A lightning early warning method, system and terminal based on wavelet decomposition

Abstract

The application provides a lightning early warning method and system based on wavelet decomposition and an early warning terminal, and relates to the technical field of lightning early warning. The method comprises the following steps: obtaining original atmospheric data in a preset time period; selecting a wavelet information to perform wavelet decomposition on a noisy signal; configuring a preset threshold value for each layer coefficient obtained through decomposition; performing inverse wavelet transformation on the coefficient after noise reduction processing to obtain reconstructed atmospheric data after noise reduction; obtaining a low-frequency component representing an approximate trend of atmospheric electric field change and a high-frequency component representing a detailed part of the atmospheric electric field through 5-layer decomposition of a sym4 wavelet function; zero-filling the d5 coefficient to reconstruct the d5 coefficient component under the condition that the sampled atmospheric data meets the Nyquist criterion; and using the d5 coefficient component to detect the sudden change of the atmospheric electric field and to early warn the lightning, so that the lightning occurring in the thunderstorm passage stage can be early warned, the time point of the thunderstorm early warning can be determined, the lightning early warning requirement can be met, and the early warning precision is improved.

Description

Technical Field

[0001] This invention relates to the field of lightning warning technology, and in particular to a lightning warning method, system and warning terminal based on wavelet decomposition. Background Technology

[0002] The atmospheric electric field is a fundamental phenomenon that always exists in the atmosphere, exhibiting different characteristics under different weather conditions. It is a basic element of atmospheric physics. Besides being influenced by factors such as cosmic radiation, solar ultraviolet radiation, and global thunderstorm activity, atmospheric electric parameters are also closely related to local climate and environment. Its intensity is controlled by multiple factors, including significant influences from surface meteorological conditions such as temperature, humidity, pressure, wind, and weather patterns. It exhibits a strong response to weather events such as thunderstorms, precipitation processes, and severe weather events like sandstorms. Under clear weather conditions, the potential difference between the ionosphere and the Earth's surface is approximately 300 kV, with the electric field pointing vertically from the ionosphere towards the ground. The electric field strength increases from high to low altitudes, reaching approximately 200 V / m at the ground. Under the influence of the atmospheric electric field, the directional movement of charged atmospheric particles continuously consumes electrical energy. The electrical energy accumulated during global thunderstorm activity flows into the ionosphere in the form of discharges, maintaining a stable atmospheric ionospheric potential and the intensity of the atmospheric electric field under clear weather conditions. Lü Daren et al. used three years of electric field records from 1990 to 1992 to analyze the electric field characteristics of various time scales, seasonal variations, typical daily variations, and typical weather conditions. They pointed out that the electric field spectrum characteristics on a one-year time scale have obvious universality, and its spectrum can be expressed as a two-segment linear decreasing law on a logarithmic scale.

[0003] The periodic variation of the atmospheric electric field under clear-sky conditions was previously presented by Xu Bin et al. based on the ARGO-YBJ experiment in Tibet. They analyzed the periodic variation of the near-surface atmospheric electric field intensity in the Yangbajing area of ​​Tibet during clear-sky conditions from December 2005 to September 2007, focusing on the phase variation of its solar diurnal cycle. The results showed that the atmospheric electric field intensity exhibited significant semi-diurnal and solar diurnal variations under clear-sky conditions. The phase variation of the solar diurnal cycle exhibited a complex continental pattern (double peaks, double troughs), and the diurnal variation curve varied seasonally. In winter, the electric field intensity was higher, with a maximum daily variation of approximately 30%; in spring and summer, the electric field intensity was lower, with a maximum daily variation of approximately 15%. Regarding the characteristics of the surface atmospheric electric field during thunderstorm activity, Yang Bo et al. used atmospheric electric field data accumulated in Taishan and Xishuangbanna to discuss the reasons for the positive and negative jumps in the atmospheric electric field during thunderstorms. Through statistical analysis of the atmospheric electric field data, they presented the mean distribution and fluctuation range of the atmospheric electric field during thunderstorm activity. The study compared and analyzed the similarities and differences in atmospheric electric field changes during thunderstorm activities in Mount Tai and Xishuangbanna.

[0004] Therefore, as one of the most commonly used atmospheric electric field detection devices, the atmospheric electric field meter can continuously measure the intensity and polarity of the ground atmospheric electric field for a long time. It can record the atmospheric electric state on clear days, as well as the electrical activity before thunderstorms and the changes in the electric field during thunderstorms. It has important application value in many aspects such as lightning monitoring and early warning.

[0005] In addition, to study the atmospheric electric field, various electric field sensors have been developed both domestically and internationally to monitor electric field strength. These sensors can directly measure the electric field strength on the ground or in the air, enabling effective monitoring and analysis. Stolzenburg et al. used a balloon-borne dual-sphere electric field meter and a rocket-borne field mill electric field meter to obtain electric field sounding curves within the convection regions of mesoscale convective systems, isolated thunderstorms, and neo-Mexican mountain thunderstorms. They discovered that these three types of thunderstorms share a common fundamental charge structure: four charge regions with sequentially changing polarity exist within the updraft region, with the lowest charge region being positive; outside the updraft region, six charge regions with sequentially changing polarity exist, with the lowest charge region being positive. However, domestic research in this area is relatively rare. As early as the 1960s, Yuan Zhen et al. obtained five electric field detection results within thunderstorm clouds in the Nanjing area of ​​my country and briefly analyzed the possible electrification mechanism. This research was interrupted for approximately 30 years until the 1990s, when Zhuang Hongchun et al. successfully developed a dual-sphere atmospheric electric field meter, but no results were found for thunderstorm cloud electric field detection.

