A method, system, electronic device, and storage medium for dynamic real-time display and storage of pressure data.

Abstract

This application provides a method, system, electronic device, and storage medium for dynamic real-time display and storage of pressure data, relating to the field of pressure monitoring technology. This application receives multi-channel digital pressure signals, extracts their spatiotemporal correlation features to construct a three-dimensional pressure correlation matrix; divides channel groups according to fluctuation patterns and performs differentiated encoding on the real-time data stream; generates a control set using the channel dimension, parses the encoded data stream, dynamically constructs a visualization interface, and displays channel identifiers, pressure waveforms, and measured values ​​in real time; predicts future trends based on historical data in the time dimension, generating a four-dimensional intelligent index composed of timestamps, channel identifiers, measured values, and predicted values; continuously monitors storage load and fluctuation frequency, dynamically adjusts routing accordingly, stores the four-dimensional index embedded in a specific file header, and generates a text index mapping table. This enables dynamic visualization, trend prediction, and intelligent index storage of multi-channel pressure data, ensuring second-level data backtracking.

Description

Technical Field

[0001] This application relates to the field of pressure monitoring technology, and in particular to a method, system, electronic device and storage medium for dynamic real-time display and storage of pressure data. Background Technology

[0002] In the scenario of ultra-high frequency pressure monitoring of industrial equipment, the technical requirements are concentrated in three points: First, real-time dynamic visualization needs to simultaneously present multi-channel waveforms, values ​​and spatiotemporal correlations between channels; second, the storage system needs to have a high throughput writing capability of millions of data points per second and support efficient compression; third, it must realize second-level intelligent backtracking based on time, channel, measured value and predicted trend, while dynamically sensing storage load and data fluctuation frequency to optimize resource allocation.

[0003] The current mainstream solution adopts a distributed stream processing framework combined with a time-series database architecture: raw data streams are received through Kafka, Flink calculates statistical features (such as covariance) between channels in real time, and separates and stores the feature data and raw stress values ​​in relational databases and time-series databases; the front end subscribes to data through WebSocket and renders waveforms using a visualization library, and the storage layer adopts a static sharding routing strategy based on channel hashes.

[0004] However, this existing solution has significant drawbacks: First, the separate storage of feature data and raw data requires cross-database queries for dynamic visualization, which introduces high latency in high-frequency scenarios, causing interface lag and making it difficult to meet real-time interaction requirements; second, the static routing strategy cannot respond to sudden fluctuations in frequency and changes in storage load, resulting in uneven node load and a sharp drop in overall throughput or even data loss during high-pressure write periods; finally, the time-series database lacks a joint index for predicting trends, and complex condition backtracking requires minute-level scanning, which cannot meet the second-level retrieval requirements. Summary of the Invention

[0005] The purpose of this application is to provide a method, system, electronic device and storage medium for dynamic real-time display and storage of pressure data, so as to solve the problems of interface delay, throughput drop and inefficient retrieval caused by feature data separation, static routing strategy and single index mechanism in the prior art.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a method for dynamic real-time display and storage of pressure data, comprising:

[0007] The system receives multi-channel digital pressure signals and extracts the spatiotemporal correlation features between channels in real time. Based on the spatiotemporal correlation features, a three-dimensional pressure correlation matrix is ​​constructed. The three-dimensional pressure correlation matrix includes channel dimension, time dimension, and fluctuation mode dimension.

[0008] The channel group types are divided by the fluctuation pattern dimension of the three-dimensional pressure correlation matrix, and the multi-channel digital pressure signals are simultaneously converted into real-time data streams. The real-time data streams are then differentially encoded according to the channel group types.

[0009] Based on the channel dimensions in the three-dimensional pressure correlation matrix, a set of channel controls is generated. The real-time data stream after differential encoding is parsed synchronously to dynamically construct a visual interactive display interface. The visual interactive display interface displays the channel identifiers corresponding to the control set, the pressure waveform restored by the real-time data stream, and the actual measured values ​​corresponding to the pressure waveform in real time.

[0010] Based on the historical pressure sequence stored in the time dimension of the three-dimensional pressure correlation matrix, the future pressure change trend is predicted, and a four-dimensional intelligent index is generated by combining the future pressure change trend. The four-dimensional intelligent index consists of timestamp, channel identifier, actual measured value and predicted trend value.

[0011] The system continuously monitors the real-time load status of the storage medium, performs dynamic routing decisions based on the real-time load status and the real-time fluctuation frequency in the fluctuation mode dimension, simultaneously embeds the four-dimensional intelligent index into a specific structured file header, and generates a text-formatted index mapping table based on the four-dimensional intelligent index to achieve second-level backtracking of stored data.

[0012] Optionally, the continuous monitoring of the real-time load status of the storage medium, the execution of dynamic routing decisions based on the real-time load status and the real-time fluctuation frequency in the fluctuation pattern dimension, the simultaneous embedding of the four-dimensional intelligent index into a specific structured file header, and the generation of a text-formatted index mapping table based on the four-dimensional intelligent index to achieve second-level backtracking of stored data, includes:

[0013] The queue length of the data to be written to the storage medium is continuously monitored, and the ratio of the remaining storage space is marked as a high load state when the queue length exceeds a set threshold or the ratio of the remaining storage space is lower than a critical ratio.

[0014] The real-time fluctuation frequency of each channel is calculated based on the repetitive fluctuation patterns in the fluctuation pattern dimension.

[0015] Under the high load condition, when the real-time fluctuation frequency of the channel exceeds the frequency threshold, the corresponding real-time data stream is marked as a high-frequency fluctuation data stream and is preferentially written to the memory image block. When the real-time fluctuation frequency of the channel is lower than the frequency threshold, the corresponding real-time data stream is marked as a low-frequency trend data stream and is directly written to the solid-state storage chip.

[0016] Under non-high load conditions, write the real-time data streams corresponding to all channels to the solid-state storage chip;

[0017] The four-dimensional intelligent index is converted into a binary format and embedded into a specific structured file header of the data file. Each index unit in the four-dimensional intelligent index is converted into a text entry in the order of timestamp, channel identifier, actual measurement value, and predicted trend value. All text entries are aggregated to form a text format index mapping table for independent storage, enabling second-level backtracking of stored data.

[0018] Optionally, the step of classifying channel group types based on the fluctuation pattern dimension of the three-dimensional pressure correlation matrix and simultaneously converting the multi-channel digital pressure signal into a real-time data stream includes:

[0019] Based on the repetitive fluctuation patterns stored in the fluctuation pattern dimension of the three-dimensional pressure correlation matrix, the synchronicity intensity of any two channel fluctuation patterns is calculated.

[0020] Based on the synchronization intensity, the channel group types are divided into three categories: channels with synchronization intensity significantly higher than the average synchronization threshold are classified as strongly correlated channels; channels with synchronization intensity close to the average synchronization threshold are classified as weakly correlated channels; and channels with synchronization intensity significantly lower than the average synchronization threshold are classified as independent channels.

[0021] The digital pressure signal of each channel is converted into a continuous binary data block in time sequence, and the binary data blocks of all channels are spliced ​​together according to the acquisition time sequence to form a real-time data stream.

[0022] Optionally, the step of generating a channel control set based on the channel dimensions in the three-dimensional pressure correlation matrix, and simultaneously parsing the differentiated encoded real-time data stream, dynamically constructs a visual interactive display interface. This interface displays in real-time the channel identifiers corresponding to the control set, the pressure waveform reconstructed from the real-time data stream, and the actual measured values ​​corresponding to the pressure waveform, including:

[0023] Extract channel identifiers and their corresponding physical layout order from the channel dimensions of the three-dimensional pressure correlation matrix to generate a selectable set of channel controls, with each control bound to a unique channel identifier;

[0024] Based on the coding rules of the channel group type, the real-time data stream after differential coding is parsed, the direct pressure value restoration is performed on the main channel of the strongly correlated channel group, the residual superposition restoration is performed on the slave channel, the waveform reconstruction is performed on the weakly correlated channel group, and the entropy optimization data recovery is performed on the independent channel.

[0025] Create a dynamic visual interactive display interface. When the user selects the control corresponding to the channel identifier, the waveform display area is activated in the visual interactive display interface. The analyzed pressure waveform is drawn in real time in the waveform display area, and the actual measured value of the latest pressure value is dynamically marked.

[0026] When the user cancels the selection of the control corresponding to the channel identifier, the waveform display area is closed and the parsing process of the corresponding channel is terminated.

[0027] Optionally, based on the historical pressure sequence stored in the time dimension of the three-dimensional pressure correlation matrix, the future pressure change trend is predicted, and a four-dimensional intelligent index is generated in combination with the future pressure change trend. The four-dimensional intelligent index consists of a timestamp, a channel identifier, actual measured values, and predicted trend values, including:

[0028] Historical pressure sequences of continuous time sections are extracted from the time dimension of the three-dimensional pressure correlation matrix. The historical pressure sequences contain measured pressure values ​​of each channel at multiple time sections.

