Smart grid communication noise analysis and ground noise visualization method

Abstract

The invention discloses a smart grid communication noise analysis and ground noise visualization method, and belongs to the technical field of smart grid communication. The method comprises the following steps: collecting original ground noise data of smart grid communication; performing first preprocessing on the original bottom noise data to obtain first preprocessed data; performing noise analysis including impulse noise, periodic noise and white noise on the first preprocessed data, and outputting noise features; performing second preprocessing on the original bottom noise data to obtain second preprocessed data; performing FFT calculation on the second pre-processed data, and performing post-processing on the frequency domain data after the FFT calculation to obtain post-processed data; and performing calibration compensation on the post-processing data to obtain data for background noise visualization, and performing background noise graph drawing. The method can fill up double gaps of laboratory groping and on-site operation and maintenance, reduce equipment and analysis cost, and improve analysis accuracy.

Description

Technical Field

[0001] This invention belongs to the field of smart grid communication technology, and in particular relates to a method for smart grid communication noise analysis and background noise visualization. Background Technology

[0002] The smart grid communication architecture, with smart meters, concentrators, and master stations as its core layers, forms a complete data interaction system. In this architecture, noise floor is a key factor affecting communication quality, and its collection and feature analysis need to cover two major scenarios: "initial laboratory assessment" and "later on-site operation and maintenance." Current work on this topic has significant shortcomings.

[0003] 1. The need for rapid assessment of laboratory performance in the early stages of the project was not met.

[0004] In the laboratory verification phase before the selection and implementation of smart grid communication solutions (such as PLCs and wireless communication modules), it is necessary to quickly evaluate the performance of communication equipment to mitigate potential risks in the field later. Noise floor is a key factor affecting communication quality, requiring extensive data collection and analysis. However, existing laboratory tests largely rely on specialized instruments and equipment, which are expensive. The specialized knowledge required for instrument operation and result analysis makes this test difficult to conduct frequently.

[0005] 2. Gaps in the main station's diagnosis of background noise-related anomalies during on-site operation and maintenance.

[0006] During the on-site operation phase after project implementation, while the master station can locate some explicit faults (such as slave nodes not being on the network or master nodes being busy) through alarm pushes and data monitoring, its remote diagnostic capabilities are significantly insufficient for special anomaly scenarios related to background noise: First, for random interference in local communication links (such as pulse noise generated by motor start-stop or narrowband noise generated by frequency converters), the master station can only monitor the increase in data error rate, but cannot capture the time-domain characteristics and frequency-domain distribution of noise, making it difficult to locate the interference source; Second, in scenarios where regional signal attenuation is superimposed with noise, the master station can only determine that the data reception quality is poor and the communication success rate is low, but cannot distinguish the dominant role of signal attenuation and noise interference, nor can it determine the distribution differences of noise at different line nodes; Third, noise in the actual environment is often composed of multiple noises superimposed, such as pulse noise, periodic noise, and white noise. The spectrum of the superimposed noise is difficult to distinguish by the naked eye, and it is difficult to calculate the key information. Traditional manual on-site inspection relies on instruments and equipment and personal experience, which is inefficient, inaccurate, costly, and poses safety risks of live operation. Summary of the Invention

[0007] Purpose of the invention: The technical problem to be solved by the present invention is to provide a method for analyzing communication noise and visualizing background noise in smart grids, addressing the shortcomings of existing technologies.

[0008] To address the aforementioned technical problems, this invention discloses a method for analyzing communication noise and visualizing background noise in smart grids, comprising the following steps:

[0009] Collect raw noise floor data of smart grid communication;

[0010] The original noise floor data is subjected to a first preprocessing to obtain first preprocessed data;

[0011] The first preprocessed data is subjected to noise analysis, including impulse noise, periodic noise, and white noise, and the noise characteristics are output.

[0012] The original noise floor data is subjected to a second preprocessing to obtain second preprocessed data;

[0013] Perform FFT calculation on the second preprocessed data, and then perform post-processing on the frequency domain data after FFT calculation to obtain post-processed data;

[0014] The post-processed data is calibrated and compensated to obtain data for noise floor visualization, and a noise floor map is drawn.