[0006] Many scholars have conducted research on using changes in atmospheric electric fields for lightning warnings. Ding Deping statistically analyzed 32 thunderstorm events and obtained the three-level lightning warning thresholds for the Beijing area as 2.84, 5.58, and 8.29 kV / m, respectively. Wang Zhenhui proposed a near-term forecast method for the first ground flash within a certain range of a monitoring station. This method first determines whether the electric field data reflecting the charge of the thunderstorm cloud has reached a set electric field amplitude threshold and an electric field difference threshold. If the electric field meets either of these thresholds, it then determines whether the radar echo within the warning range before and after that moment reaches a set threshold to issue a warning. Yang Bo proposed setting an electric field warning value by detecting jumps in the measured values ​​of an electric field meter. The terminal software of the electric field meter compares two adjacent electric field samples; when the difference between the two samples exceeds a certain value, it is considered that an electric field jump has occurred, indicating that lightning is occurring nearby, and the electric field meter begins to alarm. Chai Rui studied the rapid fluctuation phase of the atmospheric electric field, proposing that when the absolute value of the electric field exceeds 3.5kV for 6.5 minutes, lightning is very likely to occur in the protected area after about 15.9 minutes. In summary, current methods for using atmospheric electric fields for lightning warning typically involve setting a warning value based on statistical analysis of the electric field strength. However, this method has two drawbacks: First, the changes in the atmospheric electric field during thunderstorms are very complex, especially as thunderstorms approach, where the change curve exhibits strong randomness and lacks a unified classification model. Therefore, simply setting multiple threshold levels is insufficient to achieve the desired warning effect. Second, due to the influence of factors such as terrain on the atmospheric electric field, atmospheric electric field meters will exhibit certain differences under different installation environments. These differences will lead to variations in measurement data from different electric field meters under the same weather conditions, resulting in a lack of universality in the set warning thresholds. Summary of the Invention

[0007] This invention provides a lightning warning method based on wavelet decomposition. The warning method can determine the time point of thunderstorm warning by setting a threshold using a hard threshold method, thus meeting the requirements of lightning warning and improving the accuracy of the warning.

[0008] The methods include:

[0009] S1. Obtain raw atmospheric data within a preset time period;

[0010] S2. Select wavelet information and perform N-level wavelet decomposition on the noisy signal;

[0011] S3. Configure preset thresholds for the coefficients of each layer obtained from the decomposition;

[0012] S4. Perform inverse wavelet transform on the denoised coefficients to obtain the reconstructed atmospheric data after denoising.

[0013] S5. Using the 5-level decomposition of the sym4 wavelet function, the low-frequency component 'a' representing the approximate trend of atmospheric electric field variation and the high-frequency component 'd' representing the detailed part of atmospheric electric field variation were obtained. The atmospheric data of atmospheric electric field satisfy the following expression:

[0014] S = a5 + d1 + d2 + d3 + d4 + d5;

[0015] Where S is the raw data of the atmospheric electric field, a5 is the low-frequency component, and d1 to d5 are the high-frequency components.

[0016] S6. The d5 coefficients were zero-padding to reconstruct the d5 coefficient components under the condition that the sampled atmospheric data satisfies the Nyquist law. The d5 coefficient components are used to detect abrupt changes in the atmospheric electric field and provide early warning of lightning.

[0017] It should be further noted that the steps for obtaining raw atmospheric data within the preset time period also include:

[0018] The atmospheric data of a preset area is selected as the raw atmospheric data based on a preset date.

[0019] It acquires a preset amount of raw atmospheric data every minute and stores it as a file on a daily basis.

[0020] It should be further noted that the file is stored in a format that stores one raw atmospheric data point per second, with 60 raw atmospheric data points stored per minute on a single line.

[0021] It should be further noted that the method also includes data completion rules;

[0022] The data completion rules are as follows: when the number of missing data lines per day is less than 60, the missing atmospheric data values ​​will be replaced by the average of the 10 data lines preceding that time; when the number of missing atmospheric data lines per day is more than 60, the entire file will be removed.

[0023] It should be further explained that the electric field changes during a thunderstorm are divided into three stages: the approaching stage, the passing stage, and the post-passing stage.

[0024] It should be further explained that the method's characteristic in detecting abrupt change signals is that it analyzes the changes in the d5 coefficient and d5 reconstruction sequence of the atmospheric electric field during the approach and transit phases of thunderstorms using waveform images, thereby resolving the threshold of abrupt change points during the approach phase of thunderstorms and providing early warning of lightning occurring during the transit phase of thunderstorms.

[0025] The present invention also provides a lightning early warning system based on wavelet decomposition, the system comprising: a data acquisition module, a wavelet decomposition and reconstruction module, a wavelet decomposition module, and a lightning early warning module;

[0026] The data acquisition module is used to acquire raw atmospheric data within a preset time period;

[0027] The wavelet decomposition and reconstruction module is used to select wavelet information, perform N-level wavelet decomposition on the noisy signal, configure preset thresholds for the coefficients of each level obtained by decomposition, and perform inverse wavelet transform on the denoised coefficients to obtain the reconstructed atmospheric data after denoising.