[0029] Based on the historical pressure sequence, the pressure change factor of adjacent time sections is calculated, and the predicted pressure value of the next time section is extrapolated by the pressure change factor to predict the future pressure change trend.

[0030] A four-dimensional intelligent index unit is generated by combining the future pressure change trend. The four-dimensional intelligent index unit includes the timestamp of the current time section, the channel identifier of the corresponding channel, the measured pressure value of the current time section, and the predicted pressure value of the next time section.

[0031] The four-dimensional intelligent index units of all channels are aggregated and stored in ascending order of timestamp to form a four-dimensional intelligent index set.

[0032] Optionally, the step of differentially encoding the real-time data stream according to the channel group type includes:

[0033] For strongly correlated channel groups, master-slave channel collaborative coding is implemented. A master channel is designated within the strongly correlated channel group, and the original pressure value of the master channel in the real-time data stream is saved. The slave channels only save the difference between their own pressure value and the pressure value of the master channel at the same time.

[0034] For weakly correlated channel groups, trend feature preservation and compression are implemented, recording only the direction and magnitude of pressure value changes at adjacent time points in the real-time data stream, and deleting the basic value portion of the pressure values ​​that has not changed.

[0035] For independent channels, entropy optimization compression is performed. The same pressure value that is repeated continuously in the real-time data stream is detected. Only the first pressure value is retained and the start time and number of repetitions of the pressure value are recorded.

[0036] Optionally, the step of receiving multi-channel digital pressure signals and extracting spatiotemporal correlation features between channels in real time, and constructing a three-dimensional pressure correlation matrix based on the spatiotemporal correlation features, wherein the three-dimensional pressure correlation matrix includes channel dimension, time dimension, and fluctuation mode dimension, including:

[0037] Simultaneously receive multi-channel digital pressure signals, and perform equal-interval slicing operations on the digital pressure signals of each channel under a unified time reference to form time sections aligned in the time dimension.

[0038] At each time segment, channels are paired up into channel pairs, the signal change correlation in each channel pair is calculated, the spatiotemporal correlation features of the time segment are generated, and the spatiotemporal correlation feature values ​​of all channel pairs constitute the spatiotemporal correlation feature set of the time segment.

[0039] The spatiotemporal correlation feature sets of multiple consecutive time sections are connected end to end in chronological order to form a channel spatiotemporal correlation feature sequence across time sections;

[0040] A three-dimensional pressure correlation matrix is ​​constructed based on the spatiotemporal correlation feature sequence. In the three-dimensional pressure correlation matrix, the channel identifiers of all channels are arranged in the channel dimension, the sequential number of time sections is arranged in the time dimension, and the fluctuation pattern dimension stores the repetitive fluctuation patterns identified in the spatiotemporal correlation feature sequence. The repetitive fluctuation patterns are manifested as similar waveform segments that appear at fixed intervals in the spatiotemporal correlation feature sequence.

[0041] Secondly, this application provides a dynamic real-time display and storage system for pressure data, comprising:

[0042] The extraction module is used to receive multi-channel digital pressure signals and extract the spatiotemporal correlation features between channels in real time. Based on the spatiotemporal correlation features, a three-dimensional pressure correlation matrix is ​​constructed. The three-dimensional pressure correlation matrix includes channel dimension, time dimension and fluctuation mode dimension.

[0043] The encoding module is used to classify the channel group type according to the fluctuation pattern dimension of the three-dimensional pressure correlation matrix, synchronously convert the multi-channel digital pressure signal into a real-time data stream, and perform differentiated encoding on the real-time data stream according to the channel group type.

[0044] The display module is used to generate a set of channel controls based on the channel dimensions in the three-dimensional pressure correlation matrix, and synchronously parse the real-time data stream after differential encoding to dynamically construct a visual interactive display interface. The visual interactive display interface displays the channel identifiers corresponding to the control set, the pressure waveform restored by the real-time data stream, and the actual measured values ​​corresponding to the pressure waveform in real time.

[0045] The generation module is used to predict future pressure change trends based on the historical pressure sequences stored in the time dimension of the three-dimensional pressure correlation matrix, and generate a four-dimensional intelligent index in combination with the future pressure change trends. The four-dimensional intelligent index consists of timestamps, channel identifiers, actual measured values ​​and predicted trend values.

[0046] The storage module is used to continuously monitor the real-time load status of the storage medium, perform dynamic routing decisions based on the real-time load status and the real-time fluctuation frequency in the fluctuation mode dimension, simultaneously embed the four-dimensional intelligent index into a specific structured file header, and generate a text-formatted index mapping table based on the four-dimensional intelligent index to achieve second-level backtracking of stored data.

[0047] Thirdly, this application provides an electronic device, comprising:

[0048] Memory, used to store computer programs;

[0049] A processor is configured to execute the computer program to implement the steps of the dynamic real-time display and storage method for pressure data as described in the first aspect above.

[0050] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the dynamic real-time display and storage method for pressure data as described in the first aspect above.

[0051] This application provides a method for dynamic real-time display and storage of pressure data. It receives multi-channel digital pressure signals and extracts spatiotemporal correlation features in real time to construct a three-dimensional pressure correlation matrix (channel / time / fluctuation pattern). Based on the fluctuation pattern dimension, it classifies channel groups and differentially encodes real-time data streams. Simultaneously, it generates a control set using the channel dimension, synchronously parses the encoded data stream to dynamically construct a visualization interface, displaying channel identifiers, reconstructed pressure waveforms, and actual measured values ​​in real time. Furthermore, it predicts future trends based on historical data in the time dimension, generating a four-dimensional intelligent index containing timestamps, channel identifiers, measured values, and predicted values. Finally, by continuously monitoring storage load and fluctuation frequency, it performs dynamic routing decisions, stores the four-dimensional index embedded in the structured file header, and generates a text index mapping table. This simultaneously achieves millisecond-level dynamic visualization of multi-channel pressure waveforms, storage throughput optimization in sudden fluctuation scenarios, and second-level data backtracking capabilities based on the four-dimensional composite index. Attached Figure Description

[0052] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 A flowchart illustrating a method for dynamic real-time display and storage of pressure data provided in an embodiment of this application;

[0054] Figure 2 A flowchart illustrating a method for dynamic real-time display and storage of pressure data provided in this application embodiment;

[0055] Figure 3 A scenario diagram illustrating a method for dynamic real-time display and storage of pressure data provided in an embodiment of this application;

[0056] Figure 4 This is a schematic diagram of the structure of a dynamic real-time display and storage system for pressure data provided in an embodiment of this application. Detailed Implementation

[0057] Research indicates that in the field of ultra-high frequency multi-channel stress monitoring, existing technologies face three core contradictions: the separation of time-series databases and relational databases leads to a disconnect between spatiotemporal correlation characteristics and raw data, resulting in sluggish dynamic visualization interface responses; static routing strategies cannot adapt to sudden data fluctuations and changes in storage node load, causing unstable cluster write performance; and the lack of joint support for stress trends in a single indexing mechanism results in low efficiency for backtracking complex data conditions, making it difficult to meet real-time decision-making needs.

[0058] To address the aforementioned issues, this invention proposes a method for dynamic real-time display and storage of pressure data. This method constructs a three-dimensional pressure correlation matrix that integrates channel, time, and fluctuation patterns. Channel groups are divided according to the fluctuation pattern dimension, and data streams are differentially encoded. Visual controls are dynamically generated based on the channel dimension, enabling real-time synchronous rendering of multi-channel waveforms, measurements, and correlation features. A four-dimensional intelligent index is generated based on historical data prediction trends in the time dimension, and routing paths are dynamically optimized according to storage load and fluctuation frequency. Finally, the index embedded in the structured file header and the text mapping table are stored collaboratively. This solution systematically solves the shortcomings of existing technologies: the three-dimensional matrix eliminates the problem of feature data separation, ensuring real-time visualization; the dynamic routing strategy significantly improves storage stability under sudden events; and the four-dimensional joint index mechanism overcomes complex retrieval bottlenecks, achieving second-level backtracking capability, providing a complete technical closed loop for ultra-high frequency pressure monitoring scenarios.

[0059] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0060] The core of this application is to provide a method for dynamic real-time display and storage of pressure data, and a flowchart of one specific implementation is shown below. Figure 1As shown, the method includes:

[0061] S101. Receive multi-channel digital pressure signals and extract the spatiotemporal correlation features between channels in real time. Construct a three-dimensional pressure correlation matrix based on the spatiotemporal correlation features. The three-dimensional pressure correlation matrix includes channel dimension, time dimension and fluctuation mode dimension.

[0062] Optionally, step S101 may specifically include the following steps:

[0063] S1011. Synchronously receive multi-channel digital pressure signals, and perform equal-interval slicing operations on the digital pressure signals of each channel under a unified time reference to form time sections aligned in the time dimension.

