Real-time data acquisition and processing method and system of distributed control system

Abstract

The invention relates to the technical field of data acquisition and processing, and provides a real-time data acquisition and processing method and system for a distributed control system, and the method comprises the steps: collecting a data flow in real time through a terminal sensor, analyzing the probability distribution of a data change rate through a sliding window, and obtaining a data change rate; and a sensitivity threshold is intelligently generated by combining dynamic factors such as equipment electric quantity and network load. When data mutation exceeds a threshold value, the system automatically extracts a key change section and generates time, period and variable quantity three-dimensional features, and meanwhile, the frequency domain analysis result is fused to enhance the anti-interference performance. And the event emergency degree is calculated based on the time-space correlation of the spectrum energy entropy and the characteristic parameters, and finally a processing queue sorted according to priorities is formed and the bandwidth is dynamically allocated. According to the method, through multi-dimensional data fusion and resource adaptive adjustment, the response speed of the system to emergencies, the data discrimination precision and the stability in a complex environment are improved.

Description

Technical Field

[0001] This application relates to the technical field of data acquisition and processing. Specifically, it relates to a real-time data acquisition and processing method and system for a distributed control system. Background Art

[0002] The content of this part only provides background information related to this application, and it may not constitute prior art.

[0003] In the transmission of digital information, a distributed control system continuously senses device status and environmental information through node devices spread throughout the field. These acquisition points are like the nerve endings of the system, capturing changes in real time. Data is rapidly aggregated to the processing node through a dedicated industrial network. In the transmission and processing process, the system strictly ensures consistent timing, providing an accurate basis for judgment. The aggregation point quickly cleans, transforms, and preliminarily refines the raw data. Finally, the integrated real-time information is sent to the decision-making center to support the immediate generation and issuance of control instructions. The entire process is interlocked, fast and smooth, and can handle anomalies such as network interruptions to ensure the uninterrupted operation of core functions.

[0004] In the prior art, Chinese Patent with the authorization announcement number CN104464254B discloses a distributed data synchronous acquisition device and method. This method realizes synchronous data acquisition of different devices by deploying multiple data acquisition nodes, and analyzes and processes the acquired data through a centralized processing unit.

[0005] However, this patent has some drawbacks. First, its data processing method is relatively single, mainly relying on simple synchronous acquisition and centralized processing, lacking dynamic monitoring of the data change rate and adaptive adjustment of the sensitivity threshold. This will result in the inability to timely capture changes in key data in the face of a complex and changing industrial environment, thus affecting the response speed and accuracy of the system. Second, this patent cannot effectively cope with noise interference in the data, thereby affecting the accuracy of data processing. In addition, its event priority evaluation and resource allocation mechanism is relatively simple, and it fails to fully consider the impact of factors such as devices and networks on event processing, resulting in unreasonable resource allocation and affecting the overall performance of the system.

[0006] In view of the above problems, the present invention proposes a real-time data acquisition and processing method and system for a distributed control system to improve the real-time performance, accuracy, and anti-interference ability of the system. Summary of the Invention

[0007] In order to solve the above technical problems, the purpose of this application is to provide a real-time data acquisition and processing method for a distributed control system, which improves the real-time performance, accuracy, and anti-interference ability of the system by introducing a dynamic sensitivity threshold, frequency domain analysis, and an event priority evaluation model based on a resource attenuation coefficient.

[0008] The object of the present application is achieved by the following technical solutions: In a first aspect, the present invention provides a method for real-time data acquisition and processing of a distributed control system, including: Continuously collecting a target data stream through sensors deployed on terminal nodes, and using a sliding time window to statistically analyze the probability density distribution of the instantaneous change rate of the target data stream; Selecting a preset percentile value of the probability density distribution as a basic sensitivity threshold, calculating a dynamic adjustment coefficient based on the remaining power of the corresponding device, the network load factor, and the historical transmission energy consumption, and generating a target sensitivity threshold by weighted fusion of the basic sensitivity threshold and the dynamic adjustment coefficient; When any instantaneous change rate exceeds the target sensitivity threshold, calculating the difference sequence of adjacent sampling points of the instantaneous change rate, and screening the sections with a change amplitude greater than the amplitude threshold through a preset amplitude threshold; merging the continuously co-directional change points within the section to generate a feature triple including a start timestamp, a duration period, and an accumulated change amount; Performing frequency-domain analysis on the target data stream, obtaining the effective spectral features of the current data window through Fourier transform of the sliding window, aligning the effective spectral features and the feature triple in space and time, using the signal-to-noise ratio of each frequency band in the spectral features as a dynamic weight coefficient, and generating an anti-interference enhanced feature vector by weighted fusion; Based on the anti-interference enhanced feature vector, constructing a dynamic event priority evaluation model; the evaluation model includes time-aligning the energy entropy of each frequency band in the feature vector with the duration period in the feature triple, calculating the time-series mutation intensity through the ratio of the fluctuation amplitude of the spectral energy entropy to the duration period; multiplying the time-series mutation intensity by the accumulated change amount to obtain an initial urgency, and constructing a resource attenuation coefficient in combination with the remaining power of the current node and the network load factor, and multiplying the initial urgency by the resource attenuation coefficient to generate an event urgency score; Generating a weighted event processing queue according to the descending order of the event urgency scores; dynamically adjusting the bandwidth occupancy ratio of the transmission channel according to the event urgency scores of each event in the event processing queue.

[0009] Further, before calculating the difference sequence of adjacent sampling points of the instantaneous change rate, it further includes: Performing a moving average filter on the sequence of the target data to eliminate instantaneous interference.

