Noise detection method and device for dual-mode communication network of power system

Abstract

The invention discloses a noise detection method and device for a dual-mode communication network of a power system. The noise detection method comprises the following steps: acquiring an analog signal transmitted by the dual-mode communication network of the power system in a current detection period and filtering the analog signal; sequentially carrying out signal amplification and analog-to-digital conversion on the filtered analog signal according to a preset sequence of a plurality of preset gain values to obtain a digital signal, carrying out validity verification on the digital signal, and terminating after obtaining a first preset gain value which is verified to be valid; and calculating a noise intensity value of the dual-mode communication network of the power system based on the preset gain value which is verified to be valid for the first time. Through orderly traversal of a preset gain sequence and a first effective immediate stop strategy, orderly matching of effective gains is realized, an ADC signal overflow risk is thoroughly avoided, and complete waveform capture in a strong noise environment is ensured; and meanwhile, traversal times are compressed, and detection time consumption is remarkably reduced compared with a traditional multi-stage gain scheme.

Description

Technical Field

[0001] The present invention relates to the technical field of power system communications, and in particular to a noise detection method and device for a dual-mode communication network of a power system. Background Art

[0002] As the core transmission architecture of smart grids, dual-mode communication networks in power systems generally utilize a complementary and integrated communication model: Power Line Carrier (PLC) and Micro Power Wireless (MPW). In the complex electromagnetic environment of distribution networks, broadband noise generated by switching operations, frequency conversion equipment, and lightning surges can severely degrade communication quality, leading to a surge in bit error rates and even link interruptions. According to measured data based on the Q / GDW 1866-2019 standard, noise levels in 10kV distribution cables can reach over 90dBμV, with strong bursts and a wide spectral dynamic range (3kHz-2MHz).

[0003] The industry currently generally adopts a fixed-gain noise detection scheme, which collects channel signals during the communication frame interval, amplifies them at a fixed amplification factor, and then feeds them into the analog-to-digital converter (ADC). However, this scheme has a significant drawback: in a strong noise environment, the signal amplitude can easily exceed the ADC range, causing saturation and distortion of the sampled data. For example, when the instantaneous amplitude of the surge noise reaches ±5V, if the ADC reference voltage is ±2.5V and the gain is fixed at 40dB, the output signal will lose its true noise characteristics due to hard saturation. More seriously, in the event of a communication failure (such as a slave node response timeout), the traditional frame interval acquisition method cannot obtain effective noise data due to the continuous channel occupancy, resulting in a lack of fault diagnosis information.

[0004] To mitigate saturation distortion, some improvement solutions have attempted to employ multi-level gain switching strategies. However, these solutions require traversing all gain levels to determine the optimal value, resulting in a single detection time of over 15ms, which fails to meet the 10ms real-time requirement specified in DL / T 698.45. Furthermore, single-node detection results are susceptible to local interference. Substation switching operations may generate transient spike noise at only a single node. If this triggers a network-wide communication mode switch, it will result in unnecessary bandwidth loss. Summary of the Invention

[0005] The purpose of the embodiments of the present invention is to provide a noise detection method and device for a dual-mode communication network of an electric power system. By orderly traversing a preset gain sequence and adopting a first-validity stopping strategy, the effective gain is matched in an orderly manner, the risk of ADC signal overflow is completely avoided, and the complete capture of the waveform in a strong noise environment is ensured. At the same time, the number of traversals is compressed, and the detection time is significantly reduced from >15ms in the traditional multi-level gain solution to within 5ms, fully meeting the 10ms real-time requirement of the DL / T 698.45 standard, and solving the long-standing coupling problem of saturation distortion and insufficient real-time performance in the industry.

[0006] To solve the above technical problems, a first aspect of an embodiment of the present invention provides a noise detection method for a dual-mode communication network of a power system, comprising the following steps:

[0007] Acquire and filter the analog signal transmitted by the power system dual-mode communication network in the current detection cycle;

[0008] Amplifying and analog-to-digital converting the filtered analog signal in a predetermined order of a plurality of predetermined gain values ​​to obtain a digital signal, and verifying the validity of the digital signal, and terminating after obtaining the first predetermined gain value verified to be valid;

[0009] Based on the first preset gain value verified to be valid, a noise intensity value of the power system dual-mode communication network is calculated.

[0010] Furthermore, the performing validity verification on the digital signal includes:

[0011] Acquire a valid data interval in the digital signal, where the valid data interval is a continuous sampling point of a preset length;

[0012] Determining whether a first preset number of consecutive sampling points in the valid data interval reaches a saturation threshold of the digital-to-analog converter;

[0013] If yes, determining that the preset gain value corresponding to the digital signal is an invalid value;

[0014] If not, it is determined that the preset gain value corresponding to the digital signal is a valid value, and the noise intensity value is calculated based on the valid preset gain value.

[0015] Furthermore, the determining whether a first preset number of consecutive sampling points in the valid data interval reaches a saturation threshold of the digital-to-analog converter includes:

[0016] Scanning each of the sampling points in the valid data interval in sequence, and when the sampling point values ​​are continuously equal to the maximum positive value of the digital-to-analog converter or continuously equal to the maximum negative value of the digital-to-analog converter, adding 1 to the counting variable; otherwise, clearing the counting variable;

[0017] When the counting variable reaches a first preset number, determining that there are a first preset number of consecutive sampling points that reach a saturation threshold of the digital-to-analog converter;

[0018] When the count variable does not reach the first preset number at the end of the scan, it is determined that there are no consecutive first preset number of sampling points that reach the saturation threshold of the digital-to-analog converter.

