as an example, telemetry terminals are often installed on dam slopes, dam toes, or underground corridors and rely on energy storage units for power supply. Well walls, metal well covers, humid environments, and terrain obstructions can cause wireless coverage to change with rainfall, seasons, and the condition of on-site facilities. Terminals typically collect pressure measurement values ​​according to a preset cycle and establish a wireless connection when the measurement value reaches the event reporting condition, sending the current measurement value and its sampling time to the monitoring center. To maintain the long-term operation of on-site terminals, existing telemetry systems generally adopt time-division control of acquisition and transmission, power supply stabilization, local buffering, and repeated transmission in weak coverage, so that measurement data can still be transmitted remotely even when wireless coverage fluctuates.

[0003] When wireless coverage weakens, the time required for narrowband IoT communication modules to establish connections and complete uplink transmissions may increase, potentially leading to multiple accesses or repeated transmissions. Transient load changes generated during this process may be transmitted to the pressure measurement channel via shared power and grounding paths. RF transmission signals may also couple to the analog front-end via internal wiring and installation structure, causing a delayed recovery shift in pressure measurements during and after transmission. For example, when the actual seepage pressure of a dam is slowly changing, a weak coverage transmission process may cause the pressure measurement to gradually rise and then slowly fall back after transmission. This shift is continuous in time and may meet the final... The event reporting conditions at the terminal cause the terminal to restart wireless transmission based on the measurement value affected by the transmission process. The new transmission process may prolong the recovery time of the pressure measurement channel. When the actual seepage pressure change and the recovery process after transmission overlap in time, it is difficult to determine the part formed by the actual water pressure and the part formed by the wireless transmission process based solely on the measurement value amplitude or conventional smoothing results. Therefore, how to use the state information provided by the wireless transmission process itself to identify the transmission coupling response in the pressure measurement value, prevent the response from triggering new wireless transmission, and enable the remote receiving device to obtain a verifiable real pressure range has become a technical problem that needs to be further solved in the wireless transmission of hydraulic measurement values. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides the following technical solution: A water conservancy data acquisition and transmission system based on narrowband Internet of Things includes: The data acquisition and event construction module acquires the pressure measurement sequence of the same pressure telemetry terminal and the transmission operation data of the narrowband IoT communication module, performs timing correction, forms the actual transmission load trajectory and radio frequency transmission trajectory based on the corrected transmission operation data, divides the transmission range and constructs real transmission events; and forms the pressure allowable change band based on the non-radio frequency measurement data before and after the transmission range, and performs calibration and admission determination for real transmission events. The no-RF comparison modeling module reads real transmission events that are valid for calibration and admission, controls the controllable equivalent load branch that shares the power supply path and grounding path with the narrowband IoT communication module to reproduce the incremental current trajectory, and forms no-RF load events; based on the no-RF load events, power supply coupling constraints are formed, and the power supply coupling response is subtracted from the total transmission correlation response of the real transmission events to form RF correlation coupling constraints and establish a bounded coupling model; The pressure recovery module uses a bounded coupling model to form a feasible transmission coupling interval based on the load trajectory and radio frequency transmission trajectory of the current real transmission event; it performs error expansion on the original pressure measurement value to form an observed pressure interval; it subtracts the observed pressure interval from the feasible transmission coupling interval to form a measurement-consistent pressure interval; and it intersects the measurement-consistent pressure interval with the physical pressure prediction interval formed based on the allowable pressure variation zone to form the real pressure feasible interval. The transmission control and calibration module determines the tasks to be transmitted from the identified over-limit transmission tasks and pending alarm tasks based on the positional relationship between the feasible range of the actual pressure and the event reporting threshold, and transmits them to the remote receiving device via the narrowband IoT communication module; it locks abnormal repeated transmissions triggered by the original pressure measurement value during the current actual transmission event processing; it receives the model-distinguished target returned by the remote receiving device for pending alarm tasks, and when the model distinguishes the target as pointing to the power supply coupling constraint, it selects the target sequence to perform radio frequency-free calibration, incorporates the calibration constraint into the power supply coupling constraint, and updates the bounded coupling model.

[0005] Furthermore, the transmitted operational data includes the operating current, supply voltage, and RF transmission status of the narrowband IoT communication module; The data acquisition and event construction module performs timing correction on the pressure measurement sequence and the transmitted operation data according to the fixed processing delay pre-calibrated in the pressure acquisition link and the transmission operation data acquisition link.

[0006] Furthermore, the data acquisition and event construction module forms an idle current coverage interval based on the operating current in the non-RF state, and sets the earlier of the starting time when the operating current continuously exceeds the upper limit of the idle current coverage interval and the time when the RF transmission state takes effect as the transmission start point. After the radio frequency transmission ends and the operating current and supply voltage return to their corresponding idle ranges, the moment when the fitting slope and fitting residual of the continuous pressure observation window meet the stability threshold is set as the end point of the transmission action.

[0007] Furthermore, the data acquisition and event construction module forms natural pressure constraints based on the non-radio frequency measurement data before and after the transmission range, constructs a candidate natural pressure sequence set, and forms a pressure allowable change band based on the candidate natural pressure boundary at each pressure sampling time. The maximum width of the allowable pressure variation band and the preset pressure measurement error width are divided by the continuous deviation of the original pressure measurement value to form the calibration ambiguity ratio. Real transmission events with a calibration ambiguity ratio not greater than the calibration access threshold are set as valid calibration access.

[0008] Furthermore, the no-RF comparison modeling module extracts incremental current trajectories from real transmission events that have been calibrated and approved. When the initial power supply state and pressure change state of the real transmission event and the no-RF load reproduction process to be executed meet the pairing tolerance, the controllable equivalent load branch is driven to reproduce the incremental current trajectory. When the reproduction error is no greater than the reproduction threshold, a no-RF load event is generated.

[0009] Furthermore, the non-RF control modeling module forms a power supply load regression row according to the power supply coupling memory time based on the power supply load characteristics of each pressure sampling window, and constrains the power supply regression coefficient with the pressure response interval of the non-RF load event to form a power supply coupling constraint. The power supply coupling response is subtracted from the total transmitted associated response to form the radio frequency associated response interval. The radio frequency transmission regression row is formed according to the radio frequency associated coupling memory time based on the radio frequency transmission characteristics. The radio frequency regression coefficient is constrained by the radio frequency associated response interval to form the radio frequency associated coupling constraint.

[0010] Furthermore, the no-RF control modeling module establishes a bounded coupling model when the ratio of the number of regression rows to the number of regression features reaches a preset threshold, and when the power supply regression coefficient and the RF regression coefficient have finite upper and lower bounds. An independent verification event group consisting of real transmission events not involved in modeling and their events without radio frequency load is used to verify the coverage relationship between the predicted response interval and the observed response interval. Based on the coverage relationship, the applicable state and feature scope of the bounded coupling model are determined.

[0011] Furthermore, the pressure recovery module calculates the power supply coupling response interval and the radio frequency correlation coupling response interval based on the power supply coupling constraint and the radio frequency correlation coupling constraint, and adds the intervals together to form the transmission coupling feasible interval. The measurement error interval is formed by the non-radio frequency measurement residuals before and after the transmission action interval and the pressure measurement error boundary. The original pressure measurement value is then extended using the measurement error interval to form the observation pressure interval.

[0012] Furthermore, the pressure recovery module uses the pressure allowable change zone as the physical pressure prediction interval for the first pressure sampling moment to be recovered. It forms a natural pressure increment range based on the adjacent pressure allowable change zones, adds the natural pressure increment range to the real pressure feasible interval of the previous pressure sampling moment, and then intersects it with the current pressure allowable change zone to form the current physical pressure prediction interval. The intersection of the current physical pressure prediction interval and the measured pressure interval is used to form the actual pressure feasible interval at the current pressure sampling time.

[0013] Furthermore, when the lower bound of the real pressure feasible range is continuously higher than the event reporting threshold and the confirmation time is reached, the sending control and calibration module generates a confirmed over-limit sending task. When the feasible range of real pressure crosses the event reporting threshold or no effective feasible range of real pressure is formed, an unresolved record is established; when the unresolved state reaches the preset maximum duration, or the model is inapplicable, or the upper limit of the unresolved range reaches the safety pressure threshold, or the data discontinuity state reaches the preset duration, an unresolved alarm task is generated. The remote receiving device compares the interval widths of the power supply coupling response interval and the radio frequency associated coupling response interval during the confirmation period. When the interval width of the power supply coupling response interval is not less than the interval width of the radio frequency associated coupling response interval, the model differentiation target is set as the power supply coupling constraint. The transmission control and calibration module distinguishes targets based on the model, selects target sequences, performs RF-free calibration, incorporates calibration constraints into power supply coupling constraints, and re-forms RF-related coupling constraints and bounded coupling models based on the updated power supply coupling constraints.