[0015] Further, the raw noise data is subjected to a first preprocessing to obtain first preprocessed data, including: splitting the collected hexadecimal smart grid communication raw noise data into high 4 bits and low 4 bits, converting them into integers respectively, and adding them to the first preprocessed data.

[0016] Further, impulse noise analysis is performed on the first preprocessed data, including:

[0017] The pulse position is identified by signal differential, the first-order difference of the first preprocessed data is calculated, the peak-to-average power ratio (PAPR) of the difference is calculated, and the region that is greater than the preset threshold is marked as the pulse interval.

[0018] Calculate pulse characteristics, analyze peak intervals, and calculate period, width, and frequency;

[0019] Calculate the pulse amplitude;

[0020] The reconstructed impulse noise is obtained based on the frequency, width, and amplitude of the impulse noise.

[0021] When calculating the amplitude, cubic spline interpolation was used, which can more accurately restore the pulse waveform and improve the accuracy of amplitude calculation.

[0022] Furthermore, performing impulse noise analysis on the first preprocessed data also includes:

[0023] The pulse region is zeroed out, and the first preprocessed data is reconstructed to obtain the first signal.

[0024] Furthermore, periodic noise analysis includes:

[0025] Spectral analysis is performed on the first signal to obtain the power spectral density and signal spectral distribution characteristics;

[0026] Based on power spectral density and signal spectral distribution characteristics, peak detection and fundamental frequency candidate selection are performed to obtain fundamental frequency candidates. Harmonic identification is then performed on the fundamental frequency candidates to obtain the optimal fundamental frequency and harmonic parameters.

[0027] Reconstruct the periodic noise waveform based on the optimal fundamental frequency and harmonic parameters;

[0028] The second signal is obtained by removing periodic noise waveforms from the first signal.

[0029] Conventional design methods for identifying periodic noise simply look for the fundamental frequency and ignore harmonics; alternatively, harmonics can be directly identified without looking for the fundamental frequency, such as by estimating harmonic frequencies using Welch power spectrum and then designing adaptive filters based on the harmonic bandwidth. This approach prioritizes efficiency and noise immunity. This step uses a fundamental frequency candidate followed by harmonic verification to obtain the optimal fundamental frequency and harmonic parameters. This method is more suitable for noise diagnosis and parameter source tracing, and the reconstructed waveform is closer to the real noise, resulting in more thorough filtering and more accurate subsequent white noise analysis.

[0030] Further, the white noise analysis includes: extracting white noise from the second signal, calculating white noise characteristics, analyzing the power spectral density of the white noise, and calculating the average power spectral density.

[0031] Furthermore, it also includes noise analysis verification:

[0032] Three types of signals are generated: square wave pulse, narrowband, and white noise. Noise is collected from these three combined signals. The three noisy signals are then mixed to obtain a mixed noisy signal.

[0033] The mixed noisy signal is subjected to noise analysis including impulse noise, periodic noise, and white noise, and the noise characteristics are output.

[0034] The noise features are compared with the mixed noisy signal features to obtain the verification results.

[0035] Conventional processing does not emphasize the processing order of various noises. This embodiment follows the order of impulse noise first, then periodic noise, and finally white noise. The difficulty of noise removal increases from easy to difficult, and the degree of noise interference decreases from strong to weak. This can maximize the accuracy of feature extraction and has high engineering feasibility.

[0036] Further, the raw noise data is subjected to a second preprocessing to obtain second preprocessed data, including: the raw noise data of the collected smart grid communication is stored in a text file in 4-byte hexadecimal format, read line by line and converted into a decimal integer, split into the lower sixteen bits and the higher sixteen bits, then the data is converted and adjusted for values ​​exceeding the int16 range, and the data is aggregated and added to a list;

[0037] The list is converted into a NumPy array. After removing unstable data from the NumPy array, N numbers are extracted as the second preprocessed data, where N is set according to the required sampling accuracy.

[0038] Furthermore, the step of performing FFT calculation on the second preprocessed data and post-processing the frequency domain data after FFT calculation to obtain post-processed data includes: performing FFT on the N truncated time domain points, taking the first N / 2 numbers (corresponding to the frequency domain data in the range of 0 to chip sampling rate / 2), performing amplitude calculation, and then converting them to dBm units.