[0019] Furthermore, when the sampling point value is continuously equal to the maximum positive value of the digital-to-analog converter or continuously equal to the maximum negative value of the digital-to-analog converter, adding 1 to the count variable includes:

[0020] If the current sampling point value is equal to the maximum positive value, and the previous point value is equal to the maximum positive value, then the counting variable is increased by 1;

[0021] If the current sampling point value is equal to the maximum negative value, and the previous point value is equal to the maximum negative value, the counting variable is increased by 1.

[0022] After further verifying the validity of the digital signal, the method further includes:

[0023] When no valid preset gain value is obtained after traversing all preset gain values, the preset gain value with the smallest value among the plurality of preset gain values ​​is taken as the valid preset gain value;

[0024] The noise intensity value is calculated based on the preset gain value effective value.

[0025] Furthermore, the step of amplifying and converting the filtered analog signal into analog-to-digital signals in a preset order of a plurality of preset gain values ​​to obtain a digital signal includes:

[0026] The filtered analog signal is amplified and analog-to-digital converted in descending order according to a number of preset gain values ​​to obtain a digital signal.

[0027] Furthermore, before obtaining and filtering the analog signal transmitted by the power system dual-mode communication network in the current detection cycle, the method includes:

[0028] When the dual-mode communication network master node fails to receive a response message from the dual-mode communication network slave node within a preset response timeout threshold after sending a communication message, the analog signal is acquired during the communication interruption period and a noise detection process is started.

[0029] Furthermore, after calculating the noise intensity value of the power system dual-mode communication network, the method further includes:

[0030] The noise intensity values ​​calculated by multiple detection nodes in the same detection cycle are analyzed in time and space to generate a comprehensive noise intensity value.

[0031] Furthermore, performing spatiotemporal correlation analysis on noise intensity values ​​calculated by multiple detection nodes within the same detection cycle includes:

[0032] Aligning the noise intensity values ​​of the multiple detection nodes according to the acquisition time;

[0033] Calculating a spatial correlation coefficient between the aligned plurality of noise intensity values;

[0034] If the spatial correlation coefficient is greater than a preset spatial correlation threshold, the weighted average of the noise intensity values ​​of the at least two detection nodes is used as the comprehensive noise intensity value; otherwise, the noise intensity value of the current detection cycle is discarded.

[0035] Accordingly, a second aspect of an embodiment of the present invention provides a noise detection device for a dual-mode communication network of a power system, which performs noise detection on the dual-mode communication network of the power system based on the above-mentioned noise detection method for the dual-mode communication network of the power system, including:

[0036] A signal acquisition module is used to acquire and filter the analog signal transmitted by the power system dual-mode communication network in the current detection cycle;

[0037] a signal processing module configured to amplify and convert the filtered analog signal into digital signals in a predetermined order of a plurality of predetermined gain values, and to verify the validity of the digital signal, terminating after obtaining the first predetermined gain value verified to be valid;

[0038] The noise calculation module is used to calculate the noise intensity value of the power system dual-mode communication network based on the first preset gain value verified to be valid.

[0039] Accordingly, a third aspect of an embodiment of the present invention provides an electronic device, comprising: at least one processor; and a memory connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute a noise detection method for a dual-mode communication network of a power system.

[0040] Accordingly, a fourth aspect of an embodiment of the present invention provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the above-mentioned noise detection method for the dual-mode communication network of the power system.

[0041] The above technical solutions of the embodiments of the present invention have the following beneficial technical effects:

[0042] 1. By sequentially traversing a preset descending gain sequence and adopting a "first-valid-stop" strategy, combined with precise verification of the ADC digital signal validity, this method achieves adaptive noise detection in complex electromagnetic environments, effectively avoiding the risk of ADC signal saturation and distortion caused by strong noise and ensuring complete capture of the noise waveform. Furthermore, the upper limit of the number of gain traversals is significantly reduced to a preset number, significantly reducing the time required for a single detection from over 15ms in traditional multi-level traversal schemes to under 5ms, fully meeting the stringent real-time noise detection requirements of smart grid communications (such as those in the DL / T 698.45 standard).

[0043] 2. To address the critical flaw of traditional frame-interval collection methods, which fail when the communication link is interrupted or the channel is continuously occupied, the noise collection trigger is set to occur during the communication response timeout interruption period. When the master node fails to receive a response from the slave node within the preset time, the noise collection process is initiated during the window period when the channel is inevitably idle. This ensures that even in the event of a complete communication failure, noise data reflecting the true channel conditions can still be reliably obtained. This solves the industry pain point that traditional methods cannot provide effective noise information at the most critical moment of fault diagnosis, significantly improving the completeness and reliability of fault diagnosis.

[0044] 3. Based on single-point precision detection, the noise intensity values ​​of multiple distributed nodes within the same detection cycle are further temporally and spatially correlated. By strictly aligning the data of each node according to the acquisition time and calculating their spatial correlation coefficient, abnormal noise data generated by local instantaneous interference (such as the operation of a single substation switch) can be intelligently identified and eliminated. Only when the noise of multiple nodes shows significant spatial correlation, their weighted average is used as the comprehensive noise intensity value reflecting the actual noise conditions of the entire network or region within that period. This effectively reduces the risk of false alarms in single-point detection, significantly improves the accuracy of noise source positioning, and significantly improves the reliability of network status assessment and decision-making based on noise data. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a flow chart of a noise detection method for a power system dual-mode communication network provided by an embodiment of the present invention;

[0046] Figure 2 is a signal processing flow chart provided by an embodiment of the present invention;