[0014] This invention provides a water conservancy data acquisition and transmission system based on narrowband Internet of Things, which has the following beneficial effects: 1. This invention takes real wireless transmission events as its object. When the communication module is in a non-transmitting state, it uses a controllable equivalent load branch that shares the power supply path and grounding path with the communication module to reproduce the transmission current trajectory, thus constructing a radio frequency load-free event paired with the real transmission event. In this way, the power supply load effect during the transmission process is independently extracted from the total response of the superposition of the power supply effect and the radio frequency correlation effect, so that the power supply coupling response of the pressure measurement channel has a repeatable and verifiable reference basis, avoiding the inference of the interference source based solely on the shape of the measurement curve.

[0015] 2. This invention further utilizes no-RF load events to form power supply coupling constraints and separates RF correlation coupling constraints from the total transmission correlation response of the real transmission event, establishing a coupling relationship with finite upper and lower bounds. For pressure offset during and after transmission, instead of directly outputting a single correction value, it incorporates the power supply coupling response, RF correlation response, pressure measurement error, and natural pressure change into interval calculations to form a feasible range of real pressure. This allows the remote receiving device to know the range of the current pressure measurement result that can be explained by real water pressure and to distinguish between over-limit states and undetermined states that cannot be determined yet.

[0016] 3. This invention uses the actual pressure feasible range as the basis for wireless transmission control. During the current transmission event and its delay recovery phase, it locks the original pressure measurement value to prevent direct triggering of new abnormal transmission tasks, thus cutting off the self-excited feedback chain of "transmission causing pressure deviation, deviation triggering transmission again, and then transmission extending the deviation again". For unresolved states that cross the event reporting threshold, the remote receiving device forms a model based on the uncertainty range of the power supply coupling response and the radio frequency correlation response to distinguish the target. The terminal then performs targeted non-radio frequency calibration and updates the coupling constraints, so that subsequent event processing can adapt to changes in the state of the energy storage unit, wireless coverage, and terminal installation conditions. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the transmission coupling structure of the water conservancy telemetry terminal of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 This embodiment provides a water conservancy data acquisition and transmission system based on narrowband Internet of Things, including: a data acquisition and event construction module, a radio frequency-free comparison modeling module, a pressure recovery module, and a transmission control and calibration module.

[0020] The following describes the process of establishing a model, conducting on-site operation, and updating the model of the J-03 seepage pressure monitoring well at the downstream toe of the earth-rock dam. This embodiment always uses the same pressure telemetry terminal, the same pressure acquisition channel, and the same narrowband IoT communication module as the processing object. Each transmission event is bound to the terminal identifier, pressure acquisition channel identifier, communication module identifier, and event sequence number. Data from different monitoring points, different terminals, or different acquisition channels are not merged. Unless otherwise specified, the pressure unit is kPa, the time unit is s, the current unit is mA, and the voltage unit is V.

[0021] The J-03 monitoring well is equipped with a pressure sensor with a range of 0 to 200 kPa. The pressure sensor is connected to the terminal controller via a signal conditioning circuit and an analog-to-digital conversion circuit. The terminal controller is connected to a narrowband IoT communication module, a working current acquisition circuit, a power supply voltage acquisition circuit, an RF envelope detection circuit, a local memory, a power meter for the energy storage unit, and a controllable equivalent load branch. The controllable equivalent load branch is connected to the communication module at the same common power supply node. Its return end is connected to the main grounding junction point shared by the communication module and the signal conditioning circuit, so that the equivalent load current passes through the common voltage regulator, the common power supply line, and the main common grounding impedance that the communication module actually transmits.

[0022] After installation, the pressure sensor was calibrated using a standard pressure source at 0, 50, 100, 150, and 200 kPa. The indication error at each calibration point was recorded in a segmented calibration table, and calibration was completed before the pressure data entered the event processing stage. The maximum residual indication after calibration was 0.05 kPa, and the fluctuation of repeated measurements within 30 minutes did not exceed 0.02 kPa. The minimum pressure change that the pressure sensor could stably distinguish was 0.02 kPa, and the pressure change after analog-to-digital conversion quantization step size was 0.005 kPa. Therefore, 0.02 kPa was determined as the pressure change resolution boundary to determine the distinguishable changes between adjacent pressure samples. 0.05 kPa was determined as the absolute pressure uncertainty boundary to expand the upper and lower limits of the original pressure measurement value, the measurement value without RF load, and the threshold determination related interval. The pressure change resolution boundary, the absolute pressure uncertainty boundary, and the segmented calibration table were all determined during the installation and commissioning phase and were not modified based on a single transmission event.

[0023] The terminal controller acquires the calibrated raw pressure measurement sequence through the pressure acquisition channel, the communication module operating current sequence through the current acquisition circuit, the power supply voltage sequence of the common power supply node through the voltage acquisition circuit, and the wireless access status and radio frequency transmission status through the communication module control interface and the radio frequency envelope detection circuit. The sampling time of each sequence is recorded using the same hardware timing reference.

[0024] During the installation and commissioning phase, a step signal triggered by the same control pulse is input to the pressure acquisition chain, operating current acquisition chain, power supply voltage acquisition chain, and RF status acquisition chain, respectively. This is repeated 10 times under 3.4V, 3.6V, and 3.8V power supply conditions. The arrival time is determined when the first moment when the output of each acquisition chain exceeds 50% of the step amplitude for three consecutive sampling points. The median value of each arrival time relative to the control pulse is taken as a fixed processing delay. When the difference between the maximum and minimum values ​​of the 10 results of the same acquisition chain exceeds one sampling cycle of that acquisition chain, recalibration is performed. During operation, only the calibrated fixed processing delay is called to correct the sampling time, so that the pressure, current, voltage, and RF status correspond to the same physical time axis.

[0025] The pressure sampling frequency can be 5 to 20 Hz, preferably 10 Hz; the operating current and power supply voltage sampling frequency can be 10 to 50 kHz, preferably 20 kHz. In this embodiment, the shortest load pulse that the communication module can observe is about 1 ms. When using a sampling frequency of 20 kHz, one shortest load pulse corresponds to 20 sampling points, which can preserve the load edge and the power supply voltage drop process.

[0026] The stabilization time of the pressure acquisition chain is determined to be 0.6s through step testing, the digital filter group delay is 0.15s, the upper limit of the communication module connection release time is 2.5s, and the recovery safety factor can be 1.5 to 3.0, preferably 2.0. The maximum recovery observation time is formed by multiplying the maximum value of the aforementioned three times by the recovery safety factor, and in this embodiment, it is 5.0s. The 5.0s maximum recovery observation time is calculated from the electrical recovery time and is used for selecting the pressure observation window after electrical recovery and determining the end point of the transmission action. The connection release and power restoration process between the end of RF transmission and the electrical recovery time are timed separately and are not included in the 5.0s maximum recovery observation time. The duration of the pressure observation window and the non-RF measurement segment can be 1.2 to 2.0s, preferably 1.5s. At a 10Hz pressure sampling frequency, each preferred window includes 15 pressure samples, simultaneously meeting the requirements of physical duration and no less than 12 effective pressure samples.

[0027] In this embodiment, the 90% minimum width coverage interval is formed by sorting the samples. It does not presuppose that the samples follow a specific probability distribution. The number of samples to be covered is the total number of samples multiplied by 90% and then rounded up. At least 14 out of 15 samples are covered. Uncovered samples are not directly deleted. If the continuous duration of a sample reaches the continuous response holding time, the corresponding measurement segment is not considered a stable measurement segment.

[0028] During the model establishment phase, the terminal accumulated 6 pairs of events at monitoring well J-03. Among them, E_01 to E_05 were used to determine the memory time, effective feature columns, feature scaling parameters, and bounded coupling coefficient set. E_06 was reserved as a pair of events for independent verification only after the model structure and parameters were completely fixed. The representative real transmission event was recorded as E_04, and its corresponding no-RF load event was recorded as Z_04. Model version V2.2 was put into operation after verification by E_06, and the subsequent cycle transmission formed the real transmission event E_07.

[0029] The data acquisition and event construction module generates real transmission event records and determines the reference calibration access status. The terminal identifies a periodic transmission or an abnormal first transmission as a candidate transmission event and binds the original pressure measurement sequence, communication module operating current sequence, power supply voltage sequence, wireless access status, and RF transmission status within the candidate transmission event to the corresponding event identifier. The RF transmission status is formed by the communication module sending an enable signal and the independent RF envelope detection result. The sending enable signal limits the control period during which transmission is allowed, and the RF envelope detection result confirms the actual RF burst. After fixed processing and delay compensation, the two form the RF transmission trajectory. When the RF transmission status can only be inferred from the operating current pulse, the source of the RF transmission trajectory is marked as a non-independent source. This event only calls the corresponding model when the RF transmission trajectory source saved in the existing model matches the source label. If no model version is matched, only a degraded pressure range is formed. Non-independent source events do not enter the RF correlation coupling model update chain.