[0039] Furthermore, the process of drawing the background noise graph includes: visualizing the background noise using Python, loading the Python library required for background noise visualization, initializing the chart layout, adding a frequency axis, time axis, title, labels, legend, and grid, adjusting the spacing between subgraphs, and saving and displaying the image.

[0040] Beneficial Effects: Existing technologies require specialized instruments and operators for noise analysis and background noise visualization, resulting in high costs in terms of time, equipment, and personnel. This invention, based on an existing chip platform, captures background noise in smart grid communication without the need for instruments, analyzes the noise, outputs characteristic parameters, and visualizes them. The operation is simple, filling the gap between laboratory assessments and on-site maintenance, reducing time, equipment, and personnel costs, and improving analytical accuracy.

[0041] Specifically, at the laboratory level, optimized measurement schemes lower the operational threshold and improve operational efficiency; at the field level, noise characteristic value analysis and background noise data visualization break through the blind spots of remote diagnosis at the main station, reduce manual power connection steps, reduce safety risks, and improve the efficiency and accuracy of anomaly handling; ultimately forming a closed loop of "preliminary verification - subsequent operation and maintenance" to provide background noise analysis assurance for the entire lifecycle of smart grid communication solutions. Attached Figure Description

[0042] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments, and the advantages of the present invention in the above and / or other aspects will become clearer.

[0043] Figure 1This is an example of raw background noise data in a smart grid communication noise analysis and background noise visualization method provided in an embodiment of this application.

[0044] Figure 2 This is a schematic diagram of the host computer interface in a smart grid communication noise analysis and background noise visualization method provided in an embodiment of this application.

[0045] Figure 3 The noise floor map is provided in a method for analyzing and visualizing communication noise floor in a smart grid according to an embodiment of this application.

[0046] Figure 4 A flowchart of a method for analyzing communication noise and visualizing background noise in smart grids, provided in an embodiment of this application. Detailed Implementation

[0047] The embodiments of the present invention will now be described with reference to the accompanying drawings.

[0048] In the communication architecture of smart grids, background noise is a key factor affecting communication quality.

[0049] Initial laboratory scenario for the project: Laboratory noise measurement traditionally relies on professional instruments such as spectrum analyzers, which have high price and operational barriers.

[0050] In the later stage of on-site operation and maintenance: the remote diagnosis of the main station can only monitor obvious results such as increased bit error rate and weak transmission, but lacks the real-time situation of the noise floor in the frequency and time domains, as well as the characteristic data such as noise amplitude and frequency, making it difficult to locate the source of interference; traditional manual investigation relies on personal experience and lacks targeted tools, coupled with the safety hazards of live operation on site, ultimately leading to inefficient and inaccurate investigation of noise floor-related anomalies.

[0051] This application discloses a method for analyzing communication noise and visualizing background noise in smart grids, such as... Figure 4 As shown, it includes the following steps:

[0052] Step 1: Collect raw noise floor data for smart grid communication. Specifically, this raw noise floor data is stored by the chip's ADC on the Vango chip platform. All modules equipped with the Vango chip platform can obtain the raw noise floor data by issuing commands. The raw data is as follows: Figure 1 As shown.

[0053] Step 2: Perform a first preprocessing on the original background noise data to obtain first preprocessed data;

[0054] The collected hexadecimal raw noise data of smart grid communication is split into high 4 bits and low 4 bits, converted into integers respectively, and added to the first preprocessed data.

[0055] Step 3: Perform noise analysis on the first preprocessed data, including impulse noise, periodic noise, and white noise, and output noise characteristics.

[0056] Step 3.1, Impulse Noise Analysis

[0057] Pulse location is identified through signal differential analysis. First, the first-order differential (diff) of the preprocessed data is calculated. Then, the peak-to-average power ratio (PAPR) of the differential is calculated. Regions exceeding a threshold of 20 dB are marked as pulse intervals. Typically, the PAPR of OFDM signals is around 5.6 dB, with the highest peak value found to be around 12-15 dB after testing; therefore, a threshold of 20 dB is set.