[0047] Figure 3a This is a schematic diagram of excessive gain of an analog amplifier provided by an embodiment of the present invention;

[0048] Figure 3b This is a schematic diagram of an analog amplifier with too low gain according to an embodiment of the present invention;

[0049] Figure 4This is a schematic diagram of sampling point traversal provided by an embodiment of the present invention;

[0050] Figure 5 1 is a schematic diagram of analog signal sampling timing provided by an embodiment of the present invention;

[0051] Figure 6a This is an example of the average noise level at the same location on different dates provided by the embodiment of the present invention. Figure 1 ;

[0052] Figure 6b This is an example of the average noise level at the same location on different dates provided by the embodiment of the present invention. Figure 2 ;

[0053] Figure 6c Schematic diagram 3 of average noise levels at the same location on different dates provided by an embodiment of the present invention;

[0054] Figure 6d This is an example of the average noise level at the same location on different dates provided by the embodiment of the present invention. Figure 4 ;

[0055] Figure 6e This is an example of the average noise level at the same location on different dates provided by the embodiment of the present invention. Figure 5 ;

[0056] Figure 6f 6 is a schematic diagram of average noise levels at the same location on different dates provided by an embodiment of the present invention;

[0057] Figure 7a This is an example of the average noise level at different locations at the same time provided by the embodiment of the present invention. Figure 1 ;

[0058] Figure 7b This is an example of the average noise level at different locations at the same time provided by the embodiment of the present invention. Figure 2 ;

[0059] Figure 8 This is a schematic diagram of a node failure provided by an embodiment of the present invention;

[0060] Figure 9 This is a block diagram of a noise detection device module for a dual-mode communication network of an electric power system provided by an embodiment of the present invention.

[0061] Reference numerals:

[0062] 1. Signal acquisition module, 2. Signal processing module, 3. Noise calculation module. DETAILED DESCRIPTION

[0063] To make the objectives, technical solutions, and advantages of the present invention more clearly understood, the present invention will be further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present invention.

[0064] Please refer to Figure 1 and Figure 2 A first aspect of an embodiment of the present invention provides a noise detection method for a dual-mode communication network of a power system, comprising the following steps:

[0065] S100: Acquire and filter an analog signal transmitted by the power system dual-mode communication network in a current detection cycle.

[0066] First, the original analog signal transmitted by the power system's dual-mode communication network (PLC+MPW) during the current detection cycle is acquired. This signal contains both communication messages and background noise. To focus on the target noise frequency band (780kHz-12MHz), the signal is preprocessed through a hardware-implemented bandpass filter. This bandpass filter's primary function is to filter out unwanted signals outside the target frequency band (such as low-frequency power interference and ultra-high frequency wireless signals), allowing only signals within the 780kHz to 12MHz band to pass through. This provides a preliminary cleansed analog signal foundation for subsequent accurate noise detection.

[0067] S200, amplifying and converting the filtered analog signal into analog-to-digital signals in a preset order of several preset gain values ​​to obtain a digital signal, and verifying the validity of the digital signal. The process ends after obtaining the first preset gain value verified to be valid.

[0068] Alternatively, a specific descending gain sequence (42dB, 26dB, 10dB, -6dB) is preset. The processing flow is as follows: First, the analog signal filtered by S100 is applied to the analog amplifier for amplification in descending order of preset gain values ​​(from highest to lowest). After each gain value is applied, the amplified signal is immediately sampled and converted by a 12-bit signed analog-to-digital converter (ADC), acquiring a digital signal of 16,384 consecutive sample points in a single acquisition. After the conversion is complete, the digital signal is immediately validated: the algorithm scans the valid data interval of the digital signal (sample point indexes 2048 to 10,240, avoiding the areas at the beginning and end where distortion may occur) to see whether five consecutive sample points reach the ADC's saturation threshold (positive saturation value +2047 or negative saturation value -2048). If five consecutive saturation points exist, the current gain value is deemed invalid and automatically switches to the next (lower) gain value in the sequence, repeating the amplification, conversion, and validation process. Once five consecutive saturation points are not detected, the current gain value is determined to be valid and the calculation process of the entire gain sequence is terminated immediately.

[0069] S300 , calculating a noise intensity value of the power system dual-mode communication network based on a first preset gain value verified to be valid.

[0070] Based on the effective gain value determined in S200 and the corresponding digital signal data, the voltage sampling values ​​of the valid data interval (sampling points 2048 to 10240) are first extracted. The voltage root mean square is obtained by calculating the square sum of all voltage values ​​in the interval, dividing it by the number of valid points, and then taking the square root. The voltage root mean square is then converted into a single noise intensity value in dBm by combining the effective gain value with the preset calibration constant. Within a single detection cycle, the system performs multiple single noise acquisition calculations, converts all decibel values ​​into linear energy values, accumulates and averages them, and finally converts the energy average back into a decibel value, thereby outputting an average noise intensity value that accurately reflects the comprehensive noise energy level within the cycle.

[0071] In the technical solution of the present application, the process of analog signal acquisition, amplification and analog-to-digital conversion is as follows: the analog signal in the dual-mode communication network of the power system first enters the 780K-12MHz bandpass filter to filter out interference signals outside the frequency band; the retained valid frequency band signal is then input into the analog amplifier, which uses a preset gain sequence for dynamic gain control, applies gain values ​​in sequence to amplify the signal, and immediately converts it into a digital signal through a 12-bit signed analog-to-digital converter (ADC) (16384 points are collected at a time); the converted digital signal needs to undergo validity verification (check whether there are 5 consecutive points in the valid data interval [2048,10240] that reach the ADC saturation threshold ±2048). Once the verification is passed, the gain traversal is terminated, and the current gain value is used as the effective gain to participate in the subsequent noise intensity calculation, thereby realizing accurate capture and anti-saturation processing of broadband noise at the hardware level.