[0030] During the no-radio period before the candidate transmission event occurs, the terminal collects the idle operating current of the communication module and forms an idle current coverage interval. The coverage ratio of the minimum width coverage interval can be 85% to 95%, preferably 90%. In this embodiment, the idle current samples are sorted by value, and the narrowest closed interval is selected from the closed intervals that can cover no less than 90% of the samples. When there are multiple intervals of equal width, the closed interval with the smaller lower end is selected. When the coverage ratio is less than 85%, it is easy to exclude normal idle fluctuations. When it is greater than 95%, it is easy to include the processor wake-up pulse in the idle reference.

[0031] The terminal uses the no-RF supply voltage and idle operating current before and after the candidate transmission event to perform piecewise linear interpolation according to the physical sampling time, forming a local supply voltage reference trajectory and an idle current reference trajectory respectively. The terminal forms a 90% minimum width coverage interval by the residual of the no-RF supply voltage relative to the local supply voltage reference trajectory, and extends the interval by two supply voltage sampling quantization steps at both ends to form the supply voltage idle allowable band. The local supply voltage reference trajectory, the supply voltage idle allowable band, and the idle current reference trajectory are used to eliminate the slow drift of the storage unit voltage and idle current before and after an event, and to determine whether the power supply status has been restored.

[0032] The terminal divides the pressure sampling window by the midpoint between adjacent pressure sampling times. For each pressure sampling window, the operating current of the communication module is subtracted from the idle current reference at the corresponding time. Differences less than 0 are set to 0, and the resulting incremental current is integrated over time to obtain the incremental charge. The local power supply voltage reference is subtracted from the actual power supply voltage, and the maximum value within the current pressure sampling window is taken for differences greater than 0 to obtain the maximum voltage drop. The absolute value of the incremental current change rate is integrated over time to obtain the load edge quantity. The incremental charge quantity is used to characterize the total additional load within the current window, the maximum voltage drop quantity is used to characterize the dynamic response of the common power supply path, and the load edge quantity is used to distinguish transmission processes with similar cumulative charge quantities but different pulse numbers or edge speeds.

[0033] The terminal connects the incremental charge, maximum voltage drop, and load edge quantity of each pressure sampling window according to the pressure sampling sequence to form the actual transmission load trajectory. It also saves the high-frequency operating current segment and high-frequency supply voltage segment within the transmission range. The actual transmission load trajectory is used to form the model input, and the high-frequency segment is used by the non-RF comparison modeling module to reproduce transient load and verify reproduction error.

[0034] The terminal compares the moment when the operating current of the communication module first continuously exceeds the upper limit of the idle current coverage area and the moment when the radio frequency transmission state first becomes effective, and determines the earlier one as the start of transmission. The duration of the operating current exceeding the upper limit is not less than the shortest load pulse duration of the communication module, so as to eliminate single-point sampling glitches.

[0035] After the radio frequency transmission ends and the wireless access state returns to idle, the terminal continues to detect the operating current and the supply voltage. The operating current continuously returns to the idle current coverage area, and the deviation of the supply voltage relative to the local supply voltage reference trajectory continuously returns to the supply voltage idle allowable band. After maintaining a value no less than the communication module connection release time, the electrical recovery time is determined. Starting from the electrical recovery time, the terminal sequentially selects two pressure observation windows that last for 1.5 seconds each, and performs local linear fitting with the minimum absolute deviation on the original pressure measurement values ​​in each window.

[0036] Before transmission, in the absence of radio frequency measurement, the terminal forms a local pressure slope sequence by dividing the difference between adjacent pressure samples by the corresponding actual time interval. Then, it subtracts the adjacent local pressure slopes to form a slope change sequence. The terminal uses the minimum width interval covering 90% of the slope change samples and takes the larger of the absolute values ​​of the two endpoints of this interval as the slope change boundary. When the difference between the fitted slopes of two consecutive pressure observation windows does not exceed this boundary, and the fitted residuals of the two observation windows are both within the minimum width coverage interval of the fitted residuals before transmission, the starting point of the second pressure observation window is determined as the transmission endpoint.

[0037] Candidate transmission events that fail to reach their transmission endpoint within 5.0 seconds of electrical recovery are marked as incomplete recovery events. Incomplete recovery events are defined by using the starting pressure boundary, allowable pressure slope interval, and upper limit of pressure curvature formed before transmission by the non-RF measurement segment, canceling the endpoint pressure boundary, and propagating backward to the maximum recovery observation time to form a forward physical pressure prediction interval. If an existing model is applicable and at least one valid non-RF measurement segment exists, the forward physical pressure prediction interval is intersected with the measurement-consistent pressure interval formed by the pressure recovery module; the resulting intersection is used only as a degraded pressure interval. If the model calling conditions are not met, the measurement error interval is not formed, or the intersection is empty, no definite pressure interval is formed, and only an unresolved state is output. Incomplete recovery events retain the original data but are not included in the calibration chain of the non-RF control modeling module.

[0038] Before the start of transmission, the terminal selects a no-RF measurement segment before transmission and after the end of transmission, it selects a no-RF measurement segment after recovery. Both no-RF measurement segments satisfy the following conditions: the communication module is idle, RF transmission is invalid, the operating current is within the idle current coverage range, the deviation of the supply voltage from the local supply voltage reference trajectory is within the idle allowable band of the supply voltage, the duration is 1.2 to 2.0 seconds, and there are no less than 12 effective pressure samples. In this embodiment, it is preferred to last for 1.5 seconds and obtain 15 effective pressure samples.

[0039] The terminal performs local linear fitting with the minimum absolute deviation on the non-RF measurement segments before transmission and after recovery. Based on the fitting results, residual coverage interval, and 0.05 kPa absolute pressure uncertainty boundary, it forms the pressure boundary at the start and end of the transmission action. The terminal forms a pressure slope sequence by dividing the difference between adjacent pressure samples by the actual time interval, and forms the pressure slope allowable interval by covering the interval with 90% minimum width of the slope sequence. Then, the difference between adjacent pressure slopes is divided by the interval between the corresponding times of the two slopes to form a pressure curvature sequence in kPa / s². The upper limit of the pressure curvature is formed by multiplying the upper limit of the 90% minimum width coverage interval of the absolute value of the pressure curvature by the curvature expansion coefficient. The curvature expansion coefficient can be 1.2 to 2.0, preferably 1.5. This range is used to cover the limited natural pressure acceleration or deceleration within the transmission action interval, avoiding unconditionally including continuous abrupt changes within the range of natural pressure changes.

[0040] Within the transmission range, candidate natural pressure values ​​at each pressure sampling time are used as variables to be determined. Starting pressure boundary, ending pressure boundary, upper and lower bounds of adjacent sample slopes, and upper bounds of adjacent slope changes are set to form a set of linear inequalities. The terminal first solves the feasibility of the set of inequalities. When the feasible region is empty, the current event does not enter the control calibration. When the feasible region is not empty, the minimum and maximum values ​​of candidate natural pressures are obtained for each pressure sampling time under the same constraints to form the allowable pressure change band at that time.

[0041] The continuous response hold time is taken as the larger value between the digital filter group delay and the three pressure sampling periods. In this embodiment, the digital filter group delay is 0.15s and the three pressure sampling periods are 0.30s. Therefore, the continuous response hold time is 0.30s. The terminal calculates the distance from the original pressure measurement value to the corresponding pressure allowable change zone at each pressure sampling time, and traverses the continuous sampling windows with a duration of not less than 0.30s and an interval between adjacent samples of not more than 0.30s. The minimum distance is taken in each window, and the maximum value is taken from the minimum values ​​corresponding to all windows to form the continuous response amplitude outside the interval.

[0042] The terminal calculates the maximum value of the allowable pressure change band width at each pressure sampling time. The sum of this maximum width and the absolute pressure uncertainty boundaries on both sides is divided by the sustained response amplitude outside the interval to form the calibration ambiguity ratio. The upper limit of the calibration ambiguity ratio can be 0.30 to 0.50, preferably 0.50. A value lower than 0.30 will significantly reduce the effective calibration events in weak coverage scenarios, while a value higher than 0.50 will result in insufficient distinction between natural pressure changes and transmission-related responses. When the sustained response amplitude outside the interval is 0, the calibration ambiguity ratio is not calculated, and the current event is set as invalid for calibration admission.

[0043] The representative real transmission event E_04 occurred in a weak coverage state with coverage enhancement level CE1 and transmit power configuration of 23dBm. The wireless access and uplink repeated transmission of E_04 lasted for 3.4s, the peak incremental current was 340mA, the maximum supply voltage drop was 0.18V, the maximum width of the pressure allowable variation band of E_04 was 0.08kPa, and the continuous response amplitude of the original pressure measurement value outside the pressure allowable variation band was 0.45kPa. Adding 0.08kPa and the absolute pressure uncertainty boundary of 0.10kPa on both sides and dividing by 0.45kPa, the calibration ambiguity ratio is 0.40, which is lower than the admission upper limit of 0.50.