[0028] The wavelet decomposition module utilizes the sym4 wavelet function with 5 levels of decomposition to obtain the low-frequency component 'a' representing the approximate trend of atmospheric electric field variation and the high-frequency component 'd' representing the detailed part of atmospheric electric field. The atmospheric electric field data satisfies the following expression:

[0029] S = a5 + d1 + d2 + d3 + d4 + d5

[0030] Where S is the raw data of the atmospheric electric field, a5 is the low-frequency component, and d1 to d5 are the high-frequency components.

[0031] The lightning warning module is used to zero-paste the d5 coefficients, reconstructing the d5 coefficient components under the condition that the sampled atmospheric data satisfies the Nyquist law. The d5 coefficient components are used to detect abrupt changes in the atmospheric electric field and provide early warning of lightning.

[0032] The present invention also provides an early warning terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of a lightning early warning method based on wavelet decomposition.

[0033] As can be seen from the above technical solutions, the present invention has the following advantages:

[0034] Based on the wavelet decomposition-based lightning warning method provided by this invention, in terms of selecting the lightning warning threshold, the high-frequency characteristics of the reconstructed sequence of the d5 component of the wavelet decomposition of atmospheric electric field intensity can represent the abrupt changes in atmospheric electric field before, during and after the passage of a thunderstorm. The threshold set by the hard threshold method can determine the time point of the thunderstorm warning.

[0035] The lightning warning system based on wavelet decomposition provided by this invention can summarize atmospheric data, efficiently collect, store, and process atmospheric data. It can monitor the three stages of the thunderstorm process, namely the approach stage, the transit stage, and the post-transit stage. Based on wavelet decomposition, the accuracy of lightning warning is improved, thereby realizing the timeliness and scientific nature of the whole process of lightning warning supervision and management. Attached Figure Description

[0036] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 The flowchart shows a lightning warning method based on wavelet decomposition.

[0038] Figure 2 A spatial distribution map of the three stations;

[0039] Figure 3 The graph shows the evolution of the atmospheric electric field and wavelet components during a thunderstorm at Suncun Town Station on July 30, 2016.

[0040] Figure 4 The graph shows the evolution of the atmospheric electric field and wavelet components during a thunderstorm at Suncun Town Station on August 3, 2016.

[0041] Figure 5 The graph shows the evolution of the atmospheric electric field and wavelet components during a thunderstorm at Suncun Town Station on August 14, 2016.

[0042] Figure 6 The graph shows the evolution of the atmospheric electric field and wavelet components during a thunderstorm at Suncun Town Station on August 19, 2016.

[0043] Figure 7 This is a graph showing the evolution of the atmospheric electric field and wavelet components during a thunderstorm at Suncun Town Station on July 26, 2017.

[0044] Figure 8 This is a graph showing the evolution of the atmospheric electric field and wavelet components during a thunderstorm at Suncun Town Station on August 18, 2017.

[0045] Figure 9 This is a graph showing the evolution of the atmospheric electric field and wavelet components during a thunderstorm at Suncun Town Station on August 23, 2017. Detailed Implementation

[0046] like Figure 1As shown, the lightning warning method based on wavelet decomposition provided by this invention utilizes digital twin (DT) wavelet decomposition technology. By collecting raw atmospheric data within a preset time period, and employing data transmission and processing technologies, it enables early warning of lightning in the weather, reflecting the weather conditions within a preset area. This effectively solves the problem that the atmospheric electric field changes during thunderstorms are complex, and the change curve of the atmospheric electric field is highly random as thunderstorms approach, making it difficult to achieve ideal early warning results simply by setting multiple threshold levels. Furthermore, it addresses the issue that the influence of factors such as terrain on the atmospheric electric field causes variations in the atmospheric electric field meter under different installation environments, resulting in a lack of universality in the set warning thresholds.

[0047] The wavelet decomposition-based lightning warning method provided by this invention takes into account that during lightning strikes, charges in clouds neutralize at the cloud-to-ground or cloud-to-ground levels. The abrupt change in cloud charges causes a change in the ground electric field. If the lightning intensity is sufficiently high, a significant jump in the ground electric field will occur. Conversely, excluding the accidental approach of electrostatically charged objects to the electric field meter sensor, a lightning strike corresponds to the detection of a jump in the ground electric field by the electric field meter. This is because the ground atmospheric electric field is the result of static charge induction in the atmosphere or thunderclouds. When no lightning occurs, the accumulation and movement of charges is a continuous process and does not cause a jump in the ground electric field. Only when lightning occurs will the electric field jump be caused. However, lightning that causes a rapid jump in the atmospheric electric field is not necessarily a ground-to-ground lightning strike. This is because the rapid jump in the atmospheric electric field is influenced by the background charge structure in the clouds, as well as cloud-to-ground lightning. Therefore, lightning strikes can be predicted by forecasting the entry of the atmospheric electric field into a rapid jump phase. Furthermore, since the atmospheric electric field meter can measure the magnitude and continuous change of polarity of the average atmospheric electric field, it is very sensitive to the atmospheric electric field when a thunderstorm passes overhead at close range. It can simultaneously and continuously monitor the electrostatic field generated by the thunderstorm on the ground as well as the occurrence of cloud flash and ground flash. Therefore, it can be used for local thunderstorm monitoring and early warning, and can also be used in places that are prone to static electricity or susceptible to static electricity hazards to monitor the static electricity intensity and avoid potential dangers.