[0072] Please refer to Figure 3a and Figure 3b The gain of the analog amplifier affects the true validity of the signal. Excessive gain will cause the amplified signal to exceed the valid range of the 12-bit signed digital number, resulting in signal distortion. Figure 3a As shown; too small a gain will also make the signal smaller, the waveform unclear, and the error in calculation increase, such as Figure 3b shown.

[0073] Furthermore, the validity verification of the digital signal in S200 includes:

[0074] S210 , obtaining a valid data interval in the digital signal, where the valid data interval is a continuous sampling point of a preset length.

[0075] After acquiring the analog-to-digital converted digital signal (16,384 sampling points in a single acquisition), the system first excludes the signal's beginning and end, which may be distorted by hardware transient response, and locks in a preset valid data interval (sampling points 2048 to 10,240). This interval contains 8,192 consecutive sampling points, selected based on empirical evidence from hardware circuit characteristics: the first 2,047 and last 6,144 points of the original signal are susceptible to filter transient effects, while the mid-sampling points best reflect the noise waveform characteristics, providing a reliable data foundation for subsequent validity verification.

[0076] S220 , determining whether a first preset number of consecutive sampling points in the valid data interval reaches a saturation threshold of the digital-to-analog converter.

[0077] The valid data interval is scanned point by point: After initializing the counter, each sampling point is sequentially determined to see if it reaches the saturation threshold (+2047 or -2048) of the 12-bit ADC. If the current sampling point and the previous point have the same positive saturation value (+2047) or negative saturation value (-2048), the counter increments by 1. If the point value is not saturated or the positive or negative saturation state switches, the counter is cleared. Once the counter reaches the first preset number (five consecutive points), the digital signal at the current gain is immediately determined to be invalid due to the amplitude exceeding the ADC range. If the entire scan interval still does not trigger saturation at five consecutive points, the signal is considered valid. This accurately identifies waveform clipping caused by excessive gain (for example, hard saturation of the signal to a square wave under strong noise), preventing distorted data from being included in noise calculations.

[0078] By combining effective data interval focusing with continuous saturation point dynamic detection, an anti-distortion barrier is established in the analog-to-digital conversion link of the signal chain: it not only avoids the problem of waveform feature loss that is inevitable in traditional fixed-gain solutions under strong noise, but also ensures the integrity of the noise waveform and data authenticity through hardware-adaptive interval selection (avoiding head and tail distortion) and software-efficient saturation pattern recognition (continuous 5-point threshold judgment); at the same time, the computational complexity is extremely low, and a single scan can be completed in microseconds, providing a core guarantee for the real-time performance of gain adaptive control (≤5ms).

[0079] The determination of the first preset number (i.e., the threshold for determining the number of consecutive saturation points) is based on a comprehensive consideration of hardware constraints and experimental verification results. At the hardware level, a 12-bit signed analog-to-digital converter (ADC) with a value range of -2048 to +2047 is used. The saturation threshold is strictly defined as the positive limit of +2047 (binary 0111 1111 1111) and the negative limit of -2048 (binary 1000 00000000). Experimentally, by injecting noise test data ranging from -55dBm to +10dBm (see the narrowband / Gaussian noise comparison table on Page 4-5), it was verified that saturation of five consecutive sampling points can effectively distinguish between true hardware overflow and transient interference. Fewer than five consecutive sampling points are susceptible to pulse noise misjudgment, while five consecutive saturation points can stably reflect signal distortion caused by excessive gain while ensuring detection reliability over a wide noise range. Ultimately, this value was established as the optimal technical choice.

[0080] S230: If yes, determine that the preset gain value corresponding to the digital signal is an invalid value.

[0081] If S220 detects that five consecutive sampling points within the valid data interval reach the ADC saturation threshold (+2047 or -2048), the current gain value is deemed invalid. This indicates that the signal amplification factor is too high, resulting in hard clipping of the waveform (displaying a plateau-like distortion). At this point, the collected data does not reflect the true noise characteristics. The system then switches to the next lowest gain value in the preset descending gain sequence and repeats the signal amplification, analog-to-digital conversion, and validity verification process until a valid gain is found or all preset gain values ​​have been exhausted.

[0082] S240: If not, determine that the preset gain value corresponding to the digital signal is a valid value, and calculate the noise intensity value based on the valid preset gain value.

[0083] If S220 scans the entire valid data interval (8192 points) without triggering saturation at five consecutive points, the current gain value is considered valid. This indicates that the signal amplification factor is adapted to the current noise intensity, the waveform is not clipped or distorted, and the digital signal fully retains the noise's time-domain characteristics. Subsequent iterations of the gain sequence are immediately terminated, and the current gain value is locked and output as the valid gain value (G1). Simultaneously, the digital signal data collected at this gain is passed to the noise calculation module (S300) as the sole valid input source for noise intensity calculation.

[0084] Through the rapid elimination of invalid gains and the instant locking output mechanism of effective gains, precise control of the core links of noise detection is achieved. On the one hand, based on the hardware-quantifiable saturation characteristics (5 consecutive points of clipping), distorted data is actively excluded to ensure that only undistorted waveforms are involved in the calculation; on the other hand, through the "first effective" strategy (not full traversal), the number of invalid attempts is compressed to a minimum (≤3 times). Under the premise of ensuring data authenticity, the gain matching time is compressed to within 5ms, overcoming the technical contradiction of "fidelity" and "real-time" in traditional solutions.