[0064] S1012. At each time segment, pair the channels into two groups, calculate the signal change correlation in each group of channels, generate the spatiotemporal correlation features of the time segment, and construct the spatiotemporal correlation feature set of the time segment by combining the spatiotemporal correlation feature values ​​of all the channel groups.

[0065] S1013. Connect the spatiotemporal correlation feature sets of multiple consecutive time sections in chronological order to form a channel spatiotemporal correlation feature sequence across time sections.

[0066] S1014. Construct a three-dimensional pressure correlation matrix based on the spatiotemporal correlation feature sequence. In the three-dimensional pressure correlation matrix, the channel identifiers of all channels are arranged in the channel dimension, the sequential number of time sections is arranged in the time dimension, and the repetitive fluctuation patterns identified in the spatiotemporal correlation feature sequence are stored in the fluctuation pattern dimension. The repetitive fluctuation patterns are manifested as similar waveform segments that appear at fixed intervals in the spatiotemporal correlation feature sequence.

[0067] In the above scheme, the multi-channel digital pressure signal refers to a sequence of dynamic pressure measurements synchronously acquired from multiple physical locations. It includes the time synchronization accuracy requirements and spatial distribution information between channels, used to characterize the real-time state of pressure transmission within the equipment. Spatiotemporal correlation characteristics reflect the linkage patterns of different channel pressure signals in terms of temporal evolution and spatial location. This includes the quantified correlation strength values ​​(such as covariance) of channel pairs and the evolution patterns across time sections, used to reveal the propagation path and coupling effects of pressure waves within the equipment. The three-dimensional pressure correlation matrix is ​​a three-dimensional data structure integrating channel relationships, temporal evolution, and fluctuation patterns. The channel dimension stores physical location identifiers, the time dimension records the sequence number of events, and the fluctuation pattern dimension encodes the recurring signal change characteristics, used to construct a dynamic behavior model of the pressure system.

[0068] In this embodiment, step S1011 first establishes a unified time reference for all channel pressure signals using a high-precision clock synchronization module such as the PTP protocol. After receiving multi-channel digital pressure signals, a slicing operation is performed according to a preset fixed time interval: the continuous pressure signal stream of each channel is cut into segments of equal duration and marked with time segment numbers (such as T001, T002). This process ensures that data from different channels at the same physical moment are assigned to the same time segment, forming spatiotemporally aligned data blocks. For example, at 10:00:00.000, the C1 channel of a hydraulic press unit (5 pressure sensors C1-C5) is sliced ​​into segments [15.2, 15.3, 15.1], C2 into [22.4, 22.6, 22.5]...C5 into [19.0, 19.2, 19.1], and all data belong to time segment T001.

[0069] Secondly, in step S1012, channels are paired up for each time segment. For each channel pair (e.g., C1-C2), the correlation of signal changes is calculated: the covariance algorithm is used to process all sampling points within the slice, as shown in the following formula: ,in , The value of the i-th sampling point in both channels. , For the corresponding mean, , The standard deviation n corresponds to the number of sampling points within the slice. Finally, a set of spatiotemporal correlation features for this section is generated based on the calculation results. For example, in section T001, the correlation between channels C1 and C2 is calculated to be 0.85, indicating a strong positive correlation, suggesting that the pressure changes in the two channels are highly synchronized; while the correlation between C1 and C3 is only 0.32, indicating a weak correlation, suggesting that the two have weak linkage.

[0070] Next, in step S1013, the spatiotemporal correlation feature sets of consecutive time sections are connected in chronological order to form the spatiotemporal correlation feature sequence of the channel pair. This sequence essentially describes the dynamic curve of the correlation strength between the two channels evolving over time. For example, the C1-C2 correlation quantities (0.85→0.89→...→0.12) of 100 sections from T001 to T100 are concatenated in chronological order to form the spatiotemporal correlation feature sequence of the C1-C2 channel pair.

[0071] Finally, in step S1014, a pattern scan is performed on the spatiotemporal correlation feature sequence: a sliding window peak detection algorithm is used to identify recurring similar waveform segments in the sequence. When the same waveform is detected three or more times consecutively, its period length, amplitude change slope, and waveform similarity are extracted as feature parameters to generate a unique pattern label. Next, the recognition results are stored in a structured manner: all channel identifiers are arranged in physical location order as channel dimensions, such as C1, C2, C3; time section numbers are stored in the order of event occurrence as time dimensions, such as T001 to T100; pattern labels and their trigger times are recorded in key-value pairs, such as P1: [T005, T010, T015], and associated with feature parameters, such as P1 period = 5 sections, slope = -2.8 / section. The final constructed three-dimensional pressure correlation matrix forms a three-dimensional coordinate model. The X-axis positions the sensor identifier index, such as channel C1 corresponding to position 0; the Y-axis positions the section sequence index, such as T001 corresponding to row 0; and the Z-axis stores the pattern label mapping table, such as P1 mapped to storage address 0x0A01. For example, in the monitoring of hydraulic units, the C1-C2 channel associated sequence repeatedly showed a sudden drop waveform of "0.8→0.1" at time points T003, T008, and T013, with a corresponding amplitude difference >0.7 and an interval of 5 sections. After the system identified that its similarity was >95%, it marked it as pattern P1. The channel identifier set [C1,C2,C3], the time series [T001-T100], and the P1 parameters {trigger point list, period value, slope value} were integrated and stored in a three-dimensional matrix to provide a structured fluctuation feature library for subsequent dynamic coding.

[0072] In practical applications, in a fluid pulsation monitoring system, the system synchronously receives water depth and pressure signals collected at a sampling rate of 1000Hz from three ultra-high precision pressure modules installed at points A, B, and C on the pipeline. Time slice operations are performed at 100ms intervals. At the T=12:00:00.000 section, the pressure values ​​are: point A [10.2MPa, 10.3MPa], point B [10.5MPa, 10.6MPa], and point C [9.8MPa, 9.9MPa]. The correlation between each pair of channels at the current section is calculated: the covariance of the AB channel is 0.95 (strong positive correlation, reflecting synchronous pulsation of water flow at adjacent points), and the AC channel is 0.12 (weak correlation, because point C is located in the vortex region). The AB correlation sequence of 60 consecutive sections within a 6-second period shows a pattern: every 10 sections, a sudden change in the correlation value from 0.9 to 0.3 occurs, corresponding to the pump start-up and shutdown impact. The system identifies this repetitive waveform using a dynamic time warping algorithm, labels it as "pump vibration mode P1", and constructs a three-dimensional matrix: the channel dimension stores [A,B,C], the time dimension records T001-T060, and the fluctuation mode dimension stores P1={trigger point:T010 / T020 / T030, period:10, amplitude difference threshold:0.6}, providing a basis for fluctuation characteristics for subsequent water turbine pressure control.

[0073] The overall scheme of S101 described above transforms the original dispersed multi-channel signals into a three-dimensional matrix containing spatiotemporal linkage patterns through time-aligned slicing and channel correlation calculation; its fluctuation mode dimension automatically identifies repetitive features, providing a basis for subsequent channel grouping and dynamic coding, and significantly improving the system's ability to analyze complex pressure correlation scenarios.

[0074] S102. Divide the channel group type according to the fluctuation mode dimension of the three-dimensional pressure correlation matrix, and simultaneously convert the multi-channel digital pressure signal into a real-time data stream. Then, perform differentiated encoding on the real-time data stream according to the channel group type.

[0075] Optionally, step S102 may specifically include the following steps:

[0076] S1021. Based on the repetitive fluctuation patterns stored in the fluctuation pattern dimension of the three-dimensional pressure correlation matrix, calculate the synchronicity intensity of any two channel fluctuation patterns.

[0077] S1022. Based on the synchronization intensity, classify the channel group types, classify the channels with synchronization intensity significantly higher than the average synchronization threshold as strongly correlated channel groups, classify the channels with synchronization intensity close to the average synchronization threshold as weakly correlated channel groups, and classify the channels with synchronization intensity significantly lower than the average synchronization threshold as independent channels.

[0078] S1023. Convert the digital pressure signal of each channel into a continuous binary data block in time sequence, and splice the binary data blocks of all channels into a real-time data stream according to the acquisition time sequence.

[0079] S1024. Implement master-slave channel collaborative coding for strongly correlated channel groups. Designate a master channel within the strongly correlated channel group and save the original pressure value of the master channel in the real-time data stream. The slave channels only save the difference between their respective pressure values ​​and the pressure values ​​of the master channel at the same time.

[0080] S1025. For weakly correlated channel groups, implement trend feature preservation and compression, only record the direction and magnitude of the change in pressure values ​​at adjacent time points in the real-time data stream, and delete the basic value part of the pressure values ​​that has not changed.

[0081] S1026. For independent channels, perform entropy optimization compression, detect the same pressure value that is repeated continuously in the real-time data stream, retain only the first pressure value, and record the start time and number of repetitions of the pressure value.