[0048] The wavelet decomposition-based lightning early warning method of the present invention is applied to one or more early warning terminals. The early warning terminal is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions. Its hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0049] The early warning terminal can be any electronic product that can interact with the user, such as personal computers, tablets, smartphones, personal digital assistants (PDAs), interactive network television (IPTV), etc.

[0050] The early warning terminal may also include network devices and / or user devices. The network devices include, but are not limited to, a single network server, a server group consisting of multiple network servers, or a cloud based on cloud computing consisting of a large number of hosts or network servers.

[0051] The networks where the early warning terminals are located include, but are not limited to, the Internet, wide area networks, metropolitan area networks, local area networks, and virtual private networks (VPNs).

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] In the wavelet decomposition-based lightning early warning method provided by this invention, the raw atmospheric data within a preset time period can be obtained by selecting data from the Jinan Suncun Town station from January to December 2017 as the first-step filtered data based on the dates corresponding to the clear-sky data. The raw atmospheric electric field data is stored in individual data files on a daily basis, i.e., one file per day. The file storage format is one data point per second, with 60 data points stored per minute on a single line, totaling 1440 lines per day (24 hours), and 86400 data points per file.

[0054] Due to missing data in some areas, data supplementation was performed on the data after the initial screening to meet the completeness and statistical requirements of each period. The data supplementation rules were as follows: when the number of missing data rows per day was less than 60, the missing data value was replaced by the average of the 10 data rows preceding that time period; when the number of missing data rows per day was more than 60, the entire file was removed. For all data after completing the above steps, under clear-sky atmospheric electric field conditions, the average of 60 data points per row was taken as a single sample data unit.

[0055] In embodiments of the present invention, wavelet information is selected, and N-level wavelet decomposition is performed on the noisy signal;

[0056] Specifically, the ground atmospheric electric field waveform data obtained by the atmospheric electric field meter contains not only the actual ground atmospheric electric field signal to be detected, but also various interference noises, including system random noise. Since the system random noise is a broadband signal in both the time and frequency domains, it cannot be directly removed. Filtering methods can be used to suppress the noise and increase the signal-to-noise ratio.

[0057] Therefore, wavelet decomposition can be considered for processing one-dimensional signals. Unlike Fourier analysis, wavelet analysis has the ability to characterize local signal features in both the time and frequency domains. It is a time-frequency localization analysis method where the window size is fixed but its shape can be changed, and both the time and frequency windows can be modified. Wavelet transform has high frequency resolution and low time resolution in the low-frequency region, and high time resolution and low frequency resolution in the high-frequency region. The signal denoising characteristic of wavelet transform is that it distributes the noise energy across all wavelet coefficients, while the wavelet coefficients of the signal are concentrated only in a finite portion of the frequency scale space.

[0058] The specific process of wavelet analysis for denoising is as follows: First, select a wavelet and perform N-level wavelet decomposition on the noisy signal; second, select an appropriate threshold for the coefficients of each level obtained by decomposition; third, perform inverse wavelet transform on the denoised coefficients to obtain the reconstructed signal after denoising.

[0059] In this invention, a wavelet is defined as a function that localizes a given function. A wavelet can be derived from a function defined on a finite interval. To construct, This is called the mother wavelet or the basic wavelet. It consists of a set of wavelet basis functions. It can be achieved by scaling and translating the basic wavelet. To generate,

[0060]

[0061] Where a is the scaling parameter, which reflects the width (or scale) of a specific basis function; b is the translation parameter, which specifies the position to be translated along the x-axis.

[0062] When a = 2j and b = ia, the one-dimensional wavelet basis function sequence is defined as:

[0063]

[0064] Where i is the translation parameter and j is the scaling factor.

[0065] The function f(x) is expressed in wavelet... The continuous wavelet transform with basis is defined as the function f(x) and The inner product,

[0066]

[0067] The inverse transform of the wavelet is:

[0068]

[0069] in, Mother wavelet Admissible conditions.

[0070] When using wavelet transform to denoise ground-based atmospheric electric field signals, it is important to select appropriate wavelet functions, analysis functions, and decomposition levels. Wavelet transform provides a variety of wavelet functions with different time-frequency characteristics.

[0071] This invention selects a 5-level decomposition of the sym4 wavelet basis to analyze atmospheric electric field data. After wavelet decomposition, the wavelet coefficient a5, representing the changing trend, can well correspond to the changing trend of the original curve. Moreover, since the interference of noise signals is removed while the information of the original data is preserved, the curve after wavelet decomposition is used to describe the changing trend of atmospheric electric field data more clearly.

[0072] For the observation data of this invention, the atmospheric electric field data are observation data from three stations using a DF02 electric field meter. For example, these can be the Suncun Town Station in Jinan, the Wuyingshan Station, and the Agricultural Academy Station. The locations of the three stations are distributed as follows: Figure 2 As shown.