[0085] For further information, please refer to Figure 4 The step S220 of determining whether there are a first preset number of consecutive sampling points in the valid data interval that reach the saturation threshold of the digital-to-analog converter includes:

[0086] S221, scan each sampling point in the valid data interval in sequence, when the sampling point value is continuously equal to the maximum positive value of the digital-to-analog converter or continuously equal to the maximum negative value of the digital-to-analog converter, add 1 to the counting variable, otherwise clear the counting variable.

[0087] After initializing the clear counter, it scans point by point from the start of the valid data range (2048-10240 points). If both the current sampling point and the previous point reach the positive saturation threshold (+2047) or the negative saturation threshold (-2048), the counter increments by 1. If the current point is not saturated, switches between positive and negative saturation states, or is inconsistent with the saturation state of the previous point, the counter is immediately cleared. This mechanism dynamically tracks the number of consecutive saturation points in the same direction to accurately capture waveform clipping distortion characteristics (for example, when a signal is hard-limited to a flat-top waveform under strong noise).

[0088] S222: When the counting variable reaches a first preset number, determine that there are a first preset number of consecutive sampling points that reach a saturation threshold of the digital-to-analog converter.

[0089] When the counter reaches the first preset number (five consecutive points) during a scan, it is determined that persistent saturation distortion exists at the current gain. This threshold is based on actual hardware measurements: when the gain is too high, the ADC output waveform typically exhibits a plateau-like saturation pattern for >5 consecutive sampling points, far exceeding the transient spikes characteristic of normal noise (usually ≤3 points). If this determination is established, the current gain is marked as invalid, and the gain switching process is triggered.

[0090] S223 , when the counting variable does not reach the first preset number at the end of the scan, determining that there are no consecutive first preset number of sampling points that reach the saturation threshold of the digital-to-analog converter.

[0091] If the counter never reaches 5 after scanning the entire valid data interval (8192 points) (for example, the highest continuous saturation point is 4 or less), it is determined that there is no continuous saturation distortion at the current gain, indicating that the signal amplitude falls completely within the ADC range (-2048 to +2047) and the waveform characteristics are not destroyed. After confirming that the signal is valid, the scan is immediately terminated and the current gain value is locked.

[0092] Through point-by-point state comparison and a dynamic count clearing mechanism, this solution achieves reliable identification of saturation distortion with ultra-low computational complexity: it uses five consecutive points of saturation in the same direction as the judgment threshold, which can not only avoid misjudgment of transient spike interference (such as lightning pulses <5 points), but also accurately capture continuous waveform clipping distortion caused by excessive gain. Combined with the hardware-adapted effective data interval selection, a single scan takes only tens of microseconds, providing underlying guarantees for the real-time performance of gain adaptive control (≤5ms) and data fidelity in strong noise environments.

[0093] Furthermore, in S221, when the sampling point value is continuously equal to the maximum positive value of the digital-to-analog converter or continuously equal to the maximum negative value of the digital-to-analog converter, the counting variable is increased by 1, including: if the current sampling point value is equal to the maximum positive value and the previous point value is equal to the maximum positive value, then the counting variable is increased by 1; if the current sampling point value is equal to the maximum negative value and the previous point value is equal to the maximum negative value, then the counting variable is increased by 1.

[0094] In step S221, continuous saturation point detection uses same-direction continuity as the core judgment logic: the counter variable is accumulated only when the current sampling point and the previous point are both in a positive saturation state (both +2047) or a negative saturation state (both -2048). Conversely, if the saturation state of the current point is inconsistent with the previous point (such as a positive-negative switch) or if either point does not reach the saturation threshold, the counter variable is immediately cleared. This mechanism accurately distinguishes continuous clipping distortion (continuous same-direction saturation) from transient spike interference (isolated saturation points or positive-negative jumps), providing a reliable basis for gain validity judgment.

[0095] Furthermore, after verifying the validity of the digital signal in S200, the following steps are further included:

[0096] S251 : When no valid preset gain value is obtained after traversing all preset gain values, a preset gain value with the smallest value among the preset gain values ​​is used as a valid preset gain value.

[0097] After iterating through all preset gain values ​​in descending order (42 / 26 / 10 / -6dB), if each verification is invalid due to detecting five consecutive points of saturation (i.e., all gains cause signal distortion), the minimum gain value in the sequence (-6dB) is forcibly locked as the valid value. Based on actual hardware measurements, in extremely noisy environments, the minimum gain can maximize signal amplitude suppression, increasing the probability that the signal will fall within the ADC range (-2048 to +2047), thus avoiding the problem of continuous saturation causing the complete inability to obtain waveform data.

[0098] S252: Calculate the noise intensity value based on the preset gain value effective value.

[0099] Based on the mandatory minimum gain value (-6dB), the noise calculation process is identical to that for conventional effective gain: voltage samples are extracted from the valid data interval (2048-10240 points), the voltage RMS is calculated, and then converted to a dBm-level single-shot noise intensity value using a preset linear-logarithmic conversion relationship (including fixed calibration constants). Although the signal may still be slightly saturated or compressed, this mechanism ensures that even in extreme scenarios where traditional solutions completely fail, noise intensity data with engineering reference value can still be output.