[0010] Further, the step of obtaining the effective spectral features of the current data window through Fourier transform of the sliding window specifically includes: Obtaining the spectral features of the current data window through Fourier transform of the sliding window; Extracting the main frequency component amplitude and frequency point distribution in the spectral features; Compare with the reference noise spectrum of the device, and separate the inherent noise components of the system through spectral subtraction; Combine the temperature and electromagnetic intensity parameters collected by the environmental sensor, and use the three-level wavelet packet decomposition and dynamic energy ratio detection method to identify the environmental noise frequency band; Generate a noise suppression weight matrix according to the energy distribution ratio of the inherent noise components and the environmental noise frequency band; Perform a dot product operation on the main frequency component and the noise suppression weight matrix, and output the optimized effective spectral features.

[0011] Further, the steps of obtaining the spectral features of the current data window through the Fourier transform of the sliding window specifically include: Extend the target data within the current time window forward to cover a preset percentage of the data area at the end of the previous window to form a new window with data overlap; Perform smooth transition processing on the 5% data areas at the head and tail of the new window; Perform standard spectral transform calculation on the processed window data. Based on the frequency components whose energy intensity exceeds the set standard and their distribution positions in the spectrum of the previous window, identify the frequency components whose energy intensity suddenly increases in the spectrum of the current window, and calibrate their positions; screen out the frequency components that repeatedly appear at the same frequency position in at least two consecutive windows, and merge them to form the spectral features of the current data window.

[0012] Further, the steps of dynamically adjusting the bandwidth occupancy ratio of the transmission channel specifically include: When the event urgency score exceeds the upper limit of the preset urgency threshold range, automatically suspend the occupancy of the reserved bandwidth resource channel for low-priority events, and allocate this resource channel for use by the event with an urgency score exceeding the limit; Configure a fixed step adjustment strategy corresponding to the urgency score value range, perform an incremental adjustment of the bandwidth resource according to the preset step value matching the value range, and approve the real-time adjustment amount through the two-way negotiation protocol between nodes; Deploy a centralized coordinator to monitor the network-wide load status. When the remaining bandwidth capacity of a node is lower than the system preset capacity threshold, freeze the bandwidth resource application operation for low-priority events, and at the same time allocate the minimum guaranteed bandwidth for events with an urgency score higher than the guarantee threshold.

[0013] Further, the formula for generating the target sensitivity threshold is:

[0014] Wherein, is the inverse function (quantile function) of the probability density distribution function ; is the preset percentile value; is the percentage of the remaining battery power of the current device, is the attenuation coefficient; is the real-time occupancy of the TCP / IP protocol stack buffer, is the total buffer capacity; is the energy consumption sensitivity coefficient, is the average energy consumption per unit data volume in the past 24 hours; is the basic threshold weight, is the dynamic adjustment coefficient weight, satisfying .

[0015] Furthermore, after dynamically adjusting the bandwidth ratio of the transmission channel, it further includes: Comparing the event urgency score with a preset threshold, automatically triggering an alarm signal when the score exceeds the threshold, and at the same time marking the feature vector data of the event as high priority and pushing it to the cloud server for storage.

[0016] In a second aspect, the present invention provides a real-time data acquisition and processing system for a distributed control system, including: A data stream acquisition and probability analysis module, which is used to continuously acquire the target data stream through sensors deployed on terminal nodes, and adopt a sliding time window to statistically analyze the probability density distribution of the instantaneous change rate of the target data stream; A dynamic threshold generation module, which is used to select a preset percentile value of the probability density distribution as the basic sensitivity threshold, calculate the dynamic adjustment coefficient based on the remaining battery power of the corresponding device, the network load factor, and the historical transmission energy consumption, and generate the target sensitivity threshold through weighted fusion of the basic sensitivity threshold and the dynamic adjustment coefficient; A feature triple extraction module, which is used to calculate the difference sequence of adjacent sampling points of the instantaneous change rate when any instantaneous change rate exceeds the target sensitivity threshold, and screen the sections with a change amplitude greater than the amplitude threshold through a preset amplitude threshold; merge the continuously co-directional change points within the section to generate a feature triple including the start timestamp, the duration period, and the cumulative change amount; A frequency domain enhancement processing module, which is used to perform frequency domain analysis on the target data stream, obtain the effective spectral features of the current data window through the Fourier transform of the sliding window, align the effective spectral features and the feature triple in space and time, use the signal-to-noise ratio of each frequency band in the spectral features as the dynamic weight coefficient, and generate an anti-interference enhanced feature vector through weighted fusion; A priority evaluation modeling module for constructing a dynamic event priority evaluation model based on an anti-interference enhanced feature vector; the evaluation model includes time-aligning the energy entropy of each frequency band in the feature vector with the duration period in the feature triple, and calculating the time series mutation intensity through the ratio of the spectral energy entropy fluctuation amplitude to the duration period; multiplying the time series mutation intensity by the cumulative change amount to obtain the initial urgency, constructing a resource attenuation coefficient by combining the remaining power of the current node and the network load factor, and generating an event urgency score by multiplying the initial urgency by the resource attenuation coefficient; A bandwidth dynamic scheduling module for generating a weighted event processing queue according to the descending order of the event urgency scores; dynamically adjusting the bandwidth occupation ratio of the transmission channel according to the event urgency scores of the events in the event processing queue.

[0017] In a third aspect, the present invention provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, the steps corresponding to the method in the first aspect are implemented.

[0018] In a fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps corresponding to the method in the first aspect are implemented.