[0073] Taking seven thunderstorm events from July 2016 to August 2017 as examples, based on lightning location data observed by the ATDT lightning location system, lightning occurring within a 15km radius of the center of each of the following three stations—Suncun Town Station (117.317285E, 36.695090N), Wuyingshan Station (116.97555E, 36.69N), and Agricultural Academy Station (117.064322E, 36.695938N)—was selected as the object of atmospheric electric field analysis and lightning warning threshold research for thunderstorm events. The seven thunderstorm events and lightning location data are shown in Tables 1-4.

[0074] Table 1. Thunderstorm weather processes and lightning location data within 15km of Suncun Town Station.

[0075]

[0076] Table 1. Thunderstorm weather processes and lightning location data within 15km of Suncun Town Station (continued)

[0077]

[0078] Table 2. Thunderstorm weather processes and lightning location data within 15km of Suncun Town Station (continued)

[0079]

[0080] Table 3. Thunderstorm weather processes and lightning location data within 15km of Wuyingshan Station

[0081]

[0082] Table 3. Thunderstorm weather processes and lightning location data within 15km of Wuyingshan Station (continued)

[0083]

[0084] Table 4. Thunderstorm weather processes and lightning location data within 15km of the Agricultural Science Academy station.

[0085]

[0086] Table 4. Thunderstorm weather processes and lightning location data within 15km of the Agricultural Science Academy station (continued)

[0087]

[0088] Table 4. Thunderstorm weather processes and lightning location data within 15km of the Agricultural Science Academy station (continued)

[0089]

[0090] This invention also analyzes the atmospheric electric field characteristics of thunderstorm weather processes. Specifically, based on the distance of the thunderstorm process relative to the atmospheric electric field meter station during its movement, the electric field changes of the thunderstorm process can be divided into three stages: the approach stage, the passing stage, and the post-passing stage.

[0091] Using the sym4 wavelet function with 5-level decomposition, the low-frequency component (a) representing the approximate trend of atmospheric electric field variation and the high-frequency component (d) representing the detailed part of atmospheric electric field variation are obtained. The original signal of atmospheric electric field satisfies the following expression: S=a5+d1+d2+d3+d4+d5

[0092] Where S represents the original data of the atmospheric electric field, a5 represents the low-frequency component, and d1 to d5 represent the high-frequency components. When applying the d5 coefficient to the lightning warning threshold of this invention, in order to obtain a time series that corresponds one-to-one with the original data, the d5 coefficient is padded with zeros so that the d5 coefficient component can be reconstructed under the condition that the sampled signal satisfies the Nyquist law. The d5 coefficient component can be used to detect abrupt changes in the atmospheric electric field, and thus for lightning warning.

[0093] Figure 3 — Figure 9The figure presents the time-varying changes in the average atmospheric electric field at Suncun Town station before and after seven thunderstorm events. From top to bottom, the figures show the atmospheric electric field intensity (raw data), approximate component coefficients of the 5-level symmetric wavelet decomposition, reconstructed detail component coefficients of the 5-level symmetric wavelet decomposition, and detail component coefficients of the 5-level symmetric wavelet decomposition. The vertical lines in the figure represent ground flashes observed at the corresponding times (in 1 second).

[0094] Since the time of the lightning data is GPS time, while the time of the atmospheric electric field meter is the computer host time, although the time difference between the two is small, they cannot be completely correlated. The focus of this invention is on the changing characteristics of the atmospheric electric field during the approaching and passing stages of a thunderstorm, and on finding suitable indicators in the approaching stage of a thunderstorm to indicate the stage of drastic changes in the atmospheric electric field during the passing stage of a thunderstorm, thereby providing lightning warnings. Rather than studying a one-to-one correspondence between the occurrence of lightning and changes in the atmospheric electric field in time, the time of the lightning data is corrected to the time axis of the atmospheric electric field data and used as a graph to visually express the changes in the atmospheric electric field during a thunderstorm with lightning. The inconsistency in time between the lightning data and the atmospheric electric field data will not affect the results of this invention.

[0095] Overall, the atmospheric electric field undergoes significant changes throughout the entire process of thunderstorm activity. As a thunderstorm approaches, the atmospheric electric field begins to change noticeably from a stable state, primarily manifested as a gradual increase in amplitude or a reversal of direction. As the thunderstorm reaches the observation point, the amplitude of the atmospheric electric field intensity experiences a large and rapid jump, accompanied by a reversal of direction, resulting in very drastic fluctuations. After the thunderstorm passes, these fluctuations gradually decrease until they return to the stable state before the thunderstorm. These characteristics are closely related to changes in the charge structure within thunderstorm clouds. The amount of charge at the bottom layer of the thunderstorm cloud determines the magnitude of the atmospheric electric field, while the polarity of the charge at the bottom layer plays a decisive role in the direction of the atmospheric electric field. During the thunderstorm's passage, the large and rapid fluctuations in the intensity and polarity of the atmospheric electric field reflect the complex charge structure and charge discharge within the atmospheric electric field. This charge discharge includes not only the observed cloud-to-ground discharge but also charge discharge within and between clouds.

[0096] This invention utilizes the 5-level decomposition of the sym4 wavelet function to obtain the low-frequency component (a) representing the approximate trend of atmospheric electric field variation and the high-frequency component (d) representing the detailed part of atmospheric electric field variation.