[0082] In the above scheme, synchronicity strength refers to an indicator that quantifies the degree of linkage between fluctuation modes in channels. It includes characteristics such as mode period consistency, waveform overlap, and similarity of evolution trends, and is used to objectively classify the level of coordination between channels. A strongly correlated channel group refers to a set of channels with highly coordinated fluctuation modes, whose signal changes exhibit master-slave following characteristics. The complete data of the slave channel can be restored using the master channel's reference value. Master-slave collaborative encoding refers to a compression method for strongly correlated channel groups, including the storage of the master channel's original value and a slave channel difference recording mechanism, significantly reducing data volume by utilizing signal redundancy between channels. Trend feature preservation compression refers to an optimization strategy for weakly correlated channel groups, including the extraction of key features such as the direction and magnitude of signal changes, used to preserve dynamic evolution patterns within limited storage space. Entropy optimization compression refers to an efficient encoding technique for processing independent channels, including continuous repeating value detection and run-length counting mechanisms, improving storage efficiency by eliminating steady-state data redundancy.

[0083] In this embodiment, step S1021 first extracts the repetitive fluctuation characteristic parameters of each channel from the fluctuation pattern dimension of the three-dimensional pressure correlation matrix, including the fluctuation period length, waveform morphology sequence, and trigger time point. For any two channels (e.g., channels C1 and C2), a dynamic time warping (DTW) algorithm is used for waveform alignment: the fluctuation waveforms of the two channels are treated as time series, and the path with the minimum cumulative distance is found through dynamic programming. This distance value directly reflects the waveform similarity. Simultaneously, the period matching coefficient, i.e., the ratio of the fluctuation period lengths of the two channels, is calculated. When the periods are completely identical, the coefficient is 1.0. Finally, the synchronization strength is synthesized according to the formula: For example, channel C1 exhibits a "sudden rise-gradual fall" pattern with a period of 5 time segments, corresponding to the waveform sequence [0.8, 0.9, 0.2, 0.3, 0.7]. Channel C2 exhibits a similar periodic pattern, corresponding to the waveform sequence [0.75, 0.88, 0.18, 0.28, 0.68]. The DTW distance is calculated to be 0.05, and the period matching coefficient is 1.0. .

[0084] Secondly, step S1022 calculates the global average synchronization strength based on the synchronization strength calculation results of all channel pairs. A dynamic threshold classification rule is set: channel pairs with a synchronization strength ≥ 0.8 are classified as strongly correlated channel groups, indicating highly coordinated fluctuations, such as C1-C2 with a strength of 0.952; synchronization strengths between 0.4 and 0.8 are classified as weakly correlated channel groups, indicating partially coordinated fluctuations, such as C1-C3 with a strength of 0.58; channels with a synchronization strength < 0.4 are considered independent channels, indicating no significant correlation in fluctuations, such as C4 with strengths < 0.3 for all other channels. Finally, a channel group mapping table is generated: for example, C1-C2 is bound to a strongly correlated group and C1 is designated as the main channel, C1-C3 is included in the weakly correlated group, and C4 is marked as an independent channel.

[0085] Next, in step S1023, the raw pressure signal stream of each channel is converted into continuous binary data blocks in chronological order: First, the floating-point pressure values ​​are converted into IEEE 754 standard 32-bit binary format, and the data block of each time segment contains the binary concatenation of all sampling points within that segment. Then, strictly following the acquisition time order, the binary blocks of different channels are concatenated end to end to form a global real-time data stream. For example, in the T001 time segment: the binary blocks of channel C1 (12.3 → block A), C2 (18.6 → block B), C3 (15.1 → block C), and C4 (20.0 → block D) are concatenated in sequence to form a continuous binary stream of "block A + block B + block C + block D".

[0086] Finally, the real-time data stream is encoded according to the channel group type through steps S1024, S1025, and S1026. For strongly correlated groups, the system implements master-slave collaborative compression: the master channel is selected, and its original binary data block is completely preserved; for the slave channel, the numerical difference between it and the master channel at the same time segment is calculated, using the following formula: The difference is converted into binary for storage. During decoding, simply adding the difference to the main channel value restores the original value of the secondary channel. For example, in section T001: C1 value is 12.3 MPa, C2 value is 18.6 MPa. Selecting C1 as the main channel and preserving the original binary data block of C1, with C2 as the secondary channel, the difference between the secondary channel C2 and the main channel C1 is +6.3 MPa, corresponding to the binary values ​​0100000110010011 00110011 00110011. For weakly correlated channels, the system uses trend feature compression: comparing pressure values ​​at adjacent time sections and calculating the change. And extract the direction of change. The time corresponds to "↑", The time corresponds to "↓"; only the sign and the absolute value of the change are encoded in binary, and the unchanged base value is deleted. During restoration, the change must be accumulated based on the previous value. For example, if C3 changes by +0.2 MPa from T001 to T002, it is encoded as "↑0.2", corresponding to binary 0000000100000110 00000000 00000000. The first 16 bits indicate the direction and magnitude, and the last 16 bits are left blank. For independent channels, entropy optimization compression is performed: the pressure values ​​of continuous time sections are scanned, and the number of consecutive repetitions of the same value is detected. When the number of repetitions is ≥2, only the complete value of the first occurrence and the number of repetitions are retained, and intermediate repetitive data blocks are deleted. During restoration, the consecutive repetitive values ​​are reconstructed based on the start time and the number of repetitions. For example, the value of C4 in T001-T003 is 20.0 MPa, which is encoded as "start time T001|value 20.0|repetition count 3", corresponding to binary 10100000 00000011 00000000 00000000, with the first 16 bits recording the value and the last 16 bits recording the count.

[0087] In practical applications, during turbine fluid pulsation monitoring, the system receives signals from three ultra-high precision pressure sensors: C1 at the volute inlet, C2 at the guide vane outlet, and C3 at the draft tube, with a sampling rate of 1000Hz. First, the synchronization intensity is calculated based on a three-dimensional matrix wave pattern: both C1 and C2 exhibit pulsating waveforms with a period of 0.1 seconds (DTW distance 0.08, period ratio 1.0 → intensity 0.93); the waveforms of C1 and C3 show significant differences (DTW distance 0.75, period ratio 0.8 → intensity 0.32). Next, channel groups are defined: C1-C2 with an intensity of 0.93 > 0.8 are grouped into a strongly correlated group, with C1 designated as the main channel, and C3 with an intensity of 0.32 < 0.4 classified as an independent channel. The original signal was then converted into a real-time data stream: T=1.000 second cross section, C1 value 1520.3kPa → IEEE754 binary "01000100 1011110000110011 00110011", C2 value 1538.6kPa → "01000100 11000000 10011001 10011010", C3 value 101.0kPa → "01000010 11001010 00000000 00000000". Then, encoding is performed according to the channel group type: Strongly correlated group encoding is performed on C1 and C2: the difference between C2 and C1 + 18.3 kPa → converted to binary "01000011 1001001001100110 01100110". Independent channel encoding is performed on C3: if three consecutive cross-sectional values ​​of C3 are all 101.0 kPa → encoded as "101.0@T001×3", corresponding to binary "01000010 11001010 00000000 00000011". The first 16 bits are the value, and the last 16 bits are the repetition count. The final output is a mixed encoded stream: the complete value of the main channel C1 + the difference of the secondary channel C2 + the C3 repetition compression mark.

[0088] The overall solution in S102 described above precisely divides channel groups based on the synchronicity of fluctuation patterns and implements targeted differentiated coding strategies: strongly correlated groups significantly reduce data volume by utilizing inter-channel redundancy, weakly correlated groups retain key trend characteristics to reduce storage requirements, and independent channels eliminate redundant duplicate values. This significantly improves overall data compression efficiency while ensuring that critical information is lossless and recoverable.

[0089] S103. Generate a channel control set based on the channel dimension in the three-dimensional pressure correlation matrix, and simultaneously parse the real-time data stream after differential encoding to dynamically construct a visual interactive display interface. The visual interactive display interface displays the channel identifier corresponding to the control set, the pressure waveform restored by the real-time data stream, and the actual measured value corresponding to the pressure waveform in real time.

[0090] Optionally, step S103 may specifically include the following steps:

[0091] S1031. Extract channel identifiers and the physical layout order corresponding to the channel identifiers from the channel dimension of the three-dimensional pressure correlation matrix, generate a selectable set of channel controls, and bind a unique channel identifier to each control.

[0092] S1032. Parse the real-time data stream after differential encoding according to the encoding rules of the channel group type, perform direct pressure value restoration on the main channel of the strongly correlated channel group, perform residual superposition restoration on the slave channel, perform waveform reconstruction on the weakly correlated channel group, and perform entropy optimization data recovery on the independent channel.

[0093] S1033. Create a dynamic visual interactive display interface. When the user selects the control corresponding to the channel identifier, the waveform display area is activated in the visual interactive display interface. The analyzed pressure waveform is drawn in real time in the waveform display area, and the actual measured value of the latest pressure value is dynamically marked.