[0019] In summary, the technical solutions of the embodiments of the present application have at least the following advantages and beneficial effects: The present invention continuously collects data streams through terminal sensors, analyzes the probability distribution characteristics of the instantaneous change rate using a sliding time window, combines dynamic parameters such as device power and network load to generate an adaptive sensitivity threshold, and realizes the accurate capture of abnormal changes. Through difference sequence analysis and continuous same-direction change point merging, triple data with time-dimensional characteristics is extracted, and at the same time, spectral characteristics are extracted using frequency-domain analysis and aligned with spatio-temporal characteristics, and an anti-interference feature vector is constructed by dynamically weighting with the signal-to-noise ratio. Based on the calculation of the time series mutation intensity of the spectral energy entropy and the duration period, a priority evaluation model is constructed by combining the resource state, and finally a weighted event processing queue is formed and the transmission bandwidth is dynamically allocated. This method improves the real-time response of the system to emergencies, the accuracy of data discrimination, and the anti-interference ability in complex environments through multi-dimensional feature fusion and dynamic resource adaptation mechanisms. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a flowchart of a method for real-time data acquisition and processing of a distributed control system provided by the present invention; Figure 2 is a structural schematic diagram of a real-time data acquisition and processing system of a distributed control system provided by the present invention; Figure 3Schematic diagram of the structure of an electronic device provided by the present invention. Specific embodiments

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. Components of the embodiments of the present application usually described and illustrated in the drawings here can be arranged and designed in various different configurations.

[0022] As Figure 1 shown, a real-time data acquisition and processing method for a distributed control system proposed in an embodiment of the present application includes: S101, continuously collect the target data stream through sensors deployed at the terminal nodes, and use a sliding time window to statistically calculate the probability density distribution of the instantaneous change rate of the target data stream.

[0023] Specifically, multi-type sensors (such as temperature, pressure, or vibration sensors, etc.) deployed at the terminal nodes continuously collect the target data stream at a fixed sampling frequency, and the sampling interval is set to milliseconds to seconds according to the application scenario. The sliding time window uses a dynamically adjustable window length, typically set to 10 to 30 sampling periods, and the window sliding step size is a single sampling interval to ensure data continuity. When calculating the instantaneous change rate within the window, the backward difference method is used to obtain the numerical change gradient of adjacent sampling points to form a change rate time series.

[0024] In addition, the probability density distribution statistics are realized through the kernel density estimation algorithm. A Gaussian kernel function is selected to smooth all the instantaneous change rate data within the window to generate a continuous probability distribution curve. This curve reflects the dynamic characteristics of the current monitoring parameter within the time window. For example, in the vibration monitoring of industrial equipment, the change rate is concentrated in the range of ±0.5 m / s² under normal working conditions, and a long-tailed distribution will appear under abnormal working conditions. Its beneficial effect is that the statistical law of the dynamic behavior of the system is quantitatively characterized by the probability density distribution, which is more adaptable than a fixed threshold. At the same time, the sliding window mechanism achieves a balance between the real-time nature and historical relevance of data processing, avoiding false triggering of single-point mutations. Its calculation formula is as follows: (1) where is the instantaneous change rate is the probability density function; is the number of sampling periods included in the sliding time window (i.e., the sliding window length); is the number of sampling points within the window, ; is the bandwidth parameter, controlling the smoothing degree; is the instantaneous change rate of the th sampling point within the window; is the input variable (change rate value); is the serial number of the sampling point corresponding to the starting moment of the sliding window (integer index).

[0025] S102: Select the preset percentile value of the probability density distribution as the basic sensitivity threshold, calculate the dynamic adjustment coefficient based on the remaining battery power of the corresponding device, the network load factor, and the historical transmission energy consumption, and generate the target sensitivity threshold through weighted fusion of the basic sensitivity threshold and the dynamic adjustment coefficient.

[0026] Specifically, first extract the preset percentile value (such as the 90th percentile) from the probability density distribution curve generated in step S101 as the basic sensitivity threshold, which reflects the statistical boundary characteristics of the monitoring parameters under normal working conditions. The calculation of the dynamic adjustment coefficient comprehensively considers three key parameters: the remaining battery power of the device, the network load factor, and the historical transmission energy consumption. The remaining battery power of the device is quantified using an exponential decay model to ensure a reduced acquisition intensity at low battery levels. The network load factor is calculated by real-time monitoring of the buffer occupancy rate of the TCP / IP protocol stack. The historical transmission energy consumption is weighted based on the average energy consumption per unit data volume within the past 24 hours. The weighted fusion process is implemented using a fuzzy logic controller, and the specific values are automatically adapted according to the device type to finally generate the target sensitivity threshold. This step improves the traditional fixed threshold to a dynamic threshold system with environmental perception ability, realizing adaptive data acquisition optimization. The specific calculation process is as follows: The basic sensitivity threshold is (2) where is the inverse function (quantile function) of the probability density distribution function ; is the preset percentile value (for example, the 90th percentile corresponds to being 0.9) The dynamic adjustment coefficient is: (3) where is the remaining battery power factor, is the network load factor, is the historical transmission energy consumption factor.

[0027] The expression of the remaining battery power factor is: (4) where is the current percentage of the remaining battery power of the device (0 ≤ ≤ 1), is the attenuation coefficient.

[0028] Network load factor is expressed as: (5) where is the real-time occupancy of the TCP / IP protocol stack buffer, is the total buffer capacity.

[0029] Historical transmission energy consumption factor is expressed as: (6) where is the energy consumption sensitivity coefficient, is the average energy consumption per unit data volume in the past 24 hours.

[0030] Target sensitivity threshold is expressed as: (7) where is the basic threshold weight, is the dynamic adjustment coefficient weight, satisfying .