[0097] Due to the characteristics of the d5 coefficient in detecting abrupt changes, the changes in the d5 coefficient and d5 reconstruction sequence of the atmospheric electric field during the approach and transit phases of thunderstorms can be analyzed by waveform image analysis to identify the threshold of abrupt change points during the approach phase of thunderstorms, thereby providing early warning of possible lightning during the transit phase of thunderstorms.

[0098] Taking the thunderstorm process on August 3, 2016 as an example, from... Figure 4 It can be seen that between 16:00 and 16:55, the intensity of the atmospheric electric field showed a gradual increasing trend, which was the approaching stage of the thunderstorm; between 16:55 and 17:35, the atmospheric electric field changed rapidly, which was the passing stage of the thunderstorm, during which 22 ground flashes were detected; between 17:35 and 18:00, the atmospheric electric field fluctuations returned to stability, which was the post-passage stage of the thunderstorm.

[0099] from Figure 4 The two graphs of atmospheric electric field intensity and a5 coefficient show that the upward trend of the atmospheric electric field from approximately 16:00 to 16:50 is a relatively continuous process. However, the two graphs of d5 and d5 reconstruction show a stable high-frequency signal accompanied by many numerical jumps. Significant amplitude and frequency changes occur around 16:47 during the upward phase of the atmospheric electric field. In the d5 reconstruction sequence, all corresponding values ​​within 30 seconds before and after 16:47 were identified, and the absolute value of the instantaneous maximum was determined to be 66.9, occurring at 16:47:50. During the phase of drastic changes in the atmospheric electric field beginning around 16:55, the same method was used to identify the absolute value of the instantaneous maximum of the atmospheric electric field representing the entry into the thunderstorm stage; this value was 510.7, occurring at 16:59:00. The warning time is set at 16:47:50, which corresponds to the instantaneous maximum value of the thunderstorm approaching the stage. The warning time is set at 16:59:00, which corresponds to the time when the thunderstorm enters the passing stage. The warning time lead is 11 minutes and 10 seconds.

[0100] Table 5 shows the selection of absolute values ​​of instantaneous maximum values ​​used for early warning and the corresponding warning times for seven thunderstorm events. Among the four thunderstorm events in 2016, the maximum instantaneous maximum value for thunderstorm imminent warning was 90.8 V / m, the minimum was 45.4 V / m, and the average was 66.42 V / m; the maximum instantaneous maximum value for warning at the start of thunderstorm passage was 510.7 V / m, the minimum was 238.3 V / m, and the average was 366.225 V / m; the maximum warning lead time was 32 minutes and 31 seconds, the minimum was 11 minutes and 10 seconds, and the average was 24 minutes and 11 seconds. Therefore, the threshold and warning strategy for lightning warning based on changes in the atmospheric electric field are determined as follows: The average value of the d5 coefficients of the wavelet decomposition is calculated sequentially every 10 seconds according to the observation frequency. 45.4 V / m and 90.8 V / m are respectively used as the lower limit and upper limit of the warning value for the approaching thunderstorm, and 238.3 V / m and 510.7 V / m are respectively used as the lower limit and upper limit of the warning value for the start of the thunderstorm passage.

[0101] Table 5. Warning value selection and corresponding warning time for four thunderstorm events in 2016.

[0102] Thunderstorm process date Thunderstorm imminent warning value v / m Thunderstorm transit warning value v / m Thunderstorm approaching warning time point Thunderstorm transit warning time point Warning lead time 20160730 45.4 238.3 13:35:04 14:07:35 00:32:31 20160803 66.9 510.7 16:47:50 16:59:00 00:11:10 20160814 90.8 314.3 14:37:25 14:57:53 00:20:28 20160819 60.9 439.2 6:49:42 7:10:13 00:20:31 average value 66.42 366.225 — — 00:21:10

[0103] The wavelet decomposition-based lightning warning method described above can determine the timing of thunderstorm warnings, meeting the requirements for lightning warnings and improving the accuracy of warnings. To verify the accuracy of the method, this invention also examines the wavelet decomposition-based lightning warning method.

[0104] from Figure 3 , Figure 5 and Figure 9 It can be seen that after a period of drastic change in the atmospheric electric field and a gradual return to stability, the thunderstorm process is followed by another fluctuation in the atmospheric electric field. To address this potential situation or multiple thunderstorms occurring within a single day, a condition for ending the warning has been added to the previous warning model.

[0105] After a thunderstorm transit warning, when the absolute value of the average d5 coefficient of the wavelet decomposition over a consecutive 10-minute period is less than 5, this point is defined as the warning end time. Then, the previous warning pattern is repeated for a second warning. To verify the warning effect, the modified warning method was used to monitor three thunderstorm events that occurred in 2017. Figure 7 — Figure 9 Thunderstorm warnings were issued, and the warning results are shown in Table 6.

[0106] Compare separately Figure 7 — Figure 9 The changing trends of the atmospheric electric field and the stages of ground flashes show that the timing of the warnings corresponds well with the time periods when the thunderstorms are approaching and passing through. In terms of the lead time, the longest warning time for the three thunderstorm events was 33 minutes and 6 seconds, the shortest was 11 minutes and 43 seconds, and the average was 20 minutes and 28 seconds, indicating that the overall warning effect was quite satisfactory.