[0094] S1034. When the user cancels the control corresponding to the channel identifier, the waveform display area is closed and the parsing process of the corresponding channel is terminated.

[0095] In the above scheme, the channel control set refers to an interactive button group generated according to the physical layout order of the sensors, containing channel identifiers and unique binding relationships arranged according to the flow path, used to activate the visualization of pressure data at specific locations. Residual superposition restoration refers to a data reconstruction technique for strongly correlated channel groups, including loading the main channel reference value and dynamic difference superposition calculation, used to accurately restore the transient values ​​of subordinate nodes during the propagation of fluid pressure waves. Waveform reconstruction refers to a data recovery method for weakly correlated channel groups, including base value initialization and adjacent change accumulation mechanisms, used to reproduce the gradual trend characteristics of pressure pulsation attenuation regions. Entropy-optimized data recovery refers to an efficient decoding technique for processing independent channels, including initial value loading and repetition count parsing processes, used to reconstruct the steady-state reference pressure.

[0096] In this embodiment, step S1031 first reads the list of channel identifiers and their physical installation position order in the channel dimension of the three-dimensional pressure correlation matrix. Based on this order, a set of interactive controls for the visual interface is generated: a grid layout algorithm is used to automatically arrange the button positions according to the spatial coordinates of the actual equipment structure diagram. Each button is bound to a unique channel ID and labeled with an abbreviation of its physical location. For example, in a water turbine pressure monitoring scenario, three horizontally arranged circular buttons with a diameter of 30px are generated according to the physical distances of 0m (channel P1) → 3.2m (channel P2) → 6.5m (channel P3), with the spacing proportional to the actual distance.

[0097] Secondly, in step S1032, when the user activates a control, the system immediately starts real-time data stream parsing for the corresponding channel: For strongly correlated groups, the current value of the main channel is loaded, and the stored difference is superimposed to obtain the true value. For example, the binary difference code "01000001 10010011" of the channel is parsed as +18.3kPa, which is superimposed with the main channel value of 1520.3kPa to obtain 1538.6kPa; For weakly correlated groups, a new value is obtained based on the cumulative change of the previous value. For example, if the value at time T001 is 1520.3kPa, and the change at time T002 is ↑0.2kPa, then a new value of 1520.5kPa is obtained; For independent channels, the repeated markers are parsed, and constant values ​​are automatically filled in the time interval. For example, if the corresponding value of a certain channel is 101.0kPa×3, then a constant value is automatically filled in the time interval of T001-T003.

[0098] Next, after the control is activated via step S1033, a dynamic canvas is created in a designated area of ​​the interface, such as a 300×200px rectangle below the control. First, coordinates are initialized with the X-axis representing time (0-5 seconds) and the Y-axis representing pressure value (1400-1600 kPa). Then, the resolution value is acquired every 20 ms, and a Bézier curve is used to connect adjacent points to form a smooth waveform for real-time plotting. Simultaneously, data markers are displayed floating at the top of the canvas to show the latest value. For example, after the P1 channel is activated, 50 data points are plotted per second, forming a continuous fluctuating curve, with the latest value "1520.5 kPa" refreshed in real-time.

[0099] Finally, in step S1034, when the user cancels the control and selects the close button: the data stream is truncated to terminate the data parsing thread of that channel; the corresponding canvas object is destroyed and GPU memory is reclaimed to release rendering resources; cache cleanup is performed simultaneously to clear unprocessed data packets for that channel; then load balancing is performed, dynamically allocating computing resources to other active channels. For example, when the P3 channel is closed, the system immediately releases 3.2MB of video memory and transfers 15% of the CPU resources originally used for that channel to the data processing of the wormhole channel.

[0100] In practical applications, in a wellhead pressure monitoring system, the system generates a set of controls based on a three-dimensional matrix channel dimension, including wellhead sensor P1, mid-well section sensor P2, and standard reference module P3: buttons are arranged vertically according to depth ["Wellhead-300m", "Mid-well-1500m", "Standard Pressure Module"]. When the engineer selects the "Mid-well-1500m" control, the system parses the real-time data stream: this channel belongs to the strongly correlated group slave channel (the main channel is wellhead P1), loads the current wellhead value of 35.2MPa, superimposes the +12.8MPa parsed from the binary difference code "01000010 10010000", and restores the actual pressure in the well to 48.0MPa; at the same time, the waveform display area below the control is activated, and a fluid pulsation curve from 48.0MPa to 48.3MPa is plotted with a time window of 1 second, updating a data point every 20ms, and floating the latest value "T=12:05:03.020: 48.2MPa". When the "Standard Pressure Module" control is selected, the system performs entropy optimization recovery on the independent channel: parsing the binary marker "10100000 00000101" to correspond to 20.0 MPa is repeated 5 times, generating a horizontal straight line waveform on the interface and labeling it "Constant Pressure 20.0 MPa". After closing the "Wellhead-300m" control, the system immediately releases the 3D rendering resources for that channel and terminates the decoding thread.

[0101] The overall solution of S103 described above generates intuitive controls through physical layout sorting and achieves millisecond-level waveform restoration by combining differentiated data parsing strategies; the user interactive selection mechanism optimizes resource allocation and ensures the real-time visualization accuracy of key pressure data in high-concurrency scenarios, meeting the needs of fluid machinery pulsation monitoring and high-precision testing.

[0102] S104. Based on the historical pressure sequence stored in the time dimension of the three-dimensional pressure correlation matrix, predict the future pressure change trend, and generate a four-dimensional intelligent index in combination with the future pressure change trend. The four-dimensional intelligent index consists of timestamp, channel identifier, actual measured value and predicted trend value.

[0103] Optionally, step S104 may specifically include the following steps:

[0104] S1041. Extract the historical pressure sequence of continuous time sections from the time dimension of the three-dimensional pressure correlation matrix. The historical pressure sequence includes the measured pressure values ​​of each channel at multiple time sections.

[0105] S1042. Calculate the pressure change factor for adjacent time sections based on the historical pressure sequence, and extrapolate the predicted pressure value for the next time section using the pressure change factor to predict the future pressure change trend.

[0106] S1043. A four-dimensional intelligent index unit is generated by combining the future pressure change trend. The four-dimensional intelligent index unit includes the timestamp of the current time section, the channel identifier of the corresponding channel, the measured pressure value of the current time section, and the predicted pressure value of the next time section.

[0107] S1044. Aggregate all channels' four-dimensional intelligent index units, arrange them in ascending order of timestamps, and store them to form a four-dimensional intelligent index set.

[0108] In the above scheme, the historical pressure sequence refers to the continuous measured value records stored in the three-dimensional pressure correlation matrix in chronological order. It includes the pressure fluctuation trajectory of each channel within a fixed time interval and its spatial correlation characteristics, which can be used to analyze the dynamic evolution law of the equipment pressure system. The pressure change factor refers to the key parameter that quantifies the rate of pressure change between adjacent time sections. It includes the correlation characteristics between the pressure value change and the time interval, and can be used to extrapolate short-term pressure change trends. The four-dimensional intelligent index unit refers to an atomic data structure that integrates time reference, spatial location, measured state, and future trends: the timestamp marks the precise moment of data acquisition; the channel identifier locates the spatial location of the pressure source; the measured value records the actual pressure state at the current moment; the predicted value reflects the pressure evolution direction at the next moment; this structure supports multi-dimensional joint retrieval and can be used for early warning of abnormal equipment conditions. The four-dimensional intelligent index set refers to the cluster of index units that are aggregated in an orderly manner according to the time axis. It includes cross-channel and cross-time period pressure measured-predicted pairs, which can be used to trace back the complete spatiotemporal evolution process of the pressure system at the second level.

[0109] In this embodiment, firstly, historical pressure data for continuous time sections are extracted from the time dimension of the three-dimensional pressure correlation matrix in millisecond-level timestamp order via step S1041. A sliding window mechanism is used to dynamically read the latest N complete sections, for example, a window size of 100 sections covering a 2-second duration. Each channel independently acquires its spatiotemporal sequence: the physical location is located using the channel dimension identifier, the time dimension index is arranged according to the acquisition order, and abnormal fluctuation markers recorded in the associated fluctuation pattern dimension are used for data verification. The extraction process retains the original accuracy; for example, the measured value of the channel in the well at section T188 (timestamp 14:05:02.980) is 48.15 MPa, and at section T189 (14:05:03.000) it is 48.20 MPa, forming a continuous spatiotemporal pressure evolution trajectory. This step ensures that subsequent predictions are based on the complete equipment operating state context.

[0110] Secondly, based on the historical pressure sequence, step S1042 calculates the pressure change factor of adjacent time sections: first, the pressure difference between the current section and the previous section is obtained. Combined with fixed sampling interval Rate of change As a pressure change factor; then extrapolating the predicted value of the next section based on the pressure change factor. This prediction mechanism is compatible with fluid nonlinear characteristics. For example, if the outlet pressure of a turbine guide vane changes from 48.15 MPa to 48.20 MPa, then the pressure difference between the current section and the previous section will be calculated. The generated pressure change factor is Extrapolate the predicted value of the next cross section .