[0058] Average power ratio:

[0059] Calculate pulse characteristics, pulse frequency, and pulse width, specifically:

[0060] ① Find the index that satisfies the PAPR condition in the first-order difference array, and denote it as the pulse interval;

[0061] ② Based on empirical values, select indices in the first-order difference array that are greater than 80% of the peak value, and denote them as peak indices;

[0062] ③ Calculate the peak index interval within the pulse interval;

[0063] ④ Calculate the index width of a single pulse interval within the pulse interval;

[0064] ⑤ Calculate the pulse frequency using the peak index interval, and estimate the pulse width using the index width of a single pulse interval.

[0065] Pulse frequency = ,

[0066] Pulse width estimation = ,

[0067] Where fs represents the chip sampling rate for acquiring raw noise data, period_gap represents the peak index interval, and width_gap represents the index width of a single pulse interval.

[0068] Calculate the pulse amplitude:

[0069] ① Perform sliding window analysis on the first-order difference array of the signal, calculate PAPR, and filter out the pulse intervals that meet the PAPR conditions. Within the pulse interval, locate the starting index of the rising edge and the ending index of the falling edge of the pulse based on the positive and negative values ​​of the first-order difference values.

[0070] ② Iterate through the loop, extracting a single pulse interval from the signal for analysis each time, based on the starting index of the rising edge and the ending index of the falling edge.

[0071] ③ Create a mask, filter individual pulse intervals using the mask, and perform extreme value filtering to obtain the effective signal of the pulse interval.

[0072] ④ Use the "cubic spline interpolation method" to reconstruct the effective signal in the pulse interval to obtain the reconstructed signal.

[0073] ⑤ Extract the maximum value of the reconstructed signal as the estimated signal amplitude, Signal.

[0074] ⑥ Adjust the amplitude according to the chip conditions

[0075] Amplitude = Signal / 2048 * 0.7 (related to the threshold range of the chip used)

[0076] The reconstructed impulse noise is obtained based on the frequency, width, and amplitude of the impulse noise;

[0077] The pulse region is zeroed out, the first preprocessed data is reconstructed, the pulse noise is removed, and the first signal is obtained, which facilitates the analysis of other types of noise in the future.

[0078] Step 3.2, Periodic noise analysis (including narrowband noise)

[0079] The first signal is subjected to spectral analysis (FFT), and harmonic identification is performed by combining spectral peak detection and fundamental frequency candidate screening. Waveform reconstruction is then performed.

[0080] Step 3.2.1, Spectrum Analysis:

[0081] ① Input the first signal and calculate the FFT.

[0082] ② Calculate the power based on the FFT results and convert it to dBm.

[0083] Power calculation formula

[0084] Where V represents the voltage amplitude and R represents the resistance value, and according to general practice, R = 50Ω is taken;

[0085] Here, the positive frequency component of the FFT result is extracted first, and the complete signal energy is recovered to obtain the processed voltage amplitude V. Power is calculated using the formula P. dbm =10log 10 P converted to dBm

[0086]

[0087] ③ Filter positive frequency axes using a mask

[0088] ④ Output power P dBm and the filtered positive frequency axis

[0089] Step 3.2.2, Spectrum Peak Detection (Dynamic Parameters):

[0090] ① Calculate the spectral resolution (frequency interval corresponding to the number of FFT points, Hz / sample) based on the FFT results.

[0091] ② Determine the minimum frequency interval, taking at least the larger of twice the frequency resolution or 1 Hz. To ensure accuracy, twice the resolution is necessary. However, if the original resolution is low, the frequency interval must be at least 1 Hz. Then, round up to convert the interval to the number of samples.

[0092] ③ Through power P dBm And sample interval detection spectral peak (threshold relaxed to P) dBm The maximum value is 15dB lower to adapt to weak fundamental frequencies and avoid missing weak fundamental frequencies. Find the spectral peak and peak attributes that meet the power and sample interval requirements.

[0093] Step 3.2.3, fundamental frequency candidate screening, output periodic noise parameters:

[0094] ① Extract the peak frequency array from the peak values ​​of the spectrum on the positive frequency axis, and extract the peak amplitude array from the peak attributes.

[0095] ② Sort the spectrum peak amplitude array in descending order, extract the 5 largest values ​​as candidates, remove the 0 frequency, and generate a candidate list of fundamental frequencies.