[0044] The high-frequency operating current segment, high-frequency power supply voltage segment, original pressure measurement sequence, and independent RF transmission trajectory of E_04 are all recorded completely. The RF measurement segment before transmission and the RF measurement segment after recovery both last for 1.5s and contain 15 valid pressure samples. Therefore, E_04 is identified as a calibrable transmission event. The data acquisition and event construction module outputs the real transmission event record of E_04 without updating the bounded coupling model, event reporting threshold, or historical pressure benchmark.

[0045] The no-RF comparison modeling module constructs no-RF comparison events with the same load and forms a bounded coupling model. The no-RF comparison modeling module reads the real transmission event record of E_04, subtracts the idle current reference at the corresponding moment from the communication module's operating current during the real transmission period, sets the difference less than 0 to 0, and forms the real transmission incremental current trajectory. The controllable equivalent load branch uses the incremental current trajectory as the reproduction target and does not reproduce the idle operating current of the communication module.

[0046] After E_04 is completed and the pressure measurement channel meets the recovery conditions specified by the data acquisition and event construction module, the terminal obtains the initial power supply voltage, local pressure change slope, energy storage unit status quantity, and equivalent load branch temperature before E_04 starts and before Z_04 is executed. The initial power supply voltage is the median value of the power supply voltage sample within 0.5 seconds before the event starts; the local pressure change slope is the minimum fitted slope of the absolute deviation of the pressure samples in the same time period; the energy storage unit status quantity is the percentage of state of charge output by the fuel meter; and the equivalent load branch temperature is obtained by a temperature sensor located near the current absorption device.

[0047] The initial power supply voltage pairing tolerance can be 10-30mV, preferably 15mV; the local pressure change slope pairing tolerance can be 0.02-0.05kPa / s, preferably 0.03kPa / s; the energy storage unit status quantity pairing tolerance can be 3%-8% of the full scale, preferably 5%; the equivalent load branch temperature pairing tolerance can be 2-5℃, preferably 3℃. Each tolerance is the larger of the two quantization step sizes of the corresponding acquired quantity and the upper limit of the natural drift of the same terminal in the state of no transmission, and is fixed during the installation and commissioning stage.

[0048] The terminal divides the absolute value of the four state differences by the corresponding pairing tolerance, and takes the maximum value among the four normalized differences as the pairing state distance. The initial power supply voltages of E_04 and Z_04 before execution are 3.61V and 3.60V, respectively. The difference in the slope of the local pressure change is 0.020kPa / s, the difference in the state quantity of the energy storage unit is 2%, and the difference in the temperature of the equivalent load branch is 1.5℃. The maximum value among the four normalized differences is 0.67, which meets the execution condition that the pairing state distance is not greater than 1.

[0049] If an abnormal first-time transmission task, a confirmed out-of-limit transmission task, a pending alarm task, or a device fault message task occurs during the pairing waiting period, the pairing will be terminated. Regular periodic transmission tasks will enter the waiting queue. If the power supply voltage is lower than 3.30V, pairing will be terminated immediately and low voltage protection will be implemented. 3.30V is formed by adding the minimum power supply voltage of 3.20V for normal operation of the communication module and a dynamic margin of 0.10V, and is higher than the minimum operating voltage of the pressure acquisition chain.

[0050] The controllable equivalent load branch is implemented using a controlled current absorption circuit. The high-frequency control cycle is 50μs, and the minimum current control step size is 2mA. In each control cycle, the value of the actual transmitted incremental current trajectory at the corresponding moment is used as the feedforward target. When the current supply voltage of Z_04 is higher than the supply voltage of E_04 at the corresponding moment and the difference exceeds the allowable error of the supply voltage, one current control step size is added to the feedforward target. When the supply voltage of Z_04 is lower than the supply voltage of E_04 at the corresponding moment and the difference exceeds the allowable error of the supply voltage, one current control step size is reduced from the feedforward target. When the difference is within the allowable error, the feedforward target is maintained. The corrected target current is simultaneously limited by the upper limit of current, the upper limit of current change rate, the upper limit of accumulated charge, and the upper limit of duration.

[0051] The upper limit of the allowable current of the equivalent load branch can be 350-450mA, preferably 400mA; the upper limit of the allowable current change rate can be 0.5-1.5A / ms, preferably 1.0A / ms; the upper limit of the allowable charge amount extracted in a single instance can be 1.0-2.0C, preferably 1.5C; the upper limit of the allowable duration in a single instance can be 4-8s, preferably 6s. The peak incremental current of E_04 and other valid events is not higher than 340mA. The cumulative incremental charge of E_04 is about 1.16C, lasting for 3.4s. The aforementioned preferred values ​​can cover the load process of the valid events and retain the safety margin of the device and the energy storage unit.

[0052] During the execution of Z_04, if the power supply voltage is lower than 3.30V, or if a device fault message task occurs, or if a high-level transmission task originating from a pre-processed preceding event, an independent safety sensing channel, or an external control command occurs, the terminal immediately suspends the equivalent load process. Current events that have not yet been processed by the pressure recovery module cannot generate a high-level transmission task to suspend their own recovery process. The suspended equivalent load process does not form a valid no-RF load event, and its pressure measurement data does not enter the model update chain.

[0053] The actual incremental current trajectory and the Z_04 actual current trajectory are aligned relative to each other after the time origin is the moment when the current first continuously exceeds the upper limit of the idle current coverage interval. The nonlinear time mapping does not use local stretching or compression of the time axis.

[0054] The allowable error for the current trajectory can be 3% to 5% of the peak incremental current of E_04, preferably 5%; the allowable error for the cumulative incremental charge can be 3% to 5% of the cumulative incremental charge of E_04, preferably 5%; the allowable error for the power supply voltage trajectory can be 5% to 10% of the maximum power supply voltage drop of E_04, preferably 10%; the allowable error for the load edge timing can be the larger of 5% to 10% of the shortest load pulse duration of the communication module and two high-frequency sampling periods, which is 0.10ms in this embodiment.

[0055] For each high-frequency sampling moment, the terminal divides the absolute value of the difference between the actual current of Z_04 and the incremental current of E_04 by the allowable error of the current trajectory to form the current trajectory normalization error; it divides the absolute value of the difference between the supply voltage of Z_04 and the corresponding supply voltage of E_04 by the allowable error of the supply voltage to form the supply voltage normalization error. The terminal forms a 90% minimum width coverage interval for the two types of normalization errors respectively, and takes the upper boundary of the corresponding interval as the current trajectory mismatch and the supply voltage trajectory mismatch. The cumulative incremental charge mismatch is formed by dividing the absolute value of the difference between the cumulative incremental charge of the two processes by the allowable error of the cumulative incremental charge. The load edge moment mismatch is formed by dividing the difference between the edge moments corresponding to the occurrence sequence by the allowable error of the edge moment and taking the maximum value. When the number of effective edges is inconsistent, it is directly determined as a load reproduction mismatch. The terminal takes the maximum value among the four types of mismatch as the load reproduction mismatch.

[0056] When the pairing distance is no greater than 1, the load reproduction mismatch is no greater than 1, and the equivalent load process is not stopped by the safety interlock, the process is determined as a no-RF load event Z_04.

[0057] Before Z_04 begins and after the pressure measurement channel is restored, the terminal acquires the non-RF measurement segment and forms the pressure allowable change band of Z_04 itself according to the same duration, sample number and natural pressure constraint processing method as the data acquisition and event construction module. The terminal extends the original pressure measurement value of each pressure sampling moment of Z_04 downward and upward by 0.05 kPa respectively to form the non-RF load measurement interval. The lower boundary of the non-RF load response interval is formed by subtracting the upper boundary of the pressure allowable change band from the lower boundary of the measurement interval, and the upper boundary of the non-RF load response interval is formed by subtracting the lower boundary of the pressure allowable change band from the upper boundary of the measurement interval.

[0058] At a pressure sampling moment of Z_04, the original pressure measurement value is 100.42 kPa. The pressure variation range of Z_04 itself is 100.05 to 100.12 kPa, and the resulting measurement range without RF load is 100.37 to 100.47 kPa. Thus, the response range without RF load is obtained as 0.25 to 0.42 kPa.

[0059] The incremental charge, maximum voltage drop, and load edge quantity correspond to the total additional load, the voltage response of the common power supply path, and the load edge excitation, respectively. When the initial power supply voltage, the state quantity of the energy storage unit, and the load amplitude are within the coverage range of paired events, the short-time response of the common power supply and common grounding paths to the pressure acquisition chain is represented by a local linear combination with finite memory. This relationship is only called within the continuously applicable state, discrete applicable state, load amplitude range, and effective feature value coverage range saved in the current model version. The equivalent load branch temperature is only used as the pairing execution condition of the modeling module without RF reference and is not used as the model calling condition for real transmission events.