[0111] Next, in step S1043, the prediction results and real-time data are encapsulated into 20-byte binary units: the timestamp uses Unix millisecond format, for example, 1690891503000 represents "2023-08-01 14:05:03.000", the channel identifier is compressed into 4-byte code using a hash algorithm, for example, P2-1500m→0xC7A3, and the measured value and predicted value are each stored as 4-byte floating-point numbers, for example, 48.20→0x4248CCCD, 48.25→0x42483333. A constant-voltage channel synchronous generation unit (predicted value = measured value) forms a unified data structure. This design ensures efficient transmission of tens of thousands of units per second in the low-bandwidth environment of the deep-sea platform.

[0112] Finally, in step S1044, all channel units under the same timestamp are aggregated, such as the wellhead units at time T190 {measured: 35.22, predicted: 35.25}, P2 {48.20, 48.25}, and P3 {20.00, 20.00}. After being strictly ordered by timestamp in ascending order, a three-level B+ tree index is constructed: the root node is the millisecond-level timestamp, the intermediate nodes are partitioned by channel type (P1 / P2 / P3), and the leaf nodes store the physical address of the unit. The latest 100 sets of indexes reside in the memory buffer, and historical data is written to SSD and a secondary bitmap index is built to achieve millisecond-level retrieval based on the combination of "time + channel" conditions. For example, in a deep-sea drilling scenario, it is possible to quickly locate the "sudden drop in well channel pressure during the period of 14:05:03.000-14:05:05.000".

[0113] In practical applications, within the wellhead pressure control system, the first step is to extract the historical sequence of the latest 100 time sections from the P2-1500m sensor in the three-dimensional matrix time dimension. This includes the measured value of 48.15 MPa at section T188 (timestamp 14:05:02.980) and the value of 48.20 MPa at section T189 (14:05:03.000), and associate it with the "M1 high-frequency pulsation mode" marked in the fluctuation mode dimension. Next, the pressure change factor between adjacent sections is calculated. Extrapolated predicted value of T190 section The process then proceeds to encapsulation: a 20-byte binary unit is generated (timestamp 1690891503000 / channel hash 0xC7A3 / measured value 0x4248CCCD / predicted value 0x42483333), simultaneously processing the wellhead sensor P1-300m unit and the tailrace tube pulsation monitoring point P3 unit. Finally, the three units at time T190 are aggregated and written into a B+ tree index in ascending order of timestamp—the root node "1690891503000" is divided into three leaf nodes storing physical addresses, and the latest index is cached in memory. When a sudden drop in tailrace tube pressure from 101.50MPa to 100.80MPa is detected, deviating from the predicted value by 0.73MPa, the abnormal waveform segment is located within 5 milliseconds using the condition "timestamp ≥ 14:05:03.000 & channel = P3".

[0114] The overall solution of S104 described above accurately predicts short-term trends by calculating fluid pressure change factors in real time, generates a four-dimensional intelligent index unit that integrates timestamps, channel locations, measured values ​​and predicted values, and constructs an efficient retrieval structure with millisecond-level time series aggregation. This provides a core data framework for deep-sea pressure monitoring that combines predictive capabilities with millisecond-level backtracking capabilities, significantly improving the early warning efficiency of fluid pulsation anomalies and the accuracy of operating condition analysis.

[0115] S105. Continuously monitor the real-time load status of the storage medium, perform dynamic routing decisions based on the real-time load status and the real-time fluctuation frequency in the fluctuation mode dimension, simultaneously embed the four-dimensional intelligent index into a specific structured file header, and generate a text format index mapping table based on the four-dimensional intelligent index to achieve second-level backtracking of stored data.

[0116] Optionally, step S105 may specifically include the following steps:

[0117] S1051. Continuously monitor the queue length of the data to be written and the ratio of the remaining storage space on the storage medium. When the queue length exceeds a set threshold or the ratio of the remaining storage space is lower than a critical ratio, it is marked as a high-load state.

[0118] S1052. Calculate the real-time fluctuation frequency of each channel based on the repetitive fluctuation pattern in the fluctuation pattern dimension.

[0119] S1053. Under the high load state, when the real-time fluctuation frequency of the channel exceeds the frequency threshold, the corresponding real-time data stream is marked as a high-frequency fluctuation data stream and is preferentially written to the memory image block. When the real-time fluctuation frequency of the channel is lower than the frequency threshold, the corresponding real-time data stream is marked as a low-frequency trend data stream and is directly written to the solid-state storage chip.

[0120] S1054. Write the real-time data streams corresponding to all channels to the solid-state storage chip under non-high load conditions.

[0121] S1055. The four-dimensional intelligent index is converted into a binary format embedded in a specific structured file header of the data file. Each index unit in the four-dimensional intelligent index is converted into a text entry in the order of timestamp, channel identifier, actual measurement value, and predicted trend value. All text entries are aggregated to form a text format index mapping table for independent storage, thereby achieving second-level backtracking of the stored data.

[0122] In the above scheme, high load state refers to the critical operating state of the storage system facing data backlog and insufficient space, including the dual characteristics of the write queue length exceeding the safety threshold and the remaining storage space falling below the warning ratio, used to trigger emergency storage optimization strategies. Real-time fluctuation frequency refers to the channel signal activity index extracted from the fluctuation pattern dimension, including the number of triggers of repetitive fluctuation patterns per unit time and their duration ratio, used to distinguish the instantaneous value level of fluid pulsation data. Memory mirror block refers to a temporary high-speed cache area opened in volatile storage media, including zero-latency data writing and batch transfer mechanisms, used to ensure the complete capture of high-frequency transient pressure data. Text index mapping table refers to converting the four-dimensional intelligent index into a human-machine readable structured text directory, including columnar alignment characteristics of timestamps, channel identifiers, measured values, and predicted trend values, which can be used for second-level accident backtracking analysis by deep-sea platform operators.

[0123] In the embodiments of this application, such as Figure 2 As shown, firstly, key metrics are collected twice per second via the storage controller interface in step S1051: the length of the data queue to be written and the percentage of remaining storage space, such as the number of SSD tasks to be processed and the percentage of available space. A dual threshold determination mechanism is adopted: when the queue length exceeds the safety threshold or the remaining space is lower than the critical percentage, a high-load status flag is immediately triggered. This process achieves hardware-level response through a real-time comparator circuit. For example, when a queue backlog of 1200 entries is detected and the remaining space is only 18%, the system generates a high-load alarm signal and interrupts the current storage thread.

[0124] Secondly, in step S1052, the repetitive mode parameters of each channel are extracted from the fluctuation mode dimension of the three-dimensional matrix, and the number of mode triggers is counted based on the sliding time window. The real-time fluctuation frequency is calculated using the formula: For example, the "M1 mode" of the deep-sea drilling wellhead channel is triggered 12 times between 14:05:03.000 and 14:05:04.000, corresponding to a frequency of 12Hz; the steady-state mode of the tailrace reference channel is triggered only 0.2 times, corresponding to a frequency of 0.2Hz.

[0125] Next, in step S1053, under high load conditions, i.e., when the length of the data queue to be written is >1000 and the remaining storage space ratio is <20%, hierarchical storage is performed based on the real-time fluctuation frequency: First, channels with fluctuation frequencies ≥5Hz are identified as high-frequency fluctuating data streams, which are preferentially written to the pre-allocated DDR4 memory mirror blocks. Data overflow is prevented through ring buffer management and timed transfer mechanism; at the same time, channels with frequencies <5Hz are marked as low-frequency trend data streams, and direct memory access technology (DMA) is used to directly write to the specified logic blocks of the NVMe SSD chip to ensure zero loss of high-frequency transient data and reliable disk writing of steady-state data.

[0126] Then, when the system is under low load, i.e., the length of the data queue to be written is ≤800 and the remaining storage space ratio is ≥25%, all channel data undergoes a standardized storage process: parallel transmission through PCIe 4.0×4 channels, using a write aggregation engine in memory to stitch multi-channel data into 128KB continuous data blocks, using deep queue scheduling to batch write to the 3D TLC flash memory cells of the solid-state storage chip, and simultaneously activating a wear leveling algorithm to dynamically allocate physical block addresses, for example, merging the P1 pressure value of 48.2MPa and the standard pressure value of 20.0MPa into flash page 0x5A3F, significantly improving storage throughput efficiency.

[0127] Finally, in step S1055, the four-dimensional intelligent index unit is converted into a TLV structured format: the timestamp serves as the type identifier header, the channel identifier is compressed into fixed bytecode using a hash algorithm, and the measured and predicted values ​​are encapsulated according to floating-point precision specifications. The overall configuration is embedded in a specific file header field of the data file. Simultaneously, the text index conversion engine is started to generate column-aligned plain text entries in a strict order of timestamp, channel identifier, measured value, and predicted value. All entries are aggregated through in-memory streaming processing to form an independently stored index mapping table. Based on this mapping table, a hierarchical retrieval tree structure is constructed, with the timestamp as the root node to establish a fast addressing channel, achieving millisecond-level query response and complete data frame loading for any combination of "time + channel" conditions, providing closed-loop retrieval support for transient anomaly analysis of fluid pressure systems.