[0096] ③ Iterate through all candidate fundamental frequencies, calculate the integer multiples of harmonics for each candidate fundamental frequency, and calculate the tolerance between the peak frequency and the harmonics. If the tolerance is within ±10%, it is determined to be a valid harmonic, and the candidate fundamental frequency with the most harmonics is selected as the optimal fundamental frequency.

[0097] ④ Find the frequency closest to the optimal fundamental frequency in the peak frequency array and denote it as the optimal position. Substitute the optimal position into the peak frequency array and the peak amplitude array, output the frequency of the optimal position as the narrowband frequency, and output the amplitude of the optimal position as the narrowband amplitude.

[0098] Step 3.2.4: Remove the periodic noise waveform from the first signal to obtain the second signal.

[0099] ① If an effective periodic noise fundamental frequency is detected, the Nyquist frequency is calculated and normalized for use in filter design. Then, a fourth-order band-stop filter is designed using the butter function to filter out the periodic component through bidirectional filtering.

[0100] Step 3.3, White Noise Analysis

[0101] Extract white noise from the second signal, calculate the white noise characteristics, analyze the power spectral density of the white noise, and calculate the average power spectral density.

[0102] Step 3.3.1, analyze the power spectral density of white noise.

[0103] Input a second signal, use the welch function to calculate the power spectral density array PSD[i] of the white noise data, convert the power spectral density array to dB representation, and output PSD. dB [i]

[0104] Step 3.3.2, calculate the average PSD

[0105] Select the commonly used Band 0 effective subcarrier frequency band for HPLC analysis (index 320 to 1960, corresponding to 2~12MHz of Band 0), calculate the average PSD value and convert it to dB.

[0106] Average PSD: PSDMean=

[0107] Step 4, Noise Analysis and Verification:

[0108] The instrument generates three types of signals: square wave pulse, narrowband, and white noise. The communication module is used to collect noise from the three types of signals to obtain three noisy signals. Then, the three noisy signals are mixed to obtain a mixed noisy signal.

[0109] In the specific implementation process, the instrument can use a signal generator, such as KEYSIGHT 33500B; the communication module can use any high-speed carrier module, such as Vango single-phase single-mode communication module VTM6220S-V01AXH.

[0110] The mixed noisy signal is subjected to noise analysis including impulse noise, periodic noise, and white noise, and the noise characteristics are output.

[0111] The noise features are compared with the mixed noisy signal features to obtain the verification results.

[0112] The output sample is as follows:

[0113] Noise characteristic values:

[0114] Pulse amplitude: 1.25945 V; Actual amplitude: 1.00000 V

[0115] Pulse frequency: 100.00000 Hz; Actual frequency: 100.00000 Hz

[0116] Pulse width estimate: 1.12000 us; Actual pulse width: 1.00000 us

[0117] Narrowband amplitude: -23.48678 dBm; Actual amplitude: -20.00000 dBm

[0118] Narrowband frequency: 1.01318 MHz; Actual frequency: 1.00000 MHz

[0119] White noise average PSD: -133.81151 dBm; Actual average PSD: -133.00000 dBm

[0120] As can be seen from the above output examples, the deviation between the noise characteristic value and the actual value is within an acceptable range, proving that the noise analysis process of the present invention is effective and reliable, and the noise analysis results obtained based on this process can be adopted.

[0121] Step 5: Perform a second preprocessing on the original background noise data to obtain second preprocessed data;

[0122] The noise floor data collected by the chip platform is stored in a text file in 4-byte hexadecimal format. After reading it line by line, it is converted into a decimal integer, split into the lower 16 bits and the higher 16 bits, and then the data is converted and adjusted for values ​​that exceed the int16 range. Finally, the data is aggregated and added to a list.

[0123] Because the original data contains redundant parts, preprocessing is required to extract the valid data. The process involves first converting the list into a NumPy array, then removing unstable data from the NumPy array, and finally extracting N sampling points, which constitutes the second preprocessed data. The value of N is set according to the required sampling precision. For example, assuming there are initially 2017 unstable numbers, then starting from the 2017th point in the NumPy array, 8192 points are extracted, resulting in N = 8192.