[0100] Furthermore, before obtaining and filtering the analog signal transmitted by the power system dual-mode communication network in the current detection cycle in S100, the following steps are included:

[0101] S101, when the dual-mode communication network master node fails to receive a response message from the dual-mode communication network slave node within a preset response timeout threshold after sending a communication message, an analog signal is obtained during the communication interruption period and a noise detection process is started.

[0102] In the operation of a dual-mode communication network in a power system, when a master node sends a communication message to a slave node, if no valid response is received within a preset response timeout threshold (e.g., 400ms, set according to the communication protocol standard), the current communication link is determined to be in an interrupted state. At this point, the master node immediately activates the noise detection process: utilizing the window period during which the channel must be idle during the communication interruption, the analog signal acquisition module (such as the coupler and filter chain) is started to directly obtain the original analog signal of the current channel from the physical layer. This design subverts the traditional model of relying on communication frame intervals to collect noise, and transforms the communication failure event itself into a trigger condition for noise detection. Through a noise collection mechanism driven by communication interruption events, the key defect of traditional solutions that cannot obtain noise data at the most critical moment of fault diagnosis is overcome: when the communication link is interrupted or the channel is continuously occupied, the response timeout window period (at this time the channel is 100% idle) is intelligently utilized to force the start of high-fidelity noise collection. This ensures that even under extreme conditions where communication is completely paralyzed and the traditional frame interval method completely fails, noise data reflecting the actual channel degradation state can still be captured, providing direct evidence for root cause analysis of faults (such as strong noise causing communication interruption), significantly improving the integrity and accuracy of power grid fault diagnosis.

[0103] Please refer to Figure 5 In power networks, data packets are separated by interframe intervals. By combining the valid data detection interrupt processing function in the carrier chip's physical layer core with data acquisition within the interframe interval, we can effectively avoid collecting communication message signals. However, when noise levels are high or a fault occurs, valid data message signals are no longer received, making it impossible to capture noise signals within the interframe intervals. To avoid interference from communication frames, in all environments, we choose to capture noise signals during the timeout interrupt when no valid message is detected.

[0104] Furthermore, after calculating the noise intensity value of the power system dual-mode communication network in S300, the method further includes:

[0105] S400 , performing spatiotemporal correlation analysis on noise intensity values ​​calculated by multiple detection nodes in the same detection cycle to generate a comprehensive noise intensity value.

[0106] After calculating the noise intensity value for a single node, the system automatically triggers a multi-node collaborative analysis mechanism. First, multiple distributed noise detection nodes (such as meter readers) upload the noise intensity values ​​calculated within the same detection cycle to a cloud server. When the server performs spatiotemporal correlation analysis, it specifically involves three-dimensional processing: In the spatial dimension, by comparing the noise intensity differences between geographically adjacent nodes (e.g., -30dBm at monitoring point 1 and -45dBm at monitoring point 2), a noise intensity gradient model is established in conjunction with the node location topology map to identify the node closest to the noise source. In the temporal dimension, periodic data is extracted from each node's historical database (e.g., noise values ​​from 21:00-22:00 for six consecutive days), and scatter plot cluster analysis is used to identify recurring noise patterns (e.g., daily peak periods). In the event dimension, periods of sudden noise increases (e.g., a -15dBm outlier at 12:00) are matched with grid operation logs to correlate specific device start and stop events (e.g., charging station operation records). To generate the final comprehensive noise intensity value, three types of features are weighted and fused: the current period's spatial gradient is weighted at 50%, the historical noise mean is weighted at 30%, and the device-associated confidence is weighted at 20%. The output is a comprehensive value with spatial coordinate and time tags (such as "coordinate X / Y 12:00 comprehensive noise -28dBm"). This overcomes the core flaws of existing technologies, such as rough manual positioning and the inability to cover irregular noise sources. Through multi-node data collaboration and historical pattern mining, it achieves automatic and precise positioning of noise sources (such as identifying the water heater closest to monitoring point 1) and intelligent attribution of noise events (such as associating with peak kitchen electricity consumption). This allows operation and maintenance personnel to obtain a map of the spatiotemporal evolution of noise without on-site supervision, significantly improving the grid's fault prevention capabilities.

[0107] Furthermore, performing spatiotemporal correlation analysis on the noise intensity values ​​calculated by multiple detection nodes in the same detection cycle in S400 includes:

[0108] S410: Align the noise intensity values ​​of multiple detection nodes according to the acquisition time.

[0109] First, the noise intensity values ​​collected by multiple distributed detection nodes (such as meter readers) within the same detection cycle are timestamp-calibrated. Specifically, the local clock of each node is synchronized with the server clock through the 4G communication module to ensure that the noise intensity value collection time deviation of all nodes is less than 10 milliseconds. Then, the time tags in the data reported by each node are extracted, and the preset detection cycle (for example, 1 minute) is used as the time window, and the noise values ​​of all nodes in the window are arranged in ascending order according to the collection time. For time-offset data caused by network delays, linear interpolation is used to align them to standard sampling points on a unified time axis (such as the 0th, 20th, and 40th seconds of each minute) to form a spatiotemporally aligned noise intensity value matrix.

[0110] S420: Calculate the spatial correlation coefficient between the aligned noise intensity values.

[0111] Based on the aligned noise intensity value matrix, the spatial correlation coefficient is calculated. First, a node position topology map is constructed, and the spatial attenuation weight is calculated based on the distance between nodes: the weight decays by 20% for every 100 meters increase in the distance between adjacent nodes. Then, the improved Pearson correlation coefficient formula is used to introduce a weight factor to calculate the covariance of the noise values ​​of each pair of nodes. In the specific calculation, single-point sudden increase outliers are first excluded, and then a sliding window correlation analysis is performed on the noise intensity sequence of valid node pairs (with a spacing of less than 500 meters). The window width is 3 consecutive sampling points (i.e., 3 samples within 1 minute), and finally the average spatial correlation coefficient of all valid node pairs is output.

