Search

CN-122027659-A - Original data characteristic analysis system in water meter detection

CN122027659ACN 122027659 ACN122027659 ACN 122027659ACN-122027659-A

Abstract

The invention relates to the technical field of industrial data analysis and discloses an original data characteristic analysis system in water meter detection, which comprises a pulse data converging unit, a time sequence difference processing unit, a characteristic decoupling shunt unit, an inherent mode mapping unit and a transient distortion judging unit, wherein the pulse data converging unit is used for acquiring an original pulse time sequence, the time sequence difference processing unit is used for receiving the original pulse time sequence and executing first-order difference operation to generate a first-order time sequence difference value sequence, the characteristic decoupling shunt unit is used for executing logic locking by using a monitoring carrier environment base line scalar to divide the first-order time sequence into a steady-state processing channel and a transient analysis channel, the inherent mode mapping unit is used for pre-storing inherent mode characteristic vectors, and the transient distortion judging unit is used for extracting frequency domain characteristic vectors and executing orthogonal projection mapping with the inherent mode characteristic vectors to output effective transient distortion tensor.

Inventors

  • LIU WENQI
  • TANG WEI

Assignees

  • 临沂市计量检定所

Dates

Publication Date
20260512
Application Date
20260318

Claims (10)

  1. 1. A system for analyzing characteristics of raw data in water meter detection, the system comprising: the pulse data aggregation unit is used for acquiring an original pulse time sequence representing the output parameter of the metering sensing unit; the time sequence difference processing unit is used for receiving the original pulse time sequence and executing first order difference operation so as to generate a first order time sequence difference value sequence representing the change rate of the monitored parameter; The characteristic decoupling shunt unit is used for performing logic locking on the first-order time sequence difference sequence by utilizing the monitoring carrier environment baseline scalar, and dividing the first-order time sequence difference sequence into a steady-state processing channel and a transient analysis channel according to the local variation coefficient when the monitoring carrier environment baseline scalar is in a preset steady-state interval; the inherent mode mapping unit is used for pre-storing inherent mode feature vectors generated by monitoring the carrier structure logic topology dimensionality reduction translation; And the input end of the transient distortion judging unit is connected with the transient analysis channel and is used for extracting frequency domain feature vectors of data in the transient analysis channel, and performing orthogonal projection mapping on the frequency domain feature vectors and the inherent mode feature vectors so as to output an effective transient distortion tensor consisting of the projection intensity of the frequency domain feature vectors on the preset mode subset and the corresponding physical excitation frequency when the orthogonal projection components cover the preset mode subset.
  2. 2. The system according to claim 1, wherein the time-series differential processing unit includes a time stamp correction module, the time stamp correction module is configured to linearly correct a time stamp interval in the original pulse time-series sequence according to a real-time temperature drift coefficient of the crystal oscillator in the measurement sensing unit before performing the first-order differential operation, and take the corrected time stamp as a time reference of the first-order differential operation.
  3. 3. The system according to claim 1, wherein the characteristic decoupling shunt unit is connected to an environment monitoring module, the environment monitoring module is used for monitoring a static pressure average value of each node in the carrier to construct a monitoring carrier environment baseline scalar, and the characteristic decoupling shunt unit is configured to unlock logic and guide the first-order sequence of time sequence difference values into the transient analysis channel in full when a change rate of the monitoring carrier environment baseline scalar from a preset steady-state interval exceeds a trigger threshold value determined based on background noise variance.
  4. 4. The system for analyzing the characteristics of the original data in the water meter detection according to claim 1, wherein the inherent mode characteristic vector is generated by finite element mode analysis of geometric topology parameters and material stiffness parameters in a computer aided design model, the geometric topology parameters comprise elbow angles, pipe diameter mutation rates and branch topological coupling degrees, and the inherent mode characteristic vector is composed of static characteristic data structures after the normalization processing of the resonant frequency of the first five-order structure of the monitoring carrier.
  5. 5. The system according to claim 4, wherein the determination rule for the transient distortion determination unit to perform the orthogonal projection mapping is defined by: the specific implementation of the scalar mapping operation is as follows: Wherein ρ is a cosine similarity factor between the frequency domain feature vector V f and the natural mode feature vector V m , n is a dimension of the feature vector, V fi is an i-th component of the frequency domain feature vector V f , V mi is an i-th component of the natural mode feature vector V m , τ is a preset physical excitation correlation intensity threshold, and when the cosine similarity factor ρ is greater than the physical excitation correlation intensity threshold τ, it is determined that the signal in the transient analysis channel is a feature for monitoring the physical structure excitation of the carrier.
  6. 6. The system according to claim 1, wherein the transient analysis channel is provided with an asynchronous double-track buffer structure, and the asynchronous double-track buffer structure is configured to perform logical tag mapping on the original pulse timing sequence in the corresponding time window and the valid transient distortion tensor when the valid transient distortion tensor is generated.
  7. 7. The system of claim 1, wherein the steady-state processing channel is configured to perform low frequency component extraction and mean shift compensation on a first order sequence of time series differences that corresponds to a predetermined steady-state interval to generate a steady-state reference value that characterizes measurement accuracy of the meter sensing unit, and wherein the system further comprises a hierarchical encoding module configured to encode the steady-state reference value and the effective transient distortion tensor in separate logical memory domains.
  8. 8. The system for analyzing characteristics of raw data in water meter detection according to claim 1, further comprising an edge distributed monitoring unit, wherein the edge distributed monitoring unit is deployed at a plurality of industrial field nodes, and each industrial field node is configured with an independent characteristic decoupling shunt unit and an inherent mode mapping unit and is connected with a central server through a narrowband communication protocol.
  9. 9. The system of claim 1, wherein the transient distortion determination unit further comprises a spatiotemporal correction module for generating a lag correction factor based on system time consumption when the system responds to the effective transient distortion tensor, and performing a reverse hedging correction on a timestamp of the effective transient distortion tensor based on the lag correction factor.
  10. 10. The system of claim 1, further comprising a data response terminal configured to receive the effective transient distortion tensor, and output a control command for the monitoring carrier when the projection intensity of the subset of preset modes in the effective transient distortion tensor exceeds a safety precaution threshold of 15%.