[0124] Step 6: Perform FFT calculation on the second preprocessed data, and post-process the frequency domain data after FFT calculation to obtain post-processed data. For example, if 8192 sampling points were obtained previously and the chip sampling rate is 50MHz, then these 8192 sampling points correspond to -25MHz to +25MHz.

[0125] Perform an FFT on the 8192 time-domain points, take the first 4096 numbers (corresponding to the 0~25MHz frequency range), calculate the amplitude, and convert them to dBm units, which is more in line with human visual perception habits.

[0126] Step 7: Perform calibration compensation on the post-processed data to obtain data for noise floor visualization, and draw the noise floor map;

[0127] The calibration compensation process involves subtracting the pre-calibrated calibration value of -156 dBm from the post-processed data to obtain data for noise floor visualization. The pre-calibration data was obtained through extensive experiments, ensuring that the final image is essentially consistent with the values ​​obtained by the instrument.

[0128] This section describes how to visualize background noise using Python. It involves loading the necessary Python libraries, initializing the chart layout, adding a frequency axis, time axis, title, labels, legend, and grid, adjusting subplot spacing to improve readability, and finally saving and displaying the image.

[0129] By using the Python matplotlib library and integrating the above steps, the host computer interface is completed, further enhancing product usability. The host computer interface is shown below. Figure 2 As shown.

[0130] The completed noise floor map is as follows: Figure 3 As shown.

[0131] In the specific implementation process, in addition to using Python to complete the visualization steps, other programming languages ​​can also be used to complete the visualization steps, and this embodiment of the invention does not limit the scope of the visualization steps.

[0132] Addressing the contradiction between the high barrier to entry for specialized instruments and the insufficient accuracy of simple methods in laboratory noise floor measurement, this invention optimizes the measurement scheme, avoiding the price and operational barriers of spectrum analyzers. Noise floor acquisition no longer requires instruments; instead, it utilizes communication modules essential for smart grid projects, incurring no additional costs and making operation more convenient and efficient. This reduces initial verification costs while ensuring data reliability. Furthermore, manual analysis suffers from low data accuracy due to visual observation and is prone to master station bias. This invention automatically performs noise analysis, providing noise characteristic values ​​to aid in noise result inference.

[0133] Addressing the pain points of on-site operation and maintenance, it directly fills the feature data gap in remote diagnosis of the main station by processing background noise data and presenting it in the time domain and frequency domain. This eliminates the need for manual experience in locating interference sources and distinguishing fault types. Furthermore, by replacing manual inspections while the circuit is powered on, it not only eliminates safety hazards but also transforms "inefficient blind inspections" into "precise diagnosis," ultimately achieving a triple improvement in operation and maintenance efficiency, accuracy, and safety. The closed-loop design of "preliminary laboratory verification - subsequent on-site operation and maintenance" allows background noise analysis to be integrated throughout the entire lifecycle of the communication solution, further amplifying the synergistic effect of the advantages of each link.

[0134] In its specific implementation, this application provides a computer storage medium and a corresponding data processing unit. The computer storage medium is capable of storing a computer program, which, when executed by the data processing unit, can run the invention's content regarding a method for analyzing and visualizing noise floor in smart grid communication, as well as some or all of the steps in various embodiments. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0135] Those skilled in the art will clearly understand that the technical solutions in the embodiments of the present invention can be implemented using computer programs and their corresponding general-purpose hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, or the parts that contribute to the prior art, can be embodied in the form of computer programs, i.e., software products. These computer program software products can be stored in a storage medium and include several instructions to cause a device containing a data processing unit (which may be a personal computer, server, microcontroller, MUU, or network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments of the present invention.

[0136] This invention provides a method for analyzing communication noise and visualizing background noise in smart grids. Many methods and approaches exist for implementing this technical solution; the above description is merely a preferred embodiment. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this invention, and these improvements and modifications should also be considered within the scope of protection of this invention. All components not explicitly stated in this embodiment can be implemented using existing technologies.

Claims

1. A method for analyzing communication noise and visualizing background noise in smart grids, characterized in that, Includes the following steps: Collect raw noise floor data of smart grid communication; The original noise floor data is subjected to a first preprocessing to obtain first preprocessed data; The first preprocessed data is subjected to noise analysis, including impulse noise, periodic noise, and white noise, and the noise characteristics are output. The original noise floor data is subjected to a second preprocessing to obtain second preprocessed data; Perform FFT calculation on the second preprocessed data, and then perform post-processing on the frequency domain data after FFT calculation to obtain post-processed data; The post-processed data is calibrated and compensated to obtain data for noise floor visualization, and a noise floor map is drawn.