[0128] In practical applications, within a centrifugal pump unit pressure monitoring system, the storage controller detects in real-time that the length of the data queue to be written to the solid-state drive has reached 1300 entries and the remaining storage space is only 17%, triggering a high-load status flag. The system then calculates the activity of each channel from the perspective of the three-dimensional matrix fluctuation pattern. The "cavitation pulsation mode" of the impeller outlet channel is triggered 15 times per second (15Hz), while the steady-state channel at the suction end is only 0.5Hz. Based on the dynamic routing strategy, the 15Hz high-frequency pulsating data is written to the DDR4 memory mirror block (address 0x9000-0x97FF), and the 0.5Hz steady-state data is directly written to the NVMe solid-state drive. When the load is low... Once the load returns to normal (i.e., the length of the data queue to be written drops to 600 records and the remaining storage space is greater than 30%), all data flows through the PCIe channel and are aggregated into 128KB blocks for writing to flash memory. Simultaneously, the four-dimensional index is converted to TLV format and embedded into the 0x50 field of the file header, and a text index entry "1690900000000,ImpellerOutlet,1.52,1.55" is generated and stored in a CSV mapping table. When the impeller pressure is detected to drop sharply from 1.52MPa to 1.20MPa, the operator can locate the "cavitation pulsation abnormality" data segment within 2 seconds using the text index to complete the fault analysis.

[0129] The overall solution of S105 described above intelligently determines the differentiated storage path for high-frequency transient data and low-frequency trend data by sensing the dynamic coupling relationship between the storage load status and the frequency of fluid pressure fluctuations in real time, ensuring zero loss capture of key pulsation features; at the same time, it constructs a dual mapping mechanism of binary embedded index and text-readable index to realize second-level accurate backtracking and multi-dimensional correlation analysis of fluid pressure system under any complex working conditions.

[0130] The following is a complete example for steps 101-105, such as Figure 3 As shown, a hydraulic station pressure monitoring system receives 1000Hz sampling signals from three channels: the main pump outlet (C1), the control valve group (C2), and the oil tank circuit (C3). Time segments are cut at 20ms intervals. At segment T001, the covariance correlation between C1 and C2 is calculated to be 0.92 (strong correlation), and the correlation between C1 and C3 is 0.15 (weak correlation). The C1-C2 correlation sequence across 100 consecutive segments shows a sharp drop pattern of "0.9→0.1" every 5 segments. A three-dimensional matrix is ​​constructed, where the channel dimension corresponds to [C1, C2, C3], the time dimension corresponds to [T001-T100], and the fluctuation pattern dimension corresponds to [pattern M1].

[0131] Secondly, based on the synchronization of M1 mode, C1-C2 are grouped into a strongly correlated group, and C3 is an independent channel; after the original signal is converted into a binary stream, master-slave encoding is implemented for the C1-C2 group. The C1 master channel stores the original value of 15.2MPa, C2 stores the difference of +3.5MPa, and C3 uses entropy optimization compression. Detecting 3 consecutive 20.0MPa results in marking "20.0×3".

[0132] Next, the system generates controls [Channel 1, Channel 2, Channel 3] according to the physical layout; when the Channel 2 control is selected, the C2 difference stream is parsed: the current value of C1 of 15.3MPa is superimposed to restore the value of C2 of 18.8MPa, the pulsation curve is dynamically drawn and labeled "T002:18.8MPa"; after the control is closed, the rendering resources are released immediately.

[0133] Then, the latest 100 cross-sectional sequence [15.1, 15.2...15.3 MPa] of C1 is extracted from the three-dimensional matrix, and the cross-sectional variation factor is calculated. The next value is predicted to be 15.35 MPa; a four-dimensional cell {timestamp: 1690900000000, channel: Valve, measured: 15.30, predicted: 15.35} is generated and stored in order of timestamp with other channel cells.

[0134] Finally, when the storage system detected a backlog of 1500 SSD queue entries, it calculated that C1 experienced a 12Hz high-frequency fluctuation due to a sudden load change; triggering a high-load routing strategy: writing the high-frequency stream of C1 to the memory mirror area, and directly writing the steady-state stream of C3 to the SSD; simultaneously, converting the four-dimensional unit into a binary embedded file header (0x4C field), and generating the text index "1690900000000,Valve,15.30,15.35". When the main pump pressure abnormally dropped, engineers used the text index to locate the data frame at time "1690900000000" within 1.2 seconds to complete the fault diagnosis.

[0135] Figure 4 This is a schematic diagram illustrating a specific implementation of a dynamic real-time display and storage system for pressure data provided in this application. (Refer to...) Figure 4 The system may include:

[0136] Extraction module 41 is used to receive multi-channel digital pressure signals and extract the spatiotemporal correlation features between channels in real time, and construct a three-dimensional pressure correlation matrix based on the spatiotemporal correlation features. The three-dimensional pressure correlation matrix includes channel dimension, time dimension and fluctuation mode dimension.

[0137] Encoding module 42 is used to classify channel group types according to the fluctuation pattern dimension of the three-dimensional pressure correlation matrix, synchronously convert the multi-channel digital pressure signal into a real-time data stream, and perform differentiated encoding on the real-time data stream according to the channel group type.

[0138] Display module 43 is used to generate a set of channel controls based on the channel dimensions in the three-dimensional pressure correlation matrix, and synchronously parse the real-time data stream after differential encoding to dynamically construct a visual interactive display interface. The visual interactive display interface displays the channel identifiers corresponding to the control set, the pressure waveform restored by the real-time data stream, and the actual measured values ​​corresponding to the pressure waveform in real time.

[0139] The generation module 44 is used to predict future pressure change trends based on the historical pressure sequences stored in the time dimension of the three-dimensional pressure correlation matrix, and generate a four-dimensional intelligent index in combination with the future pressure change trends. The four-dimensional intelligent index consists of timestamps, channel identifiers, actual measured values ​​and predicted trend values.

[0140] The storage module 45 is used to continuously monitor the real-time load status of the storage medium, perform dynamic routing decisions based on the real-time load status and the real-time fluctuation frequency in the fluctuation mode dimension, simultaneously embed the four-dimensional intelligent index into a specific structured file header, and generate a text-formatted index mapping table based on the four-dimensional intelligent index to achieve second-level backtracking of stored data.

[0141] The dynamic real-time display and storage system for pressure data in this application embodiment is used to implement the aforementioned dynamic real-time display and storage method for pressure data. Therefore, the specific implementation of the dynamic real-time display and storage system for pressure data can be found in the embodiment section of the dynamic real-time display and storage method for pressure data above. The specific implementation can be referred to the description of the corresponding embodiments, and will not be repeated here.

[0142] This application also provides an electronic device, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the dynamic real-time display and storage method for pressure data described above.

[0143] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the above-described methods for dynamic real-time display and storage of pressure data.

[0144] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0145] Embodiments of the present invention also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in the embodiments of the dynamic real-time display and storage method for any of the pressure data described above.

[0146] Those skilled in the art will further 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, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0147] The foregoing has provided a detailed description of a method, system, electronic device, and storage medium for dynamic real-time display and storage of pressure data provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and core ideas of this application. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for dynamic real-time display and storage of pressure data, characterized in that, include: The system receives multi-channel digital pressure signals and extracts the spatiotemporal correlation features between channels in real time. Based on the spatiotemporal correlation features, a three-dimensional pressure correlation matrix is ​​constructed. The three-dimensional pressure correlation matrix includes channel dimension, time dimension, and fluctuation mode dimension. The channel group types are divided by the fluctuation pattern dimension of the three-dimensional pressure correlation matrix, and the multi-channel digital pressure signals are simultaneously converted into real-time data streams. The real-time data streams are then differentially encoded according to the channel group types. Based on the channel dimensions in the three-dimensional pressure correlation matrix, a set of channel controls is generated. The real-time data stream after differential encoding is parsed synchronously to dynamically construct a visual interactive display interface. The visual interactive display interface displays the channel identifiers corresponding to the control set, the pressure waveform restored by the real-time data stream, and the actual measured values ​​corresponding to the pressure waveform in real time. Based on the historical pressure sequence stored in the time dimension of the three-dimensional pressure correlation matrix, the future pressure change trend is predicted, and a four-dimensional intelligent index is generated by combining the future pressure change trend. The four-dimensional intelligent index consists of timestamp, channel identifier, actual measured value and predicted trend value. The system continuously monitors the real-time load status of the storage medium, performs dynamic routing decisions based on the real-time load status and the real-time fluctuation frequency in the fluctuation mode dimension, simultaneously embeds the four-dimensional intelligent index into a specific structured file header, and generates a text-formatted index mapping table based on the four-dimensional intelligent index to achieve second-level backtracking of stored data.