[0112] S430: If the spatial correlation coefficient is greater than a preset spatial correlation threshold, a weighted average of the noise intensity values ​​of at least two detection nodes is used as a comprehensive noise intensity value. Otherwise, the noise intensity value of the current detection cycle is discarded.

[0113] When the spatial correlation coefficient is greater than the preset threshold of 0.7 (the threshold can be verified and determined through historical data), the current noise event is determined to be wide-area interference. At this time, the node noise value is weighted and averaged based on the node location weight (the closer to the noise source, the higher the weight) and the noise intensity confidence (for example, the confidence of the value that passes the gain validity check is increased by 15%): weight factor = location weight × confidence coefficient. The final output is the comprehensive noise intensity value with geographical tags. If the correlation coefficient is ≤0.7, the data of that period is discarded and the re-collection mechanism is triggered - the collection time is automatically extended by 20% through the meter reading controller to avoid the loss of valid data due to transient interference (such as switching power supply pulses).

[0114] This implementation solves the problem of rough noise source location caused by a lack of multi-site coordination. It eliminates differences in acquisition timing between nodes through spatiotemporal alignment, combines spatial correlation analysis to distinguish local interference from wide-area noise events, and uses weighted averaging to generate a highly reliable composite value. Operations and maintenance personnel can directly locate the noise source based on this composite value, eliminating the need to manually compare data from multiple sites, improving troubleshooting efficiency by more than three times.

[0115] Please refer to Figure 6a-6f , the scatter plot of historical data obtained from March 19-24 at the same monitoring location, Figure 6a-6fAnalysis shows that the noise values ​​at the monitoring point have been ranging from -15 to -32 for six consecutive days, mostly staying around -30dBm. There is an increase between 21:00 and 22:00, and a sudden increase around 12:00. The noise point graph is stable and similar. The noise between 21:00 and 22:00 is speculated to be caused by the water heater or other charger equipment; the noise around 12:00 may be caused by kitchen electrical equipment. Please refer to Figure 7a and Figure 7b Comparing the data of the two monitoring stations on March 20, it can be seen that the noise data of monitoring point 2 is lower than that of monitoring point 1 as a whole, indicating that monitoring point 1 is closer to the noise source.

[0116] Please refer to Figure 8 The fault sites are numbered 4, 6, 7, 9, 11, and 13. The curve data for each site are missing for several moments, which may be due to the noise generated by equipment that is started at irregular intervals. Monitoring equipment is installed near each site to collect noise data, save the data, and upload it to the host computer to form a historical data graph, which solves the problem of operation and maintenance personnel being on duty at each site and manually recording the dynamic changes in noise data.

[0117] Accordingly, please refer to Figure 9 A second aspect of an embodiment of the present invention provides a noise detection device for a dual-mode communication network of a power system, which performs noise detection on the dual-mode communication network of the power system based on the above-mentioned noise detection method for the dual-mode communication network of the power system, including:

[0118] Signal acquisition module 1, which is used to acquire and filter the analog signal transmitted by the power system dual-mode communication network in the current detection cycle;

[0119] Signal processing module 2, which is used to amplify and convert the filtered analog signal into digital signals in a preset order of several preset gain values, and verify the validity of the digital signal. The traversal is terminated after obtaining the first preset gain value verified to be valid;

[0120] The noise calculation module 3 is used to calculate the noise intensity value of the power system dual-mode communication network based on the first preset gain value verified to be valid.

[0121] Accordingly, a third aspect of an embodiment of the present invention provides an electronic device comprising: at least one processor; and a memory connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor executes a noise detection method for a dual-mode communication network of a power system.

[0122] Accordingly, a fourth aspect of an embodiment of the present invention provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the above-mentioned noise detection method for the dual-mode communication network of the power system.

[0123] The embodiment of the present invention aims to provide a noise detection method for a dual-mode communication network of an electric power system, which has the following effects:

[0124] 1. By sequentially traversing a preset descending gain sequence and adopting a "first-valid-stop" strategy, combined with precise verification of the ADC digital signal validity, this method achieves adaptive noise detection in complex electromagnetic environments, effectively avoiding the risk of ADC signal saturation and distortion caused by strong noise and ensuring complete capture of the noise waveform. Furthermore, the upper limit of the number of gain traversals is significantly reduced to a preset number, significantly reducing the time required for a single detection from over 15ms in traditional multi-level traversal schemes to under 5ms, fully meeting the stringent real-time noise detection requirements of smart grid communications (such as those in the DL / T 698.45 standard).

[0125] 2. To address the critical flaw of traditional frame-interval collection methods, which fail when the communication link is interrupted or the channel is continuously occupied, the noise collection trigger is set to occur during the communication response timeout interruption period. When the master node fails to receive a response from the slave node within the preset time, the noise collection process is initiated during the window period when the channel is inevitably idle. This ensures that even in the event of a complete communication failure, noise data reflecting the true channel conditions can still be reliably obtained. This solves the industry pain point that traditional methods cannot provide effective noise information at the most critical moment of fault diagnosis, significantly improving the completeness and reliability of fault diagnosis.