2. The method for analyzing communication noise and visualizing background noise in smart grids according to claim 1, characterized in that, The raw noise data is subjected to a first preprocessing step to obtain first preprocessed data, which includes: splitting the collected hexadecimal smart grid communication raw noise data into high 4 bits and low 4 bits, converting them into integers and adding them to the first preprocessed data.

3. The method for analyzing communication noise and visualizing background noise in smart grids according to claim 2, characterized in that, The first preprocessed data is subjected to impulse noise analysis, including: The pulse position is identified by signal differential, the first-order difference of the first preprocessed data is calculated, the peak-to-average power ratio (PAPR) of the difference is calculated, and the region that is greater than the preset threshold is marked as the pulse interval. Calculate pulse characteristics, analyze peak intervals, and calculate period, width, and frequency; Calculate the pulse amplitude; The reconstructed impulse noise is obtained based on the frequency, width, and amplitude of the impulse noise.

4. The method for analyzing communication noise and visualizing background noise in smart grids according to claim 3, characterized in that, The method of performing impulse noise analysis on the first preprocessed data further includes: The pulse region is zeroed out, and the first preprocessed data is reconstructed to obtain the first signal.

5. The method for analyzing communication noise and visualizing background noise in smart grids according to claim 4, characterized in that, Periodic noise analysis includes: Spectral analysis is performed on the first signal to obtain the power spectral density and signal spectral distribution characteristics; Based on power spectral density and signal spectral distribution characteristics, spectral peak detection and fundamental frequency candidate screening are performed to obtain fundamental frequency candidates; harmonic identification is then performed on the fundamental frequency candidates to obtain the optimal fundamental frequency and harmonic parameters. Reconstruct the periodic noise waveform based on the optimal fundamental frequency and harmonic parameters; The second signal is obtained by removing periodic noise waveforms from the first signal.

6. The method for analyzing communication noise and visualizing background noise in smart grids according to claim 5, characterized in that, White noise analysis includes: extracting white noise from the second signal, calculating white noise characteristics, analyzing the power spectral density of white noise, and calculating the average power spectral density.

7. The method for analyzing communication noise and visualizing background noise in smart grids according to claim 6, characterized in that, It also includes noise analysis verification: Square wave pulse, narrowband, and white noise signals are generated respectively. Noise is collected from the three types of signals to obtain three noisy signals. The three noisy signals are then mixed to obtain a mixed noisy signal. The mixed noisy signal is subjected to noise analysis including impulse noise, periodic noise, and white noise, and the noise characteristics are output. The noise features are compared with the mixed noisy signal features to obtain the verification results.

8. The method for analyzing communication noise and visualizing background noise in smart grids according to claim 7, characterized in that, The raw noise data is subjected to a second preprocessing to obtain second preprocessed data, including: the raw noise data of the smart grid communication collected is stored in a text file in 4-byte hexadecimal format, read line by line and converted into decimal integers, split into low sixteen bits and high sixteen bits, then the data is converted and adjusted for values ​​exceeding the int16 range, and the data is aggregated and added to a list; The list is converted into a NumPy array. After removing unstable data from the NumPy array, N numbers are extracted as the second preprocessed data, where N is set according to the required sampling accuracy.

9. The method for analyzing communication noise and visualizing background noise in smart grids according to claim 8, characterized in that, The step of performing FFT calculation on the second preprocessed data and post-processing the frequency domain data after FFT calculation to obtain post-processed data includes: performing FFT on the N truncated time domain points, taking the first N / 2 numbers, performing amplitude calculation, and converting them to dBm units.

10. The method for analyzing communication noise and visualizing background noise in smart grids according to claim 9, characterized in that, The process of creating a background noise graph includes: using Python to visualize the background noise, loading the Python libraries required for background noise visualization, initializing the chart layout, adding a frequency axis, time axis, title, labels, legend, and grid, adjusting the spacing between subgraphs, and saving and displaying the image.

Images