2. The method according to claim 1, characterized in that, The continuous monitoring of the real-time load status of the storage medium, the execution of dynamic routing decisions based on the real-time load status and the real-time fluctuation frequency in the fluctuation pattern dimension, the simultaneous embedding of the four-dimensional intelligent index into a specific structured file header, and the generation of a text-formatted index mapping table based on the four-dimensional intelligent index to achieve second-level backtracking of stored data, including: The queue length of the data to be written to the storage medium is continuously monitored, and the ratio of the remaining storage space is marked as a high load state when the queue length exceeds a set threshold or the ratio of the remaining storage space is lower than a critical ratio. The real-time fluctuation frequency of each channel is calculated based on the repetitive fluctuation patterns in the fluctuation pattern dimension. Under the high load condition, when the real-time fluctuation frequency of the channel exceeds the frequency threshold, the corresponding real-time data stream is marked as a high-frequency fluctuation data stream and is preferentially written to the memory image block. When the real-time fluctuation frequency of the channel is lower than the frequency threshold, the corresponding real-time data stream is marked as a low-frequency trend data stream and is directly written to the solid-state storage chip. Under non-high load conditions, write the real-time data streams corresponding to all channels to the solid-state storage chip; The four-dimensional intelligent index is converted into a binary format and embedded into a specific structured file header of the data file. Each index unit in the four-dimensional intelligent index is converted into a text entry in the order of timestamp, channel identifier, actual measurement value, and predicted trend value. All text entries are aggregated to form a text format index mapping table for independent storage, enabling second-level backtracking of stored data.

3. The method according to claim 1, characterized in that, The process of classifying channel groups by the fluctuation pattern dimension of the three-dimensional pressure correlation matrix and simultaneously converting the multi-channel digital pressure signals into real-time data streams includes: Based on the repetitive fluctuation patterns stored in the fluctuation pattern dimension of the three-dimensional pressure correlation matrix, the synchronicity intensity of any two channel fluctuation patterns is calculated. Based on the synchronization intensity, the channel group types are divided into three categories: channels with synchronization intensity significantly higher than the average synchronization threshold are classified as strongly correlated channels; channels with synchronization intensity close to the average synchronization threshold are classified as weakly correlated channels; and channels with synchronization intensity significantly lower than the average synchronization threshold are classified as independent channels. The digital pressure signal of each channel is converted into a continuous binary data block in time sequence, and the binary data blocks of all channels are spliced ​​together according to the acquisition time sequence to form a real-time data stream.

4. The method according to claim 3, characterized in that, The process involves generating a set of channel controls based on the channel dimensions in the three-dimensional pressure correlation matrix, simultaneously parsing the differentiated encoded real-time data stream, and dynamically constructing a visual interactive display interface. This interface displays in real-time the channel identifiers corresponding to the control set, the pressure waveform reconstructed from the real-time data stream, and the actual measured values ​​corresponding to the pressure waveform, including: Extract channel identifiers and their corresponding physical layout order from the channel dimensions of the three-dimensional pressure correlation matrix to generate a selectable set of channel controls, with each control bound to a unique channel identifier; Based on the coding rules of the channel group type, the real-time data stream after differential coding is parsed, the direct pressure value restoration is performed on the main channel of the strongly correlated channel group, the residual superposition restoration is performed on the slave channel, the waveform reconstruction is performed on the weakly correlated channel group, and the entropy optimization data recovery is performed on the independent channel. Create a dynamic visual interactive display interface. When the user selects the control corresponding to the channel identifier, the waveform display area is activated in the visual interactive display interface. The analyzed pressure waveform is drawn in real time in the waveform display area, and the actual measured value of the latest pressure value is dynamically marked. When the user cancels the selection of the control corresponding to the channel identifier, the waveform display area is closed and the parsing process of the corresponding channel is terminated.

5. The method according to claim 1, characterized in that, The historical pressure sequence stored in the time dimension of the three-dimensional pressure correlation matrix is ​​used to predict future pressure change trends. A four-dimensional intelligent index is generated based on these future pressure change trends. This four-dimensional intelligent index consists of a timestamp, channel identifier, actual measured value, and predicted trend value, including: Historical pressure sequences of continuous time sections are extracted from the time dimension of the three-dimensional pressure correlation matrix. The historical pressure sequences contain measured pressure values ​​of each channel at multiple time sections. Based on the historical pressure sequence, the pressure change factor of adjacent time sections is calculated, and the predicted pressure value of the next time section is extrapolated by the pressure change factor to predict the future pressure change trend. A four-dimensional intelligent index unit is generated by combining the future pressure change trend. The four-dimensional intelligent index unit includes the timestamp of the current time section, the channel identifier of the corresponding channel, the measured pressure value of the current time section, and the predicted pressure value of the next time section. The four-dimensional intelligent index units of all channels are aggregated and stored in ascending order of timestamp to form a four-dimensional intelligent index set.

6. The method according to claim 3, characterized in that, The step of differentially encoding the real-time data stream according to the channel group type includes: For strongly correlated channel groups, master-slave channel collaborative coding is implemented. A master channel is designated in the strongly correlated channel group, and the original pressure value of the master channel in the real-time data stream is saved. The slave channels only save the difference between their own pressure value and the pressure value of the master channel at the same time. For weakly correlated channel groups, trend feature preservation and compression are implemented, recording only the direction and magnitude of pressure value changes at adjacent time points in the real-time data stream, and deleting the basic value portion of the pressure values ​​that has not changed. For independent channels, entropy optimization compression is performed. The same pressure value that is repeated continuously in the real-time data stream is detected. Only the first pressure value is retained and the start time and number of repetitions of the pressure value are recorded.

7. The method according to claim 1, characterized in that, The system receives multi-channel digital pressure signals and extracts spatiotemporal correlation features between channels in real time. Based on these spatiotemporal correlation features, a three-dimensional pressure correlation matrix is ​​constructed. This three-dimensional pressure correlation matrix includes channel dimension, time dimension, and fluctuation mode dimension, including: Simultaneously receive multi-channel digital pressure signals, and perform equal-interval slicing operations on the digital pressure signals of each channel under a unified time reference to form time sections aligned in the time dimension. At each time segment, channels are paired up into channel pairs, the signal change correlation in each channel pair is calculated, the spatiotemporal correlation features of the time segment are generated, and the spatiotemporal correlation feature values ​​of all channel pairs constitute the spatiotemporal correlation feature set of the time segment. The spatiotemporal correlation feature sets of multiple consecutive time sections are connected end to end in chronological order to form a channel spatiotemporal correlation feature sequence across time sections; A three-dimensional pressure correlation matrix is ​​constructed based on the spatiotemporal correlation feature sequence. In the three-dimensional pressure correlation matrix, the channel identifiers of all channels are arranged in the channel dimension, the sequential number of time sections is arranged in the time dimension, and the fluctuation pattern dimension stores the repetitive fluctuation patterns identified in the spatiotemporal correlation feature sequence. The repetitive fluctuation patterns are manifested as similar waveform segments that appear at fixed intervals in the spatiotemporal correlation feature sequence.

8. A dynamic real-time display and storage system for pressure data, characterized in that, include: The extraction module is used to receive multi-channel digital pressure signals and extract the spatiotemporal correlation features between channels in real time. Based on the spatiotemporal correlation features, a three-dimensional pressure correlation matrix is ​​constructed. The three-dimensional pressure correlation matrix includes channel dimension, time dimension and fluctuation mode dimension. The encoding module is used to classify the channel group type according to the fluctuation pattern dimension of the three-dimensional pressure correlation matrix, synchronously convert the multi-channel digital pressure signal into a real-time data stream, and perform differentiated encoding on the real-time data stream according to the channel group type. The display module is used to generate a set of channel controls based on the channel dimensions in the three-dimensional pressure correlation matrix, and synchronously parse the real-time data stream after differential encoding to dynamically construct a visual interactive display interface. The visual interactive display interface displays the channel identifiers corresponding to the control set, the pressure waveform restored by the real-time data stream, and the actual measured values ​​corresponding to the pressure waveform in real time. The generation module is used to predict future pressure change trends based on the historical pressure sequences stored in the time dimension of the three-dimensional pressure correlation matrix, and generate a four-dimensional intelligent index in combination with the future pressure change trends. The four-dimensional intelligent index consists of timestamps, channel identifiers, actual measured values ​​and predicted trend values. The storage module is used to continuously monitor the real-time load status of the storage medium, perform dynamic routing decisions based on the real-time load status and the real-time fluctuation frequency in the fluctuation mode dimension, simultaneously embed the four-dimensional intelligent index into a specific structured file header, and generate a text-formatted index mapping table based on the four-dimensional intelligent index to achieve second-level backtracking of stored data.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the dynamic real-time display and storage method for pressure data as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, enables a method for dynamic real-time display and storage of pressure data as described in any one of claims 1 to 7.

Images