[0060] The power supply coupling memory time is determined using only E_01 to E_05. In the no-RF load events corresponding to E_01 to E_05, the duration of the last non-zero period in the response interval after the load ends is 0.63 to 0.74 s, and it continuously contains zero for at least 0.30 s thereafter. Therefore, the power supply coupling memory time can be taken as 0.7 to 1.0 s, preferably 0.8 s. At a pressure sampling frequency of 10 Hz, 0.8 s corresponds to 8 preceding pressure sampling windows. The power supply load regression row is formed by arranging the incremental charge, maximum voltage drop, and load edge quantity in the current pressure sampling window and the previous 8 pressure sampling windows in a fixed order, with a total of 27 candidate feature columns.

[0061] The terminal selects regression rows from non-RF load events with different load durations, pulse counts, or load edge structures to form a power supply load regression matrix. Each valid regression row corresponds to a unique response pressure sampling window, which does not overlap with the response pressure sampling windows of other valid regression rows. The preceding memory windows of different regression rows can share historical sampling data, but the shared portion is not counted as a new response sample. The terminal performs dimensionless scaling on each candidate feature column using the width of the minimum width coverage interval of that column. If the column width is less than three quantization steps of the feature, the column is deleted. The retained valid feature columns should have valid changes exceeding three quantization steps in at least three pairs of events used for modeling. Feature columns that do not meet this condition are deleted. In this embodiment, six feature columns with insufficient changes or insufficient changes across events are deleted, and 21 valid feature columns are retained. The scaling parameters and training sample value coverage intervals of each valid feature column are written into the model version.

[0062] The total number of valid paired events can be 4 to 8, preferably 6. Among them, no less than 4 pairs of paired events are used for solving the coefficient set, and at least 1 pair is reserved not to participate in the coefficient set solution and is used for independent verification. The ratio of the number of valid regression rows to the number of valid feature columns can be 3 to 6, preferably not less than 4. The terminal performs singular value decomposition on the dimensionless power supply load regression matrix, and uses the ratio of the minimum singular value to the maximum singular value as the power supply model identification margin. When the maximum singular value is 0, the power supply model identification margin is 0. The lower limit of the power supply model identification margin can be 0.05 to 0.10, preferably 0.10. In the J-03 model establishment stage, E_01 to E_05 form 96 valid power supply regression rows with unique response pressure sampling windows. The ratio of 96 to 21 is greater than 4, and the power supply model identification margin is 0.12.

[0063] After meeting the requirements for the number of valid events, the ratio of valid rows to columns, and the identification margin, the terminal establishes a linear constraint for each valid regression row, ensuring that the predicted response is not lower than the lower bound of the corresponding no-RF load response interval and not higher than its upper bound. All power supply coupling coefficient vectors that satisfy the constraints form a power supply coupling coefficient set. The terminal performs maximization and minimization solutions with each coefficient direction as the objective. If a finite upper bound or a finite lower bound cannot be obtained in any direction, the current data does not form a callable bounded power supply coupling model.

[0064] After the bounded power supply coupling model is formed, the terminal forms the total transmission-related response interval based on the original pressure measurement values ​​and allowable pressure variation bands of each real transmission event in E_01~E_05, and forms the power supply coupling prediction interval based on the corresponding actual transmission load trajectory and power supply coupling coefficient set. Let the total transmission-related response interval be its lower bound to upper bound, and the power supply coupling prediction interval be its lower bound to upper bound. Then, the lower bound of the radio frequency (RF)-related coupling response interval is formed by subtracting the upper bound of the power supply coupling prediction interval from the lower bound of the total transmission-related response interval, and the upper bound of the RF-related coupling response interval is formed by subtracting the lower bound of the power supply coupling prediction interval from the upper bound of the total transmission-related response interval. The RF-related coupling response interval represents the remaining transmission-related response that cannot be explained by natural pressure changes and the effects of common power supply and common grounding. Its range simultaneously includes the RF effect and the interactive residuals not described separately by the power supply model, and it is not determined as a pure RF radiation response.

[0065] The radio frequency (RF) associated coupling memory time is determined using only E_01 to E_05. Within E_01 to E_05, the duration of the last RF associated coupling response interval after the RF transmission ends without containing 0 is 0.36 to 0.44 seconds, and it continuously contains 0 for at least 0.30 seconds thereafter. Therefore, the RF associated coupling memory time can be taken as 0.4 to 0.7 seconds, preferably 0.5 seconds. At a pressure sampling frequency of 10 Hz, the RF transmission regression row is formed by the RF duty cycle and the number of RF start-stop cycles in the current pressure sampling window and the previous 5 pressure sampling windows, with a total of 12 candidate feature columns. In this embodiment, all 12 feature columns are retained, and 72 effective RF regression rows with unique response pressure sampling windows are formed by E_01 to E_05. The RF model identification margin is formed by the ratio of the minimum singular value to the maximum singular value of the RF transmission regression matrix, which is 0.11 in this embodiment.

[0066] Only real transmission events with the same coverage enhancement level, transmit power configuration, and RF transmit trajectory source label are written into the same RF correlation coupling model, and the RF correlation coupling coefficient set is formed according to the same boundedness, effective row number, and identification margin rules as the power supply model.

[0067] After the power supply coupling memory time, radio frequency correlation coupling memory time, effective feature columns, feature scaling parameters, and two types of coefficient sets are all determined and frozen by E_01 to E_05, E_06 is read to perform independent verification. E_06 does not participate in the formation of memory time, effective feature columns, scaling parameters, applicable range, or coefficient sets. For each effective regression row in E_06, a power supply prediction response interval and a radio frequency correlation prediction response interval are formed respectively. Only when the corresponding observed response interval is completely within the prediction response interval is the regression row considered to have passed the verification. After all effective regression rows in E_06 have passed the verification, model version V2.2 is generated.

[0068] V2.2 saves two types of coefficient sets, the order of effective feature columns, feature scaling parameters, the coverage range of effective feature values, power supply coupling memory time, radio frequency correlation coupling memory time, fixed processing delay, pressure sampling window division rules, and model applicable states. The continuous applicable states of V2.2 include the initial power supply voltage of 3.55 to 3.68V and the state of charge of the energy storage unit of 60% to 80%. The discrete applicable states include CE1, 23dBm, and independent radio frequency envelope sources.

[0069] When a new valid paired event matches the applicable state of the V2.2 model, the constraints formed by the new event are intersected with the original coefficient set. After the power supply coupling coefficient set changes, the power supply coupling prediction interval, radio frequency associated coupling response interval, and radio frequency associated coupling coefficient set of the actual transmission event are recalculated. If the intersection result is empty, V2.2 is retained, and the conflicting events are grouped according to the initial power supply voltage, energy storage unit state quantity, equivalent load branch temperature, coverage enhancement level, and transmission power configuration. If the grouped samples do not meet the complete identification conditions, no new model version is generated.

[0070] The pressure recovery module recovers the feasible range of real pressure based on the bounded coupling model. After model version V2.2 is put into operation, the J-03 terminal performs a periodic transmission according to the existing monitoring plan. This transmission forms a real transmission event E_07. The non-radio stable pressure measurement value in the 10 seconds before E_07 transmission is 100.42~100.46kPa, which is lower than the 100.50kPa event reporting threshold. During E_07 transmission, the original pressure measurement value reaches a maximum of 100.80kPa. This change occurs during the wireless transmission process. The terminal writes it into the E_07 record and does not create a new abnormal transmission task based on this.

[0071] After the E_07 RF transmission ends, the terminal completes power restoration, pressure measurement channel restoration, and post-restoration acquisition of no-RF measurement segments. Then, it reads model version V2.2. The initial power supply voltage and energy storage unit status of E_07 are within the continuous applicable range saved in V2.2. The sources of CE1, 23dBm, and independent RF envelope are consistent with the discrete applicable states saved in V2.2. Therefore, V2.2 is invoked. The measurement data after the event is completed is only used to process E_07 and its pending records, and is not written back as real-time control input during the E_07 transmission. The equivalent load branch temperature is only used for pairing with the no-RF reference modeling module and is not used as a model invocation condition for the actual transmission event.

[0072] The terminal generates power supply load regression rows and RF transmission regression rows for each pressure sampling moment in E_07 according to the fixed processing delay, pressure sampling window division rules, effective feature column order, feature scaling parameters, and two types of memory time saved in V2.2. The terminal does not regenerate feature scaling parameters based on E_07. When all effective feature values ​​are within the coverage range of the training sample values ​​saved in V2.2, it is processed as a normal model state. When any effective feature value exceeds the corresponding coverage range, the current pressure sampling moment is determined as a model input out-of-range state, and a normal real pressure feasible range is not formed through model extrapolation.

[0073] When the preceding memory window is located before the start of transmission, the preceding features are supplemented with 0 only when the corresponding pressure sampling window belongs to the segment without radio frequency measurement before transmission, and the incremental operating current is 0, the radio frequency transmission state is invalid, the operating current is located in the idle current coverage area, and the deviation of the supply voltage relative to the local supply voltage reference trajectory is located in the idle allowable band of the supply voltage; if the aforementioned states cannot be confirmed, the current pressure sampling time is marked as incomplete model input.