Description

Original data characteristic analysis system in water meter detection Technical Field The invention relates to an original data characteristic analysis system in water meter detection, and belongs to the technical field of industrial data analysis. Background The water service monitoring system in the current urban water supply network generally adopts a water meter detection terminal to collect an original pulse sequence, a data analysis center counts the total pulse quantity according to a preset time window, the instantaneous flow rate and the accumulated flow rate of the water supply network are converted, the flow constitutes a main flow mode of the current industrial water service monitoring, the main flow metering plays a core role, as the concurrent scale of pipe network nodes is continuously expanded, the flow transient impact caused by valve opening and closing or pump station switching in the pipe network causes the original pulse sequence to generate nonlinear high-frequency jitter, the prior art generally adopts average smoothing treatment or fixed bandwidth filtering, noise is suppressed, meanwhile, the characteristic information entropy which reflects the mechanical abrasion and the fluid dynamics characteristics of the pipe network is lost, the system is difficult to prejudge the long-period operation risk of the pipe network, the sampling frequency is simply increased or the statistical window is reduced, the communication load index level of the data center is increased, and more edge node hardware thermal noise is introduced, the distortion is caused by calculation of the steady state flow, the existing industrial data analysis mode has the limitation of the information processing domain and the physical design domain cutting flow, the generated high-frequency essentially is the limitation of fluid acting on the specific physical topology, the physical topology is caused by the fluid acts on the physical topology, the physical topology is not influenced by the physical topology, the physical topology is not really different from the physical topology of the physical impulse system, the prior art is designed, and the physical topology is difficult to be ignored, and the physical impulse physical system is difficult to distinguish the physical impulse has no physical impulse algorithm, and has no physical fluctuation, and has a physical fluctuation has a physical fluctuation stage. When the hardware improvement encounters bottlenecks, algorithm optimization is difficult to be qualified for complex industrial sites, for example, chinese patent application publication No. CN114637960B discloses a signal processing method of a time-frequency domain digital turbine flow sensor, band-pass filter parameters are determined in an assisted mode through FFT frequency estimation, measurement accuracy and instantaneity are improved by combining time domain linear interpolation and zero crossing detection, the technology is established on the premise that signal characteristics are single and the ideal sinusoidal distribution is met, signal fluctuation is simplified into a mathematical signal processing problem, in an actual water meter detection working condition, high-frequency jitter is nonlinear structural resonance excited by physical structures such as pipe diameter mutation, elbow topology and the like, fluid kinetic energy acts on the physical structures, physical pipe network geometric constraint is eliminated, dimension reduction is performed by adopting a general filtering and statistical algorithm, edge hardware random electrical noise and physical transient impact cannot be distinguished, a mapping relation between signal distortion and physical structural working conditions cannot be established, and steady state flow measurement and transient physical characteristic decoupling is difficult to realize in the cross-period working condition evolution of the system due to information processing domain and physical design domain cutting flow. Therefore, how to construct a data decoupling architecture with physical perceptibility, and to guide the original time sequence to execute feature dimension reduction by introducing the physical prior vector at the bottom layer becomes the technical problem to be solved by the invention. Disclosure of Invention In order to solve the problems in the background art, the technical scheme of the invention is as follows: a raw data feature analysis system in water meter detection, the system comprising: the pulse data aggregation unit is used for acquiring an original pulse time sequence representing the output parameter of the metering sensing unit; the time sequence difference processing unit is used for receiving the original pulse time sequence and executing first order difference operation so as to generate a first order time sequence difference value sequence representing the change rate of the monitored parameter; The characteristic decoupling shunt un