[0031] Substituting formulas (2), (3), (4), (5) and (6) into formula (7) gives: (8) S103. When any instantaneous change rate exceeds the target sensitivity threshold, calculate the difference sequence of adjacent sampling points of the instantaneous change rate, and screen out the sections with a change amplitude greater than the amplitude threshold through a preset amplitude threshold; merge the continuously same-direction change points within the section to generate a feature triple containing the start timestamp, duration period, and cumulative change amount.

[0032] Specifically, when the instantaneous change rate of the sensor exceeds the step target sensitivity threshold, the system first calculates the difference sequence of adjacent sampling points of the instantaneous change rate, and this difference sequence reflects the local fluctuation characteristics of the monitoring parameter. By using a preset amplitude threshold (such as set to 0.3 m / s² in industrial vibration monitoring) to screen out the sections with significant change amplitudes, the random noise interference can be effectively filtered out. For the screened sections, the system automatically merges the continuously same-direction change points (such as the pressure sensor shows an upward trend for 5 consecutive sampling periods), and generates a feature triple containing the start timestamp, duration period, and cumulative change amount. This triple completely records the spatio-temporal characteristics and intensity evolution process of the abnormal event.

[0033] Its technical principle lies in: the difference sequence calculation adopts the backward difference method to maintain consistency with the algorithm in step S101; the amplitude threshold setting needs to combine the physical dimension of the specific application scenario. For example, in the pressure monitoring of an oil pipeline, it is set to 0.2 MPa; the merging of same-direction change points adopts the sliding window matching algorithm to ensure the accurate division of event boundaries. Taking the vibration monitoring of a wind turbine as an example, when the vibration change rate exceeds the threshold due to blade bearing wear, the system will capture a characteristic triple that lasts for 8 seconds and has an accumulated change of 4.7 m / s², accurately reflecting the fault development process. Thus, the dynamic threshold trigger mechanism reduces the amount of invalid data transmission; the use of the characteristic triple structure upgrades the event description dimension from single-point alarm to process tracking, improving the fault diagnosis accuracy; the accurate recording of timestamps and continuous periods provides basic data for the spatio-temporal alignment of subsequent steps. This step highly coincides with the data compression requirements of the distributed system, and each terminal node can independently complete feature extraction, thereby reducing the network transmission load.

[0034] The calculation process is as follows: ‌Difference sequence is: (9) where is the instantaneous change rate difference of the th sampling point, the instantaneous change rate of the th sampling point within the sliding window, is the starting sampling point serial number of the sliding window, is the length of the sliding window.

[0035] The amplitude threshold screening condition is set as: (10) where is the preset amplitude threshold.

[0036] The start timestamp in the characteristic triple is: (11) where is the start time of the abnormal event, is the time of the first sampling point that meets the amplitude threshold.

[0037] The continuous period in the characteristic triple is: (12) where is the duration of the abnormal event; is the termination sampling point serial number of the continuous same-direction change section, is the starting sampling point serial number of the continuous same-direction change section; is the sampling interval.

[0038] Cumulative change in the feature triple: (13) where is the algebraic sum of all differences within the section; is an index variable representing the serial number of the sampling point; represents the th difference in the instantaneous change rate of the sampling point, defined as the difference in the change rate between adjacent sampling points.

[0039] It should be noted that before calculating the difference sequence of adjacent sampling points of the instantaneous change rate, the sequence of the target data can also be subjected to moving average filtering to eliminate instantaneous interference.

[0040] S104. Perform frequency domain analysis on the target data stream, obtain the effective spectral features of the current data window through Fourier transform of the sliding window, align the effective spectral features with the feature triple in space-time, use the signal-to-noise ratio of each frequency band in the spectral features as the dynamic weight coefficient, and generate an anti-interference enhanced feature vector through weighted fusion.

[0041] This step performs an operation to generate an anti-interference enhanced feature vector for the target data stream. The core lies in constructing a composite feature representation with noise suppression ability through sliding window frequency domain analysis and space-time alignment of the feature triple. Specifically, obtain the effective spectral features of the current data window through Fourier transform of the sliding window: extend the target data within the current time window forward to cover a data area with a preset ratio at the end of the previous window, forming a new window with data overlap. This design uses the continuity of historical data to avoid the spectral leakage phenomenon generated by the conventional sliding window at the boundary. Subsequently, perform smooth transition processing on the data areas at the head and tail edges of the new window, and use the window function weighting technique to suppress the high-frequency interference components introduced by data jumps to ensure the stability of spectral analysis.

[0042] Then, perform standard spectral transform calculation on the preprocessed window data, establish a dynamic frequency tracking template based on the frequency components whose energy intensity exceeds the set standard and their distribution positions in the spectrum of the previous window. Complete the frequency position calibration by comparing the spatial distribution relationship between the frequency components with abnormally increased energy intensity in the current window spectrum and the template. Screen the frequency components that repeatedly appear at the same frequency position in at least two consecutive windows, and merge them to form effective spectral features reflecting the true dynamics of the system. This mechanism effectively filters out transient random noise and retains the fault feature frequencies with time persistence.

[0043] Subsequently, extract the amplitude and frequency point distribution of the main frequency component in the effective spectral features, perform spectral subtraction operation in combination with the equipment reference noise map, and separate the inherent mechanical noise components of the system. At the same time, synchronously access the temperature and electromagnetic intensity parameters collected by the environmental sensor, use the three-level wavelet packet decomposition technology to strip the signal frequency band layer by layer, and quantitatively identify the dominant frequency band of environmental noise through the dynamic energy ratio detection method. According to the energy ratio distribution of the inherent noise and environmental noise in the frequency domain, generate a noise suppression weight matrix in the frequency dimension. This matrix assigns a high weight value to the fault feature frequency band and strongly attenuates the noise aggregation frequency band. Finally, perform weighted combination of the main frequency component amplitude and the noise suppression weight matrix, and output the optimized effective spectral features to achieve directional enhancement of the fault features.