[0074] For each effective power supply load return row, the terminal calculates the maximum response in the current input direction and the maximum response in the opposite input direction under the constraint of the power supply coupling coefficient set. The maximum response in the current input direction is used as the upper bound of the power supply coupling response, and the maximum response in the opposite input direction is taken as the lower bound of the power supply coupling response. The radio frequency associated coupling response interval is formed in the same way. When the terminal uses pre-calculated support values, the pre-calculated result is the outer envelope containing the actual support values.

[0075] The terminal adds the lower bound of the power supply coupling response interval to the lower bound of the radio frequency correlation coupling response interval, and adds the corresponding upper bound to form the transmission coupling feasible interval. This interval is a conservative outer envelope of all feasible combinations of the two types of coefficient sets, and the midpoint of the interval is not used as the only transmission correction quantity.

[0076] The terminal forms a 90% minimum width coverage interval for the local fitting residuals of the non-RF measurement segment before transmission and the non-RF measurement segment after recovery. When both non-RF measurement segments are valid, the minimum envelope of the two residual coverage intervals and the 0 value is taken, and 0.05 kPa is subtracted from the lower bound and 0.05 kPa is added to the upper bound to form the measurement error interval for the current event. When only one non-RF measurement segment is valid, the minimum envelope of the residual coverage interval of that measurement segment and the 0 value is taken and expanded in the same way to form a degraded measurement error interval. Only degraded pressure intervals are allowed to be formed for the corresponding event. When both non-RF measurement segments are invalid, no measurement error interval is formed, and no feasible interval for true pressure is output. If the original pressure sample shows analog-to-digital conversion saturation, sensor fault identification, or missing sampling timestamp, the sample is marked as invalid.

[0077] For each valid pressure sampling moment, the terminal subtracts the upper bound of the measurement error interval from the original pressure measurement value to form the lower bound of the observed pressure interval, and subtracts the lower bound of the measurement error interval from the original pressure measurement value to form the upper bound of the observed pressure interval; then subtracts the upper bound of the transmission coupling feasible interval from the lower bound of the observed pressure interval to form the lower bound of the measurement consistent pressure interval, and subtracts the lower bound of the transmission coupling feasible interval from the upper bound of the observed pressure interval to form the upper bound of the measurement consistent pressure interval.

[0078] The first pressure sampling moment to be recovered within the transmission range uses the pressure allowable change band corresponding to the data acquisition and event construction module as the physical pressure prediction range. For subsequent current pressure sampling moments, the terminal subtracts the upper limit of the pressure allowable change band of the previous pressure sampling moment from the lower limit of the current pressure allowable change band to form the lower limit of the natural pressure increment, and subtracts the lower limit of the pressure allowable change band of the previous pressure sampling moment from the upper limit of the current pressure allowable change band to form the upper limit of the natural pressure increment. The range of the natural pressure increment is added to the execution range of the true pressure feasible range of the previous pressure sampling moment, and then intersected with the current pressure allowable change band to form the current physical pressure prediction range. If the true pressure feasible range is not formed at the previous pressure sampling moment, the physical pressure prediction range of the previous pressure sampling moment is used to continue the recursion, and the subsequent results are marked as degraded results, until subsequent valid non-RF measurement data re-forms the measurement anchor.

[0079] The intersection of the predicted physical pressure range and the measured consistent pressure range is determined as the feasible range of the actual pressure. Taking a pressure sampling moment during the E_07 transmission as an example, the original pressure measurement value is 100.80 kPa. The core residual range before transmission is -0.01 to 0.02 kPa, and the core residual range after recovery is -0.01 to 0.01 kPa. Combined with the 0.05 kPa absolute pressure uncertainty boundary, the measurement error range is -0.06 to 0.07 kPa, the observed pressure range is 100.73 to 100.86 kPa, the power supply coupling response range formed by V2.2 is -0.35 to 0.55 kPa, the radio frequency correlation coupling response range is -0.12 to 0.24 kPa, and the transmission coupling feasible range is -0.47 to 0.79 kPa. Thus, the measured consistent pressure range of 99.94 to 101.33 kPa is obtained.

[0080] The data acquisition and event construction module forms a physical pressure prediction range of 100.43–100.65 kPa based on the natural pressure state before and after the E_07 transmission. The intersection of the physical pressure prediction range and the measured pressure range yields the actual pressure feasible range of 100.43–100.65 kPa. This range crosses the 100.50 kPa event reporting threshold. Therefore, the original pressure measurement value of 100.80 kPa during the E_07 transmission does not form a definite conclusion of exceeding the limit.

[0081] When the physical pressure prediction interval and the measured consistent pressure interval do not intersect, the terminal forms the model consistency residual with the shortest interval between the two intervals, and forms the degraded pressure interval with the smallest closed interval that can cover the two intervals. When the local mismatch reaches 0.30s continuously, the model applicability status of V2.2 for E_07 is determined to be inapplicable. When a non-empty intersection occurs, the continuous local mismatch timer is reset to zero, and the model input out of range, model input incomplete, degraded pressure interval and mismatched samples do not enter the model coefficient update chain.

[0082] To ensure continuous data sources for post-transmission confirmation, from the end of the transmission until the start of the next RF transmission, the terminal continues to form a pressure sampling time at a sampling frequency of 10Hz. If the preceding power supply coupling memory window or the RF-related coupling memory window still contains the E_07 feature, the corresponding coupling response interval is calculated according to V2.2. Only when the communication module remains idle, the RF transmission state is invalid, the incremental operating current is 0, the operating current is within the idle current coverage area, the deviation of the power supply voltage from the local power supply voltage reference trajectory is within the power supply voltage idle allowable band, and the event features in both types of memory windows have been moved out, are the power supply coupling response interval and the RF-related coupling response interval set to [0, 0].

[0083] The terminal uses the two most recent 1.5s non-RF rolling measurement segments to form the subsequent allowable pressure change band according to the same fitting, slope, and curvature constraints as the data acquisition and event construction module. It also forms a non-RF measurement consistent pressure range based on the current measurement error range. The intersection of the subsequent allowable pressure change band and the non-RF measurement consistent pressure range is used as the real pressure feasible range for the continuous confirmation phase after transmission. After two consecutive rolling measurement segments are valid and the intersection is not empty, the aforementioned downgrade recursion flag is removed. The non-RF sample after transmission is saved as the transmission judgment data corresponding to E_07. It is not added to the E_07 real transmission event record used for identification by the non-RF comparison modeling module, and the original allowable pressure change band of E_07 is not changed. If any idle condition is not met, the current continuous confirmation is stopped, and the new communication process is established as a new real transmission event.

[0084] The transmission control and calibration module controls repeated transmissions and performs closed-loop calibration based on the actual pressure feasible range. The terminal enters a transmission control suppression state from the E_07 transmission start point until the pressure recovery module completes E_07 processing and outputs the result. During suppression, the terminal continues to collect and cache data; original pressure measurements must not create new abnormal transmission tasks. Regular periodic transmission tasks enter the waiting queue, and only the latest task is retained within the same cycle. After the pressure recovery module outputs a transmission coupling pending state, the original pressure measurements must not independently create abnormal transmission tasks until the pending record is released or converted to a definite out-of-limit state. Subsequent tasks are formed based on the state rules of the actual pressure feasible range or the degraded pressure range. Necessary protocol interactions, low voltage protection, and equipment fault message tasks within the established connection are executed according to the existing priority. High-level transmission tasks originating from completed preceding events, independent safety sensing channels, or external control commands can interrupt the current recovery process. E_07, which has not yet completed the pressure recovery module processing, cannot generate high-level transmission tasks to interrupt its own recovery process. When the recovery process is interrupted, E_07 is marked as an incomplete recovery event, and the interrupted communication process uses a new event identifier to form a real transmission event record.

[0085] After the pressure recovery module outputs, if the model is applicable and the lower bound of the real pressure feasible interval is higher than 100.50 kPa, the corresponding pressure sampling time is determined as the sampling time when the limit is exceeded; if the upper bound of the real pressure feasible interval is lower than 100.50 kPa, it is determined as the sampling time when the limit is not exceeded; if the real pressure feasible interval crosses 100.50 kPa, only a degraded pressure interval is formed, the model input is out of range, the model input is incomplete, the model has a local mismatch, the model is not applicable, or no effective pressure interval is formed, it is determined as the sampling time when transmission coupling is unresolved.

[0086] The duration of abnormal pressure confirmation can be 5 to 20 seconds, preferably 10 seconds. This range is derived from the existing abnormal confirmation requirements of the J-03 monitoring center. Less than 5 seconds may cause short-term hydraulic fluctuations to be classified as abnormal tasks, while more than 20 seconds will exceed the allowable normal over-limit confirmation delay for this monitoring point. The maximum allowable continuous interval between adjacent pressure samples can be 2 to 5 times the normal pressure sampling cycle, preferably 3 times, which is 0.30 seconds at a 10Hz sampling frequency. When the interval between adjacent samples exceeds 0.30 seconds, the current continuous state is reset.