[0107] Table 6. Warning value selection and corresponding warning time for four thunderstorm events in 2017

[0108]

[0109] The wavelet decomposition-based lightning warning method of this invention can also determine the direction of thunderstorm movement by using lightning warning times from different directional stations. By utilizing atmospheric electric field data from multiple stations to form a comprehensive monitoring network for thunderstorms and lightning activity across the entire region, it is possible to provide information on the distribution of the ground electric field, as well as the distribution and movement path of the spatial charge of thunderstorms within the monitoring area. This will help improve lightning forecasting and warning capabilities.

[0110] Based on the location relationship of the three stations, Wuyingshan Station (116.97555E, 36.69N), Agricultural Academy Station (117.064322E, 36.695938N), and Suncun Town Station (117.317285E, 36.695090N) are basically distributed from west to east. The straight-line distance between Wuyingshan Station and Agricultural Academy Station is 9 kilometers, and the distance between Agricultural Academy Station and Suncun Town Station is 22 kilometers. The latitude distribution of the three stations is basically the same. Therefore, theoretically, by observing the network of the three stations, the direction of thunderstorms moving in the east-west or slightly east-west direction can be determined based on the abnormal changes in the atmospheric electric field observed in real time and the order of warnings.

[0111] This invention utilizes a proposed lightning warning method to statistically analyze and compare the warning times at three stations during six thunderstorm events. Based on the chronological order and location of the two earliest warning stations, the east-west direction of the thunderstorm's movement is determined and compared with historical radar data of each thunderstorm event. The direction of the thunderstorm's movement, as observed by radar, is determined according to the temporal position of the thundercloud echo region. Furthermore, the direction of the thunderstorm's movement, determined by atmospheric electric field data from different stations, is determined based on the chronological order of warning issuance.

[0112] Table 7 of this invention presents the verification results of the movement direction of six thunderstorm events determined using atmospheric electric field data. The results show that the two methods were consistent in determining the east-west movement direction of the thunderstorm events in five instances. However, the thunderstorm event on August 14, 2016, could not be determined because the Wuyingshan station in the west did not issue a warning, the data from the Agricultural Academy station was incorrect, and only the Suncun Town station issued a warning. From the radar echo image of this event, it is clear that the thunderstorm's path when passing through Jinan was from west to east through the southern part of Jinan, not covering the urban area of ​​Jinan, which is why the Wuyingshan station did not issue a warning.

[0113] Table 76 shows the warning time and direction determination of thunderstorm events.

[0114] Thunderstorm process date Suncun Town Station Warning Time Agricultural Science Academy Station Warning Time Wuyingshan Station Warning Time Direction judgment Radar image and upper-level wind data direction Verification results (east-west direction) 20160803 16:47:50 / 16:59:34 East → West Southeast → Northwest Consistent 20160814 14:37:25 / — Unable to determine West → East Inconsistent 20160819 6:49:42 / 4:44:06 West → East Northwest → Southeast Consistent 20170726 14:22:48 13:40:30 13:31:23 West → East Southwest → Northeast Consistent 20170818 13:48:19 12:54:43 12:50:45 West → East Southwest → Northeast Consistent 20170823 17:25:57 — 16:48:19 West → East Southwest → Northeast Consistent

[0115] Note: " / " indicates an error in the observation data, and "—" indicates that no warning was issued.

[0116] Based on the wavelet decomposition-based lightning warning method provided by this invention, in selecting the lightning warning threshold, the high-frequency characteristics of the reconstructed sequence of the d5 component of the atmospheric electric field intensity wavelet decomposition can represent the abrupt changes in the atmospheric electric field before, during, and after the passage of a thunderstorm. The threshold set by the hard thresholding method can determine the time point of the thunderstorm warning. The typical threshold for thunderstorm warning (d5 wavelet component reconstructed sequence) has a minimum of 45.4 V / m, a maximum of 90.8 V / m, and an average of 63.3 V / m.

[0117] The present invention also uses the time sequence of atmospheric electric field data from different stations to issue early warnings to determine the direction of movement of six thunderstorm processes, and compares it with the corresponding radar echo data. The judgment results are consistent in five of them.

[0118] The following are embodiments of a lightning warning system based on wavelet decomposition provided in this disclosure. This system and the lightning warning method based on wavelet decomposition in the above embodiments belong to the same inventive concept. For details not described in detail in the embodiments of the lightning warning system based on wavelet decomposition, please refer to the embodiments of the lightning warning method based on wavelet decomposition described above.

[0119] The system includes: a data acquisition module, a wavelet decomposition and reconstruction module, a wavelet decomposition module, and a lightning early warning module;

[0120] The data acquisition module is used to acquire raw atmospheric data within a preset time period;

[0121] The wavelet decomposition and reconstruction module is used to select wavelet information, perform N-level wavelet decomposition on the noisy signal, configure preset thresholds for the coefficients of each level obtained by decomposition, and perform inverse wavelet transform on the denoised coefficients to obtain the reconstructed atmospheric data after denoising.

[0122] The wavelet decomposition module utilizes the sym4 wavelet function with 5 levels of decomposition to obtain the low-frequency component 'a' representing the approximate trend of atmospheric electric field variation and the high-frequency component 'd' representing the detailed part of atmospheric electric field. The atmospheric electric field data satisfies the following expression:

[0123] S = a5 + d1 + d2 + d3 + d4 + d5

[0124] Where S is the raw data of the atmospheric electric field, a5 is the low-frequency component, and d1 to d5 are the high-frequency components.