[0044] The beneficial effects are as follows: Align the effective spectral features with the feature triple generated in step S103 in space and time. In the alignment process, based on the start timestamp recorded by the feature triple, locate the corresponding time interval of the spectral analysis to ensure the space-time consistency of the time-domain event and the frequency-domain feature. Use the signal-to-noise ratio of each frequency band in the spectral features as the dynamic weight coefficient to perform weighted fusion on the cumulative change amount parameters in the feature triple: the fault features corresponding to the high signal-to-noise ratio frequency bands are given higher weights, while the low signal-to-noise ratio frequency bands are down-weighted to enhance the anti-interference ability of the feature vector.

[0045] The specific calculation process is as follows: The expression formula of the sliding window spectral analysis is: (14) Among them, is the complex spectral value of the th frequency component; is the time-domain data of the th sampling point within the window; Hanning window function coefficient (dimensionless); is the window length; is the frequency index; is the imaginary number.

[0046] The effective frequency band screening condition is: (15) Among them, is the effective frequency band index set; is the energy threshold ratio; is the vector of all frequency amplitudes.

[0047] The signal-to-noise ratio weighting formula is: (16) (17) wherein, is the normalized weight of the frequency band ; is the signal-to-noise ratio of the frequency band ; is the noise floor power; is the index.

[0048] (18) wherein, is the final feature vector, is the start timestamp, is the duration period, is the cumulative change amount, is the frequency index.

[0049] S105. Based on the anti-interference enhanced feature vector, construct a dynamic event priority evaluation model; the evaluation model includes time-aligning the energy entropy of each frequency band in the feature vector with the duration period in the feature triple, calculating the time series mutation intensity through the ratio of the fluctuation amplitude of the spectral energy entropy to the duration period; multiplying the time series mutation intensity by the cumulative change amount to obtain the initial urgency, combining the remaining power of the current node and the network load factor to construct a resource attenuation coefficient, and multiplying the initial urgency by the resource attenuation coefficient to generate an event urgency score.

[0050] Specifically, first, time-align the energy entropy of each frequency band in the anti-interference enhanced feature vector with the duration period parameter recorded in the feature triple to ensure that the frequency domain statistical characteristics and the time domain event process are on the same time basis. The technical principle is that the spectral energy entropy characterizes the complexity fluctuation characteristics of the signal frequency domain components, while the duration period reflects the time domain continuity of abnormal events, and the alignment of the two can achieve spatio-temporal feature fusion. Through the ratio operation of the fluctuation amplitude of the aligned spectral energy entropy to the duration period, the time series mutation intensity index is calculated, which quantifies the abnormal mutation degree of the frequency domain features per unit time and effectively distinguishes sudden faults from progressive deterioration.

[0051] Subsequently, multiply the time series mutation intensity by the cumulative change amount in the feature triple to generate the initial urgency parameter. The mathematical principle of using the multiplication operation here is that the cumulative change amount reflects the overall impact scale of abnormal events, and the time series mutation intensity characterizes the severity of event development. Multiplying the two can comprehensively evaluate the overall hazard level of the event. For example, in the wind turbine monitoring scenario, if a crack in the blade causes a sudden increase in the energy entropy of the vibration signal in a specific frequency band (high time series mutation intensity), and the crack expansion causes the cumulative change amount to continue to increase, the product of the two will significantly increase the value of the initial urgency, accurately matching the actual risk level of mechanical damage.

[0052] Furthermore, a resource attenuation coefficient is constructed by combining the remaining power of the current node and the network load factor. The design principle of the resource attenuation coefficient is based on the resource constraint characteristics of the distributed system: when the remaining power of the node is lower than the safety threshold or the network load is too high, the processing intensity of low-priority events needs to be suppressed to ensure key functions. This coefficient is dynamically generated through fuzzy logic rules. For example, the remaining power is given a non-linear weight using an exponential decay model, and the network load factor is calculated based on the real-time buffer occupancy rate of the transport layer protocol stack. The two work together to determine the resource availability level. Finally, the initial urgency is multiplied by the resource attenuation coefficient to generate an event urgency score. This operation achieves a dual optimization goal - both retaining the technical severity assessment of the event itself and embedding the current resource status constraints of the node, making the scoring result meet the actual scheduling requirements of the distributed system.

[0053] The specific calculation formula is as follows: Spectrum energy entropy The expression of (19) Among them, is the spectrum energy entropy, representing the complexity of the frequency domain components; is the normalized weight of the frequency band and is the set of effective frequency band indices.

[0054] Temporal mutation intensity The expression of (20) Among them, is the temporal mutation intensity, the fluctuation amplitude of the frequency domain entropy per unit time; is the spectrum energy entropy of the current window; is the spectrum energy entropy of the previous window; is the duration period in the feature triple.

[0055] Initial urgency The expression of (21) Among them, is the initial urgency; is the cumulative change amount in the feature triple.

[0056] Resource attenuation coefficient The expression of (22) Among them, is the resource attenuation coefficient, is the device remaining power factor, is the network load factor.

[0057] Event urgency score The expression is as follows: (23) Wherein, is the final event urgency score.

[0058] S106. Arrange in descending order according to the event urgency score to generate a weighted event processing queue; dynamically adjust the bandwidth ratio of the transmission channel according to the urgency scores of the events in the event processing queue.