[0087] When the lower bound of the feasible range of real pressure is continuously higher than 100.50 kPa for 10 seconds after transmission, and the applicable state of the model is applicable during the corresponding time period, the terminal forms a definite over-limit state and generates a definite over-limit transmission task. Before the upper bound of the feasible range of real pressure is continuously lower than 100.50 kPa for 10 seconds, the terminal maintains the same continuous over-limit state and does not repeatedly create abnormal transmission tasks due to the formation of new event identifiers by subsequent wireless transmissions.

[0088] The safety pressure threshold adopts the existing engineering safety classification configuration of J-03. In this embodiment, it is set to 105.00 kPa. When the lower limit of the actual pressure feasible range reaches 105.00 kPa, the terminal generates a high-level confirmed over-limit sending task after the pressure recovery module has completed processing and the communication module has the conditions for re-access, without waiting for the 10-second confirmation time. When only the upper limit of the actual pressure feasible range or the downgraded pressure range reaches 105.00 kPa, it remains in an unresolved state and generates a high-priority alarm with an unresolved label, without forming a high-level confirmed over-limit conclusion.

[0089] The actual feasible pressure range for E_07 is 100.43–100.65 kPa, which crosses the 100.50 kPa event reporting threshold. Therefore, the terminal establishes a transmission coupling pending record. This record is bound to the first pending sampling time, the E_07 event identifier, the subsequent associated event identifier, the actual feasible pressure range or downgraded pressure range for each pressure sampling time, the model version V2.2, the model applicability status, and the model consistency residual. Subsequent associated real transmission events are only appended to the same pending record, and the first pending sampling time is not re-determined. During the existence of the pending record, the original pressure measurement value does not independently generate new abnormal transmission tasks.

[0090] The maximum allowed delay time for reporting pending statuses can be 30-120 seconds, with 60 seconds being preferred. Less than 30 seconds may not be enough to cover a weak coverage transmission, a maximum recovery observation of 5.0 seconds, and subsequent interval processing. More than 120 seconds exceeds the allowable reporting time limit for uncertain anomalies by the J-03 monitoring center. The minimum re-access interval for the communication module can be 30-120 seconds, with 60 seconds being preferred. This interval only restricts repeated pending alarms for the same pending record and does not delay the first pending alarm.

[0091] If the E_07 pending status remains unresolved for 60 seconds, the model's applicability is determined to be inapplicable, the upper limit of the pending interval reaches 105.00 kPa, or the data discontinuity persists for 60 seconds, the terminal generates a pending alarm task. The pending alarm does not resolve the pending record, nor does it reset the first pending sampling time. If the upper limit of the subsequent real pressure feasible interval is continuously lower than 100.50 kPa for 10 seconds, the pending record is resolved. If the lower limit of the subsequent real pressure feasible interval is continuously higher than 100.50 kPa for 10 seconds, the pending record is converted to a confirmed over-limit status.

[0092] The pending alarm message is bound to the E_07 event identifier, pressure sampling time, and model version V2.2, and carries the original pressure measurement value, pressure allowable variation range, measurement error range, power supply coupling response range, radio frequency correlation coupling response range, real pressure feasible range or degraded pressure range, model consistency residual, and model applicability status. When the remote receiving device does not obtain a complete record of the real transmission event and complete model parameters, it only checks the interval subtraction, interval intersection, and threshold position relationship, and does not regenerate the two types of coupling response ranges. When the remote check result is inconsistent with the terminal result, the current record is marked as an inconsistent record, and the model coefficient set is not modified accordingly.

[0093] Within the same pending record, the remote receiving device selects sampling times when both the power supply coupling response interval and the radio frequency associated coupling response interval are valid, and forms a continuous pending window with a duration of not less than 0.30s and an interval between adjacent samples not exceeding 0.30s. For each continuous pending window, the minimum value of the width of the power supply coupling response interval and the minimum value of the width of the radio frequency associated coupling response interval within the window are taken respectively. Then, the maximum value is taken from all continuous pending windows to form the power supply persistence uncertainty width and the radio frequency associated persistence uncertainty width. When the power supply persistence uncertainty width is not less than the radio frequency associated persistence uncertainty width, the remote end forms a power supply coupling priority target; when the radio frequency associated persistence uncertainty width is greater than the power supply persistence uncertainty width, the remote end forms a radio frequency associated coupling supplementary target. When either type of response interval is invalid within the continuous pending window, the remote end does not form a model distinguishing target. The remote end only returns the model distinguishing target and the corresponding model version identifier, and does not directly send any equivalent load waveform to the terminal.

[0094] In the E_07 pending record, the power supply persistence uncertainty width is greater than the radio frequency association persistence uncertainty width. The remote end returns the power supply coupling priority target and model version V2.2. After the terminal receives the target, it selects the target equivalent load sequence from the local security calibration sequence set. Each candidate sequence in the security calibration sequence set comes from a real transmission event with valid calibration admission and meets the current upper limit, current change rate upper limit, cumulative charge upper limit and duration upper limit specified by the no-radio frequency comparison modeling module.

[0095] For each candidate sequence, the terminal forms a candidate power supply load regression row according to the valid feature column order, feature scaling parameters and power supply coupling memory time saved in V2.2, and forms a candidate power supply prediction interval on the power supply coupling coefficient set. The terminal calculates the ratio of the candidate power supply prediction interval width to the sum of the allowable change band width of the source event pressure and the absolute pressure uncertainty boundary on both sides. The minimum value of the ratio is taken within the candidate window with a duration of not less than 0.30s, and the maximum value is taken from all candidate windows to form the observable distinguishability of the candidate sequence.

[0096] Candidate sequence A originates from E_04. Its candidate power supply prediction interval width within a continuous 0.30s window is 0.28kPa, and the sum of the allowable variation band of the source event pressure and the uncertain absolute pressure boundaries on both sides is 0.18kPa, with an observable discrimination value of 1.56. The corresponding values ​​for candidate sequence B are 0.25kPa and 0.21kPa, with an observable discrimination value of 1.19. Therefore, candidate sequence A is selected. When multiple candidate sequences have the same observable discrimination value, the one with the smaller cumulative extracted charge is preferred. If the cumulative extracted charge is still the same, the one with the shorter duration is selected.

[0097] The terminal executes candidate sequence A only when E_04 calibration access is valid, the distance between the current terminal state and the pairing state of E_04 is no greater than 1, there is no abnormal first-time transmission task, no confirmed over-limit transmission task, no pending alarm task, no equipment fault message task, and the power supply voltage is higher than 3.30V. If an equipment fault message task, a high-level transmission task originating from a completed preceding event, an independent safety sensor channel or external control command occurs during execution, or a low voltage interlock occurs, the equivalent load process is immediately terminated.

[0098] After the no-RF load event formed by candidate sequence A passes the load reproduction mismatch verification of the no-RF control modeling module, it is checked together with the existing paired events of E_01 to E_05 for the number of effective events, effective row-column ratio, and identification margin. Only the power supply coupling coefficient set is directly adjusted. After the power supply coupling coefficient set changes, the no-RF control modeling module re-forms the power supply coupling prediction interval, RF correlation coupling response interval, and RF correlation coupling coefficient set of the real transmission event. E_06 continues to not participate in the model update and is independently reviewed after the model structure, feature parameters, and two types of coefficient sets are re-fixed. If the review of E_06 fails, V2.2 is retained and no new model version is generated.

[0099] After the RF correlation coupling supplementary target is formed, the terminal does not actively create a wireless transmission task for the calibration purpose, nor does it change the coverage enhancement level or transmission power configuration during normal transmission. After subsequent periodic transmissions, abnormal first transmissions, or restricted pending alarms naturally form new real transmission events, the no-RF comparison modeling module determines after the event is completed whether it has a RF transmission trajectory source consistent with the target model, an effective calibration access status, and whether it can provide a new RF start-stop structure or improve the identification margin of the RF transmission regression matrix.

[0100] After the no-RF load event formed by candidate sequence A completes the aforementioned model update verification, and all valid regression rows of E_06 pass independent review again, the no-RF control modeling module generates model version V2.3. V2.3 does not cover V2.2. The original real transmission event record of E_07 is completely preserved, and its initial power supply voltage, energy storage unit status, coverage enhancement level, transmission power configuration, RF transmission trajectory source, and valid feature values ​​meet the applicable conditions of V2.3. Therefore, the terminal uses V2.3 to reprocess the unresolved E_07 pending records and the corresponding no-RF continuous judgment data after transmission.