[0126] 3. Based on single-point precision detection, the noise intensity values ​​of multiple distributed nodes within the same detection cycle are further temporally and spatially correlated. By strictly aligning the data of each node according to the acquisition time and calculating their spatial correlation coefficient, abnormal noise data generated by local instantaneous interference (such as the operation of a single substation switch) can be intelligently identified and eliminated. Only when the noise of multiple nodes shows significant spatial correlation, their weighted average is used as the comprehensive noise intensity value reflecting the actual noise conditions of the entire network or region within that period. This effectively reduces the risk of false alarms in single-point detection, significantly improves the accuracy of noise source positioning, and significantly improves the reliability of network status assessment and decision-making based on noise data.

[0127] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0128] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0129] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A noise detection method for a dual-mode communication network of an electric power system, characterized in that: The steps include: Acquire and filter the analog signal transmitted by the power system dual-mode communication network in the current detection cycle; Amplifying and analog-to-digital converting the filtered analog signal in a predetermined order of a plurality of predetermined gain values ​​to obtain a digital signal, and verifying the validity of the digital signal, and terminating after obtaining the first predetermined gain value verified to be valid; Based on the first preset gain value verified to be valid, a noise intensity value of the power system dual-mode communication network is calculated.

2. The noise detection method for a power system dual-mode communication network according to claim 1, characterized in that: The performing validity verification on the digital signal includes: Acquire a valid data interval in the digital signal, where the valid data interval is a continuous sampling point of a preset length; Determining whether a first preset number of consecutive sampling points in the valid data interval reaches a saturation threshold of the digital-to-analog converter; If yes, determining that the preset gain value corresponding to the digital signal is an invalid value; If not, it is determined that the preset gain value corresponding to the digital signal is a valid value, and the noise intensity value is calculated based on the valid preset gain value.

3. The noise detection method for a power system dual-mode communication network according to claim 2, characterized in that: The determining whether there are a first preset number of consecutive sampling points in the valid data interval that reach a saturation threshold of the digital-to-analog converter includes: Scanning each of the sampling points in the valid data interval in sequence, and when the sampling point values ​​are continuously equal to the maximum positive value of the digital-to-analog converter or continuously equal to the maximum negative value of the digital-to-analog converter, adding 1 to the counting variable; otherwise, clearing the counting variable; When the counting variable reaches a first preset number, determining that there are a first preset number of consecutive sampling points that reach a saturation threshold of the digital-to-analog converter; When the count variable does not reach the first preset number at the end of the scan, it is determined that there are no consecutive first preset number of sampling points that reach the saturation threshold of the digital-to-analog converter.

4. The noise detection method for a power system dual-mode communication network according to claim 3, characterized in that: When the sampling point value is continuously equal to the maximum positive value of the digital-to-analog converter or continuously equal to the maximum negative value of the digital-to-analog converter, adding 1 to the count variable includes: If the current sampling point value is equal to the maximum positive value, and the previous point value is equal to the maximum positive value, then the counting variable is increased by 1; If the current sampling point value is equal to the maximum negative value, and the previous point value is equal to the maximum negative value, the counting variable is increased by 1.

5. The noise detection method for a power system dual-mode communication network according to claim 1, characterized in that: After verifying the validity of the digital signal, the method further includes: When no valid preset gain value is obtained after traversing all preset gain values, the preset gain value with the smallest value among the plurality of preset gain values ​​is taken as the valid preset gain value; The noise intensity value is calculated based on the preset gain value effective value.

6. The noise detection method for a power system dual-mode communication network according to claim 1, characterized in that: The step of amplifying and converting the filtered analog signal into digital signals in a predetermined order of a plurality of predetermined gain values ​​comprises: The filtered analog signal is amplified and analog-to-digital converted in descending order according to a number of preset gain values ​​to obtain a digital signal.

7. The noise detection method for a power system dual-mode communication network according to claim 1, characterized in that: Before obtaining and filtering the analog signal transmitted by the power system dual-mode communication network in the current detection cycle, the method includes: When the dual-mode communication network master node fails to receive a response message from the dual-mode communication network slave node within a preset response timeout threshold after sending a communication message, the analog signal is acquired during the communication interruption period and a noise detection process is started.

8. The noise detection method for a power system dual-mode communication network according to any one of claims 1 to 7, characterized in that: After calculating the noise intensity value of the power system dual-mode communication network, the method further includes: The noise intensity values ​​calculated by multiple detection nodes in the same detection cycle are analyzed in time and space to generate a comprehensive noise intensity value.

9. The noise detection method for a power system dual-mode communication network according to claim 8, characterized in that: The performing spatiotemporal correlation analysis on the noise intensity values ​​calculated by the multiple detection nodes within the same detection cycle includes: Aligning the noise intensity values ​​of the multiple detection nodes according to the acquisition time; Calculating a spatial correlation coefficient between the aligned plurality of noise intensity values; If the spatial correlation coefficient is greater than a preset spatial correlation threshold, the weighted average of the noise intensity values ​​of the at least two detection nodes is used as the comprehensive noise intensity value; otherwise, the noise intensity value of the current detection cycle is discarded.

10. A noise detection device for a dual-mode communication network of an electric power system, characterized in that: The noise detection method for a power system dual-mode communication network according to any one of claims 1 to 9 is used to perform noise detection on the power system dual-mode communication network, comprising: A signal acquisition module is used to acquire and filter the analog signal transmitted by the power system dual-mode communication network in the current detection cycle; a signal processing module configured to amplify and convert the filtered analog signal into digital signals in a predetermined order of a plurality of predetermined gain values, and to verify the validity of the digital signal, terminating after obtaining the first predetermined gain value verified to be valid; The noise calculation module is used to calculate the noise intensity value of the power system dual-mode communication network based on the first preset gain value verified to be valid.

Images