[0059] This step aims to design a dynamic resource scheduling mechanism based on the event urgency score, that is, generate a weighted event processing queue according to the descending order of the event urgency score, and then dynamically adjust the bandwidth ratio of the transmission channel according to the urgency scores of the events in the queue to achieve the efficiency and fairness of event processing in the distributed control system. Specifically, when generating the weighted event processing queue, the system uses the event urgency score of each event as the queue weight benchmark, and gives priority to processing high-score events through descending order. The principle is that the urgency score comprehensively considers the temporal mutation intensity, cumulative change amount of the event and the node resource status, and can accurately reflect the technical severity and processing priority of the event; the beneficial effect is that this queue structure overcomes the disadvantages of the traditional first-come, first-served mechanism, ensures that critical events (such as equipment failures) can obtain processing resources first, and avoids low-priority events from occupying system capabilities, thereby improving the overall response speed and resource utilization rate. For example, in a vibration monitoring scenario, if the event urgency score of a device crack is 90 points (highest priority), and the event score of environmental temperature fluctuation is 30 points (lower priority), the system automatically places the blade crack event at the front of the queue, assigns it a higher weight and gives priority to scheduling its data transmission task.

[0060] In addition, the specific steps for dynamically adjusting the bandwidth ratio of the transmission channel are as follows: First, when the event urgency score exceeds the upper limit of the preset urgency threshold interval, the system automatically suspends the occupation of the reserved bandwidth resource channel for low-priority events and reallocates this resource channel for use by the events with the urgency score exceeding the limit; the principle is that the upper limit of the preset urgency threshold interval serves as the watershed for critical events, triggering the resource preemption mechanism. By interrupting the allocation of the reserved bandwidth resource channel for low-priority events, it ensures that high-priority events can exclusively occupy additional channel resources. The beneficial effect is to reduce the processing delay of high-urgency events and prevent the loss of critical data or response lag caused by resource competition. For example, in the event of a sudden change in pressure on a pressure sensor at a critical location, if the urgency score breaks through the upper limit threshold (such as set to 80 points), the system immediately suspends the bandwidth channel for nearby temperature monitoring tasks and instead fully supports the pressure data transmission, thereby accelerating fault diagnosis.

[0061] Second, the system configures a fixed-step adjustment strategy corresponding to the urgency score value range, performs an incremental adjustment of bandwidth resources according to the preset step value matched by the value range, and approves the real-time adjustment amount through a two-way negotiation protocol between nodes; the principle is that the fixed-step adjustment strategy divides the urgency score values into multiple ranges (such as high, medium, low), and different step values are preset for each range (such as a large step value for the high range and a small step value for the low range) to achieve smooth increase and decrease of bandwidth resources. At the same time, the two-way negotiation protocol between nodes ensures that the adjustment amount is mutually approved by adjacent nodes, avoiding network conflicts caused by unilateral decisions. The beneficial effect is to maintain the stability and predictability of bandwidth adjustment, prevent resource oscillation and improve the cooperation efficiency of distributed nodes. For example, in vibration monitoring, if the urgency score of a certain node event falls into the medium range (such as 50 - 70 points), the system increases the bandwidth by 10% according to its preset small step, but needs to negotiate and approve the adjustment amount with adjacent nodes to ensure the network load balance.

[0062] Third, a centralized coordinator is deployed to monitor the network load status. When the remaining bandwidth capacity of a node is lower than the system preset capacity threshold, the coordinator freezes the bandwidth resource application operation of low-priority events, and at the same time allocates the minimum guaranteed bandwidth for events with an urgency score higher than the guarantee threshold; the principle is that the centralized coordinator aggregates the bandwidth usage data of each node in real time. When it detects that the remaining bandwidth capacity is insufficient, it preferentially freezes the application process of low-priority events and forcibly reserves the minimum bandwidth resources for high-score events. The beneficial effect is to globally optimize the network resource allocation, prevent congestion and ensure the baseline transmission capacity of key events. For example, when the remaining bandwidth capacity of all network nodes drops below the preset threshold, the coordinator freezes the new applications of all events with a score lower than 20 points, and at the same time allocates the minimum guaranteed bandwidth for bearing overheating events with a score higher than the guarantee threshold (such as 60 points) to ensure the continuous upload of their data.

[0063] Further, after dynamically adjusting the bandwidth ratio of the transmission channel, it also includes: Compare the event urgency score with a preset threshold. When the score exceeds the threshold, an alarm signal is automatically triggered, and at the same time, the feature vector data of the event is marked as high priority and pushed to the cloud server for storage.

[0064] Specifically, a real-time evaluation of the event urgency score is performed through a preset threshold comparison mechanism. When the score exceeds the preset threshold, the system automatically triggers a multi-level linkage response: first, activate the audible and visual alarm signals to prompt on-site operators. At the same time, mark the anti-interference enhanced feature vector data associated with the event as high priority, and push it to the cloud server through a dedicated transmission channel for persistent storage. The technical principle is to use the threshold determination module pre-configured in the distributed control system to perform real-time comparison on the urgency score output by the dynamic event priority evaluation model. When the score exceeds the dynamic threshold set according to the device safe operation standard, trigger the alarm and data classification processing process.