[0101] After verification, within the continuous 10-second effective pressure sampling period associated with E_07, the lower bound of the feasible pressure range was higher than 100.50 kPa. The feasible pressure range at the representative pressure sampling time was 100.56–100.66 kPa. The terminal converted the pending record of E_07 to a confirmed out-of-limit status and generated a confirmed out-of-limit transmission task. The original pending results and pending alarm records formed in V2.2 continued to be saved. V2.3 is only used for subsequent real transmission events and pending records that meet its applicable conditions and have not yet been resolved. It does not rewrite the historical pressure results that have already formed a confirmed conclusion.

[0102] In this embodiment, the actual transmission load trajectory and independent radio frequency transmission trajectory are formed by the actual transmission in E_04, and the same load without radio frequency is established by Z_04. Bounded responses and radio frequency associated residual responses of common power supply and common ground paths are formed respectively. In E_07, the terminal restores the original pressure measurement value of 100.80kPa during the transmission period to the real pressure feasible range that crosses the 100.50kPa event reporting threshold. First, it maintains the pending state and prevents the original pressure measurement value from directly creating an abnormal repetitive transmission task. Then, according to the power supply coupling priority target returned by the remote end, the model boundary is adjusted by the local security calibration sequence. After V2.3 verification, the lower limit of the real pressure feasible range crosses the event reporting threshold during the continuous confirmation period. Based on this, the terminal forms a confirmed over-limit transmission task.

[0103] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0104] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0105] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A water conservancy data acquisition and transmission system based on narrowband Internet of Things, characterized in that, include: The data acquisition and event construction module acquires the pressure measurement sequence of the same pressure telemetry terminal and the transmission operation data of the narrowband IoT communication module, performs timing correction, forms the actual transmission load trajectory and radio frequency transmission trajectory based on the corrected transmission operation data, divides the transmission action range, and constructs real transmission events. Based on the non-radio frequency measurement data before and after the transmission range, a pressure allowable change band is formed to calibrate and determine the admission criteria for real transmission events. The no-RF comparison modeling module reads real transmission events that are valid for calibration and admission, and controls the controllable equivalent load branch that shares the power supply path and grounding path with the narrowband IoT communication module to reproduce the incremental current trajectory and form no-RF load events. Based on the power supply coupling constraint formed by the no-RF load event, the power supply coupling response is subtracted from the total transmission associated response of the real transmission event to form the RF associated coupling constraint and establish a bounded coupling model. The pressure recovery module uses a bounded coupling model to form a feasible transmission coupling interval based on the load trajectory and radio frequency transmission trajectory of the current real transmission event; it performs error expansion on the original pressure measurement value to form an observed pressure interval; it subtracts the observed pressure interval from the feasible transmission coupling interval to form a measurement-consistent pressure interval; and it intersects the measurement-consistent pressure interval with the physical pressure prediction interval formed based on the allowable pressure variation zone to form the real pressure feasible interval. The transmission control and calibration module determines the tasks to be transmitted from the identified over-limit transmission tasks and pending alarm tasks based on the positional relationship between the feasible range of the actual pressure and the event reporting threshold, and transmits them to the remote receiving device via the narrowband IoT communication module; it locks abnormal repeated transmissions triggered by the original pressure measurement value during the current actual transmission event processing; it receives the model-distinguished target returned by the remote receiving device for pending alarm tasks, and when the model distinguishes the target as pointing to the power supply coupling constraint, it selects the target sequence to perform radio frequency-free calibration, incorporates the calibration constraint into the power supply coupling constraint, and updates the bounded coupling model.

2. The water conservancy data acquisition and transmission system based on narrowband Internet of Things as described in claim 1, characterized in that, The transmitted operational data includes the operating current, power supply voltage, and RF transmission status of the narrowband IoT communication module; The data acquisition and event construction module performs timing correction on the pressure measurement sequence and the transmitted operation data according to the fixed processing delay pre-calibrated in the pressure acquisition link and the transmission operation data acquisition link.

3. The water conservancy data acquisition and transmission system based on narrowband Internet of Things according to claim 2, characterized in that, The data acquisition and event construction module forms an idle current coverage interval based on the working current in the non-RF state, and sets the earlier of the start time when the working current continuously exceeds the upper limit of the idle current coverage interval and the start time when the RF transmission state takes effect as the transmission start point. After the radio frequency transmission ends and the operating current and supply voltage return to their corresponding idle ranges, the moment when the fitting slope and fitting residual of the continuous pressure observation window meet the stability threshold is set as the end point of the transmission action.

4. The water conservancy data acquisition and transmission system based on narrowband Internet of Things according to claim 3, characterized in that, The data acquisition and event construction module forms natural pressure constraints based on the non-radio frequency measurement data before and after the transmission range, constructs a set of candidate natural pressure sequences, and forms a pressure allowable change band based on the candidate natural pressure boundaries at each pressure sampling time. The maximum width of the allowable pressure variation band and the preset pressure measurement error width are divided by the continuous deviation of the original pressure measurement value to form the calibration ambiguity ratio. Real transmission events with a calibration ambiguity ratio not greater than the calibration access threshold are set as valid calibration access.

5. The water conservancy data acquisition and transmission system based on narrowband Internet of Things according to claim 1, characterized in that, The no-RF control modeling module extracts incremental current trajectories from real transmission events that have been calibrated and approved. When the initial power supply state and pressure change state of the real transmission event and the no-RF load reproduction process to be executed meet the pairing tolerance, the controllable equivalent load branch is driven to reproduce the incremental current trajectory. When the reproduction error is no greater than the reproduction threshold, a no-RF load event is generated.

6. The water conservancy data acquisition and transmission system based on narrowband Internet of Things according to claim 5, characterized in that, The non-RF control modeling module forms a power supply load regression row based on the power supply load characteristics of each pressure sampling window and the power supply coupling memory time, and constrains the power supply regression coefficient with the pressure response interval of the non-RF load event to form a power supply coupling constraint. The power supply coupling response is subtracted from the total transmitted associated response to form the radio frequency associated response interval. The radio frequency transmission regression row is formed according to the radio frequency associated coupling memory time based on the radio frequency transmission characteristics. The radio frequency regression coefficient is constrained by the radio frequency associated response interval to form the radio frequency associated coupling constraint.

7. The water conservancy data acquisition and transmission system based on narrowband Internet of Things according to claim 6, characterized in that, When the ratio of the number of regression rows to the number of regression features reaches a preset threshold, and the power supply regression coefficient and the radio frequency regression coefficient have finite upper and lower bounds, the non-RF control modeling module establishes a bounded coupling model. An independent verification event group consisting of real transmission events not involved in modeling and their events without radio frequency load is used to verify the coverage relationship between the predicted response interval and the observed response interval. Based on the coverage relationship, the applicable state and feature scope of the bounded coupling model are determined.

8. The water conservancy data acquisition and transmission system based on narrowband Internet of Things according to claim 1, characterized in that, The pressure recovery module calculates the power supply coupling response interval and the radio frequency correlation coupling response interval respectively based on the power supply coupling constraint and the radio frequency correlation coupling constraint, and adds the intervals together to form the transmission coupling feasible interval. The measurement error interval is formed by the non-radio frequency measurement residuals before and after the transmission action interval and the pressure measurement error boundary. The original pressure measurement value is then extended using the measurement error interval to form the observation pressure interval.

9. The water conservancy data acquisition and transmission system based on narrowband Internet of Things according to claim 8, characterized in that, The pressure recovery module uses the allowable pressure variation zone as the physical pressure prediction interval for the first pressure sampling moment to be recovered. It forms a natural pressure increment range based on the adjacent allowable pressure variation zones, adds the natural pressure increment range to the real pressure feasible interval of the previous pressure sampling moment, and then intersects it with the current allowable pressure variation zone to form the current physical pressure prediction interval. The intersection of the current physical pressure prediction interval and the measured pressure interval is used to form the actual pressure feasible interval at the current pressure sampling time.

10. The water conservancy data acquisition and transmission system based on narrowband Internet of Things according to claim 9, characterized in that, When the lower bound of the real pressure feasible range is continuously higher than the event reporting threshold and the confirmation time is reached, the sending control and calibration module generates a confirmed over-limit sending task. When the feasible range of real pressure crosses the event reporting threshold or no effective feasible range of real pressure is formed, an unresolved record is established; when the unresolved state reaches the preset maximum duration, or the model is inapplicable, or the upper limit of the unresolved range reaches the safety pressure threshold, or the data discontinuity state reaches the preset duration, an unresolved alarm task is generated. The remote receiving device compares the interval widths of the power supply coupling response interval and the radio frequency associated coupling response interval during the confirmation period. When the interval width of the power supply coupling response interval is not less than the interval width of the radio frequency associated coupling response interval, the model differentiation target is set as the power supply coupling constraint. The transmission control and calibration module distinguishes targets based on the model, selects target sequences, performs RF-free calibration, incorporates calibration constraints into power supply coupling constraints, and re-forms RF-related coupling constraints and bounded coupling models based on the updated power supply coupling constraints.

Images