[0125] The lightning warning module is used to zero-paste the d5 coefficients, reconstructing the d5 coefficient components under the condition that the sampled atmospheric data satisfies the Nyquist law. The d5 coefficient components are used to detect abrupt changes in the atmospheric electric field and provide early warning of lightning.

[0126] The lightning warning system based on wavelet decomposition provided by this invention can summarize atmospheric data, efficiently collect, store, and process atmospheric data. It can monitor the three stages of the thunderstorm process, namely the approach stage, the transit stage, and the post-transit stage. Based on wavelet decomposition, the accuracy of lightning warning is improved, thereby realizing the timeliness and scientific nature of the whole process of lightning warning supervision and management.

[0127] The units and algorithm steps of the various examples described in the wavelet decomposition-based lightning early warning method of the present invention can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the present invention.

[0128] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A lightning early warning method based on wavelet decomposition, characterized in that, The methods include: Acquire raw atmospheric data within a preset time period; Select wavelet information and perform N-level wavelet decomposition on the noisy signal; Configure preset thresholds for the coefficients of each layer obtained from the decomposition; Perform inverse wavelet transform on the denoised coefficients to obtain the reconstructed atmospheric data after denoising; Using the sym4 wavelet function with 5-level decomposition, the low-frequency component 'a' representing the approximate trend of atmospheric electric field variation and the high-frequency component 'd' representing the detailed part of atmospheric electric field variation were obtained. The atmospheric electric field data satisfy the following expression: S = a5 + d1 + d2 + d3 + d4 + d5 Where S is the raw data of the atmospheric electric field, a5 is the low-frequency component, and d1 to d5 are the high-frequency components. The d5 coefficients were zero-padding to reconstruct the d5 coefficient components under the condition that the sampled atmospheric data satisfies the Nyquist law. The d5 coefficient components are used to detect abrupt changes in the atmospheric electric field and provide early warning of lightning. The steps for obtaining raw atmospheric data within a preset time period also include: The atmospheric data of a preset area is selected as the raw atmospheric data based on a preset date. The system acquires a preset amount of raw atmospheric data every minute and stores it as a file on a daily basis. The file is stored in a format that stores one raw atmospheric data point per second, with 60 raw atmospheric data points stored per minute on a single line. The method also includes data completion rules; The data completion rules are as follows: when the number of missing data lines per day is less than 60, the missing atmospheric data values ​​will be replaced by the average of the 10 data lines preceding that time; when the number of missing atmospheric data lines per day is more than 60, the entire file will be removed.

2. The lightning early warning method based on wavelet decomposition according to claim 1, characterized in that, The changes in the electric field during a thunderstorm can be divided into three stages: the approaching stage, the passing stage, and the post-passing stage.

3. The lightning early warning method based on wavelet decomposition according to claim 2, characterized in that, The method's characteristic in detecting abrupt change signals is that it analyzes the changes in the d5 coefficient and d5 reconstruction sequence of the atmospheric electric field during the approach and transit phases of thunderstorms using waveform images, thereby resolving the threshold of abrupt change points during the approach phase of thunderstorms and providing early warning for lightning occurring during the transit phase of thunderstorms.

4. A lightning early warning system based on wavelet decomposition, characterized in that, The system employs the wavelet decomposition-based lightning early warning method as described in any one of claims 1 to 3; The system includes: a data acquisition module, a wavelet decomposition and reconstruction module, a wavelet decomposition module, and a lightning early warning module; The data acquisition module is used to acquire raw atmospheric data within a preset time period; The wavelet decomposition and reconstruction module is used to select wavelet information, perform N-level wavelet decomposition on the noisy signal, configure preset thresholds for the coefficients of each level obtained by decomposition, and perform inverse wavelet transform on the denoised coefficients to obtain the reconstructed atmospheric data after denoising. The wavelet decomposition module utilizes the sym4 wavelet function with 5 levels of decomposition to obtain the low-frequency component 'a' representing the approximate trend of atmospheric electric field variation and the high-frequency component 'd' representing the detailed part of atmospheric electric field. The atmospheric electric field data satisfies the following expression: S = a5 + d1 + d2 + d3 + d4 + d5 Where S is the raw data of the atmospheric electric field, a5 is the low-frequency component, and d1 to d5 are the high-frequency components. The lightning warning module is used to zero-paste the d5 coefficients so that the d5 coefficient components can be reconstructed under the condition that the sampled atmospheric data satisfies the Nyquist law. The d5 coefficient components are used to detect abrupt changes in the atmospheric electric field and provide early warning of lightning. The steps for obtaining raw atmospheric data within a preset time period also include: The atmospheric data of a preset area is selected as the raw atmospheric data based on a preset date. The system acquires a preset amount of raw atmospheric data every minute and stores it as a file on a daily basis. The file is stored in a format that stores one raw atmospheric data point per second, with 60 raw atmospheric data points stored per minute on a single line. The method also includes data completion rules; The data completion rules are as follows: when the number of missing data lines per day is less than 60, the missing atmospheric data values ​​will be replaced by the average of the 10 data lines preceding that time; when the number of missing atmospheric data lines per day is more than 60, the entire file will be removed.

5. An early warning terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the wavelet decomposition-based lightning warning method as described in any one of claims 1 to 3.

Images