[0065] In the specific implementation process, the generation of the alarm signal adopts a multi-modal output strategy, which includes both audible alarm signals to attract attention and visual alarm information to accurately describe the event characteristics. The alarm information integrates key parameters such as the start timestamp and duration period recorded by the feature triple, as well as the main frequency component characteristics obtained by spectrum analysis, providing comprehensive diagnostic basis for operation and maintenance personnel. In the data marking and transmission link, the system realizes high-priority marking by modifying the priority flag bit of the data packet header, and dynamically allocates a dedicated transmission channel based on software-defined network technology to ensure the reliable transmission of key data in a congested network environment. After receiving the data, the cloud server stores it in a time-series database with fast retrieval characteristics, and at the same time establishes an associated index with the original sampling data to support subsequent in-depth analysis and traceability. Based on the same inventive concept, as Figure 2 shown, the present invention provides a real-time data acquisition and processing system for a distributed control system, including: A data stream acquisition and probability analysis module 201, configured to continuously acquire a target data stream through sensors deployed on terminal nodes, and use a sliding time window to statistically analyze the probability density distribution of the instantaneous change rate of the target data stream; A dynamic threshold generation module 202, configured to select a preset percentile value of the probability density distribution as the basic sensitivity threshold, calculate a dynamic adjustment coefficient based on the remaining battery power, network load factor, and historical transmission energy consumption of the corresponding device, and generate a target sensitivity threshold by weighted fusion of the basic sensitivity threshold and the dynamic adjustment coefficient; A feature triple extraction module 203, configured to calculate the difference sequence of adjacent sampling points of the instantaneous change rate when any instantaneous change rate exceeds the target sensitivity threshold, and screen out the sections with a change amplitude greater than the amplitude threshold through a preset amplitude threshold; merge the continuously co-directional change points within the section to generate a feature triple including the start timestamp, duration period, and cumulative change amount; The frequency-domain enhancement processing module 204 is used to perform frequency-domain analysis on the target data stream, obtain the effective spectral features of the current data window through the Fourier transform of the sliding window, perform spatio-temporal alignment on the effective spectral features and the feature triple, use the signal-to-noise ratio of each frequency band in the spectral features as the dynamic weight coefficient, and generate an anti-interference enhanced feature vector through weighted fusion; The priority evaluation and modeling module 205 is used to construct a dynamic event priority evaluation model based on the anti-interference enhanced feature vector; the evaluation model includes performing time alignment on the energy entropy of each frequency band in the feature vector and the duration period in the feature triple, calculating the temporal mutation intensity through the ratio of the spectral energy entropy fluctuation amplitude to the duration period; performing a product operation on the temporal mutation intensity and the cumulative change amount to obtain the initial urgency, combining the remaining power of the current node and the network load factor to construct a resource attenuation coefficient, and multiplying the initial urgency by the resource attenuation coefficient to generate an event urgency score; The bandwidth dynamic scheduling module 206 is used to generate a weighted event processing queue according to the descending order of the event urgency scores; dynamically adjust the bandwidth occupancy ratio of the transmission channel according to the event urgency scores of each event in the event processing queue.

[0066] Based on the same inventive concept, the present invention provides an electronic device, as Figure 3 shown, including: a memory 302, a processor 301, and a computer program stored on the memory 302 and executable on the processor 301. When the processor 301 executes the computer program, a real-time data acquisition and processing method for a distributed control system is implemented.

[0067] Based on the same inventive concept, the present invention provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, a real-time data acquisition and processing method for a distributed control system is implemented.

[0068] The above are only the preferred embodiments of the present application and are not intended to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A real-time data acquisition and processing method for a distributed control system, characterized in that, Including: Continuously collect the target data stream through sensors deployed at the terminal nodes, and use a sliding time window to statistically analyze the probability density distribution of the instantaneous change rate of the target data stream; Select the preset percentile value of the probability density distribution as the basic sensitivity threshold, calculate the dynamic adjustment coefficient based on the remaining battery power of the corresponding device, the network load factor, and the historical transmission energy consumption, and generate the target sensitivity threshold through weighted fusion of the basic sensitivity threshold and the dynamic adjustment coefficient; When any instantaneous change rate exceeds the target sensitivity threshold, calculate the difference sequence of adjacent sampling points of the instantaneous change rate, and screen out the sections with a change amplitude greater than the amplitude threshold through the preset amplitude threshold; merge the continuously co-directional change points within the section to generate a feature triple including the start timestamp, the duration period, and the cumulative change amount; Perform frequency domain analysis on the target data stream, obtain the effective spectrum features of the current data window through Fourier transform of the sliding window, align the effective spectrum features and the feature triple in space-time, use the signal-to-noise ratio of each frequency band in the spectrum features as the dynamic weight coefficient, and generate an anti-interference enhanced feature vector through weighted fusion; Based on the anti-interference enhanced feature vector, construct a dynamic event priority evaluation model; the evaluation model includes time-aligning the energy entropy of each frequency band in the feature vector with the duration period in the feature triple, and calculating the time series mutation intensity through the ratio of the fluctuation amplitude of the spectrum energy entropy to the duration period; perform a product operation on the time series mutation intensity and the cumulative change amount to obtain the initial urgency, construct a resource attenuation coefficient by combining the remaining battery power of the current node and the network load factor, and generate an event urgency score by multiplying the initial urgency by the resource attenuation coefficient; Arrange in descending order of the scores of the event urgency scores to generate a weighted event processing queue; dynamically adjust the bandwidth occupancy ratio of the transmission channel according to the urgency scores of each event in the event processing queue.

2. The real-time data acquisition and processing method of a distributed control system according to claim 1, characterized in that Before calculating the difference sequence of adjacent sampling points of the instantaneous change rate, it further includes: Perform sliding average filtering on the sequence of target data to eliminate instantaneous interference.

3. A real-time data acquisition and processing method for a distributed control system according to claim 1, characterized in that, The step of obtaining the effective spectrum features of the current data window through Fourier transform of the sliding window specifically includes: Obtain the spectrum features of the current data window through Fourier transform of the sliding window; Extract the main frequency component amplitude and frequency point distribution in the spectrum features; Compare with the device reference noise map, and separate the inherent noise components of the system through spectral subtraction; Combine the temperature and electromagnetic intensity parameters collected by the environmental sensors, and use the three-level wavelet packet decomposition and dynamic energy ratio detection method to identify the environmental noise frequency band; Generate a noise suppression weight matrix according to the energy distribution ratio of the inherent noise components and the environmental noise frequency band; Perform a dot product operation on the main frequency component and the noise suppression weight matrix, and output the optimized effective spectrum features.

4. The real-time data acquisition and processing method of a distributed control system according to claim 3, characterized in that, The step of obtaining the spectrum features of the current data window through Fourier transform of the sliding window specifically includes: Extend the target data within the current time window forward to cover a preset percentage of the data area at the end of the previous window, and form a new window with data overlap; Perform a smooth transition process on the data regions at the beginning and end of the new window, each accounting for 5%. Perform standard spectral transformation calculations on the processed window data. Based on the frequency components in the previous window spectrum whose energy intensity exceeds the set standard and their distribution positions, identify the frequency components in the current window spectrum with a sudden increase in energy intensity and calibrate their positions. Screen out the frequency components that repeatedly appear at the same frequency positions in at least two consecutive windows and merge them to form the spectral characteristics of the current data window.

5. A real-time data acquisition and processing method for a distributed control system according to claim 1, characterized in that, The step of dynamically adjusting the bandwidth ratio of the transmission channel specifically includes: When the event urgency score exceeds the upper limit of the preset urgency threshold range, automatically suspend the occupancy of the reserved bandwidth resource channel for low-priority events and allocate this resource channel for use by the event with an exceeded urgency score. Configure a fixed-step adjustment strategy corresponding to the urgency score value range, perform an incremental adjustment of the bandwidth resource according to the preset step value matched by the value range, and approve the real-time adjustment amount through a two-way negotiation protocol between nodes. Deploy a centralized coordinator to monitor the network load status. When the remaining bandwidth capacity of a node is lower than the system preset capacity threshold, freeze the bandwidth resource application operation for low-priority events, and at the same time allocate the minimum guaranteed bandwidth for events with an urgency score higher than the guarantee threshold.

6. A real-time data acquisition and processing method for a distributed control system according to claim 1, characterized in that, The formula for generating the corresponding target sensitivity threshold is: Among them, is the inverse function (quantile function) of the probability density distribution function ; is the preset percentile value; is the percentage of the remaining battery power of the current device, is the attenuation coefficient; is the real-time occupancy of the TCP / IP protocol stack buffer, is the total buffer capacity; is the energy consumption sensitivity coefficient, is the average energy consumption per unit data volume in the past 24 hours; is the basic threshold weight, is the dynamic adjustment coefficient weight, satisfying .

7. A real-time data acquisition and processing method for a distributed control system according to claim 1, characterized in that After dynamically adjusting the bandwidth ratio of the transmission channel, it further includes: Compare the event urgency score with the preset threshold. When the score exceeds the threshold, automatically trigger an alarm signal, and at the same time mark the feature vector data of this event as high priority and push it to the cloud server for storage.

8. A real-time data acquisition and processing system for a distributed control system, characterized in that, It includes: A data stream acquisition and probability analysis module, which is used to continuously collect the target data stream through sensors deployed at the terminal nodes, and use a sliding time window to statistically analyze the probability density distribution of the instantaneous change rate of the target data stream. A dynamic threshold generation module, which is used to select a preset percentile value of the probability density distribution as the basic sensitivity threshold, calculate the dynamic adjustment coefficient based on the remaining battery power of the corresponding device, the network load factor, and the historical transmission energy consumption, and generate the target sensitivity threshold by weighted fusion of the basic sensitivity threshold and the dynamic adjustment coefficient. A feature triple extraction module, which is used to calculate the difference sequence of adjacent sampling points of the instantaneous change rate when any instantaneous change rate exceeds the target sensitivity threshold, and screen out the sections with a change amplitude greater than the amplitude threshold through a preset amplitude threshold. Merge the continuously co-directional change points within the section to generate a feature triple including the start timestamp, the duration period, and the cumulative change amount. A frequency domain enhancement processing module, which is used to perform frequency domain analysis on the target data stream, obtain the effective spectral characteristics of the current data window through the Fourier transform of the sliding window, align the effective spectral characteristics and the feature triple in space and time, use the signal-to-noise ratio of each frequency band in the spectral characteristics as the dynamic weight coefficient, and generate an anti-interference enhanced feature vector through weighted fusion. A priority evaluation modeling module, configured to build a dynamic event priority evaluation model based on the anti-interference enhanced feature vector; the evaluation model includes time-aligning the energy entropy of each frequency band in the feature vector with the duration period in the feature triple, and calculating the time-series mutation intensity through the ratio of the spectrum energy entropy fluctuation amplitude to the duration period; performing a product operation on the time-series mutation intensity and the cumulative change amount to obtain an initial urgency, constructing a resource attenuation coefficient by combining the remaining power of the current node and the network load factor, and generating an event urgency score by multiplying the initial urgency by the resource attenuation coefficient. A bandwidth dynamic scheduling module, configured to generate a weighted event processing queue according to the descending order of the event urgency scores; dynamically adjust the bandwidth occupancy ratio of the transmission channel according to the urgency scores of the events in the event processing queue.

9. An electronic device, characterized in that, The electronic device includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, it implements a real-time data acquisition and processing method of a distributed control system according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements a real-time data acquisition and processing method of a distributed control system according to any one of claims 1 to 7.

Images