An AI technology-based radar level gauge intelligent debugging system and method

Abstract

The application discloses an AI technology-based radar level meter intelligent debugging system and method, relates to the technical field of measurement and control instruments, and acquires radar echo data and pre-processes the radar echo data through Bluetooth; dust interference values are calculated based on the data, dust is warned and continuously monitored, no dust is input into working conditions, the working conditions are diagnosed and analyzed by an AI large model and uploaded to the cloud for iteration, dust recovery returns to diagnosis, otherwise, a maintenance prompt is sent out. The application calculates dust interference values by extracting three core parameters, i.e., the number of wave peaks, the average value of a noise base and the peak signal-to-noise ratio, and compares the dust interference values with threshold values adaptively set based on historical statistics, so that accurate quantitative identification and early warning of dust interference are realized, false judgments caused by fixed threshold values are avoided, and the environmental adaptability and operation stability of the radar level meter under complex working conditions are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of measurement and control instrumentation technology, specifically to an intelligent debugging system and method for radar level gauges based on AI technology. Background Technology

[0002] With the maturity of 80GHz high-frequency millimeter-wave radar technology, its penetration, anti-interference, and measurement accuracy have been significantly improved. In existing technologies, raw echo curve data of 80GHz high-frequency millimeter-wave radar is acquired via Bluetooth. Combined with user-input working condition information, parameters are automatically matched and distributed by the built-in industrial knowledge graph and large language model. Then, a deep learning model is used to diagnose and analyze the echo curve data to measure stability, and cloud-based reinforcement learning is used for continuous evolution, enabling operation without professional intervention. However, during the feeding process of cement or flour silos, if the feeding time is too long, it means that a large amount of material continuously impacts the bottom of the silo at high speed, constantly raising the accumulated powder particles and rapidly replacing the dusty air in the silo. This causes the dust to accumulate continuously without time to settle, resulting in high concentrations of dust. The strong absorption and scattering of the millimeter-wave radar signal by the dust particles cause the radar receiving energy to decrease sharply. The echo curve changes from a clear single peak to a large-area and high-amplitude irregular noise envelope. At this time, the deep learning model trained based on the original samples cannot match an effective waveform pattern in the feature space, causing diagnostic logic confusion and distorted analysis results. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides an intelligent debugging system and method for radar level gauges based on AI technology. This solves the problem that existing radar intelligent debugging systems based on deep learning and knowledge graphs suffer from logical confusion due to the mismatch between training samples and extreme operating conditions when encountering high concentrations of dust that completely noise the echo signal. This results in the inability to extract effective waveforms from irregular noise, leading to severe distortion in diagnostic analysis.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an intelligent debugging method for radar level gauges based on AI technology, comprising the following specific steps: Step 1: Establish a communication connection with the radar via Bluetooth, acquire raw radar echo data, and perform preprocessing; Step 2: Perform dust interference analysis based on the preprocessed radar echo data to obtain dust interference values, analyze whether dust interference has occurred based on the dust interference values, and if a dust interference signal is detected, issue an interference warning and proceed to Step 4; if no dust interference signal is detected, proceed to Step 3; Step 3: Input on-site working condition information in text form, and perform parameter matching and diagnostic analysis on the on-site working condition information and radar echo data based on an AI large model, output the analysis results of radar measurement stability, upload the results to the cloud for iterative analysis of the analysis strategy, and return to Step 1 or end; Step 4: Continuously monitor the change in dust concentration after the interference warning, and if the dust concentration does not return to the normal range, issue a maintenance prompt and end; if the dust concentration returns to the normal range, return to Step 3.

[0005] Furthermore, the method for analyzing whether dust interference has occurred based on the dust interference value is as follows: a dust interference threshold is set, and the dust interference value is compared with the dust interference threshold. If the dust interference value is greater than the dust interference threshold, it indicates that dust interference has occurred; if the dust interference value is less than or equal to the dust interference threshold, it indicates that dust interference has not occurred.

[0006] Furthermore, the specific setting method of the dust interference threshold is as follows: under normal circumstances, the fluctuation range of dust interference values ​​is detected and statistically analyzed to obtain a set of normal dust interference values. The average value of each normal dust interference value in the set of normal dust interference values ​​is calculated to obtain the dust interference threshold.

[0007] Furthermore, the dust interference value is obtained as follows: by analyzing the original radar echo data, the number of peaks, the mean noise floor, and the peak signal-to-noise ratio are obtained. The number of peaks, the mean noise floor, and the peak signal-to-noise ratio are standardized and calculated in a comprehensive manner to obtain the dust interference value. ;in, Indicates the dust interference value. Indicates the number of peaks. Indicates the noise floor mean. Indicates peak signal-to-noise ratio. It represents a positive real number.

[0008] Furthermore, the noise floor mean is obtained as follows: the amplitude values ​​of the detection sampling points in the original radar echo data are extracted and the arithmetic mean is calculated to obtain the noise floor mean.

[0009] Furthermore, the specific method for obtaining the number of peaks is as follows: the echo amplitude of each sampling point is detected one by one, and all local maxima points that are higher than the amplitudes of the two adjacent points are selected; then, based on the current noise floor mean, a preset amplitude threshold is superimposed, and only valid peaks whose peak amplitude exceeds the sum of the noise floor mean and the set threshold are retained. Finally, the number of valid peaks is counted to obtain the number of peaks.

[0010] Furthermore, the specific method for obtaining the peak signal-to-noise ratio is as follows: First, among all the detected local maxima, the position of the main wave corresponding to the material surface during the last measurement is used as a reference point to lock the main wave peak corresponding to the current material surface and extract its peak amplitude; then, the logarithmic difference between this peak amplitude and the mean of the noise floor is calculated to obtain the peak signal-to-noise ratio.

[0011] Furthermore, the analysis method for the changing trend of the dust interference value is as follows: a detection time period is set. When the dust interference value is greater than the dust interference threshold, the detection time period is triggered. That is, within the detection time period, the difference between the dust interference value and the dust interference threshold at each moment is calculated, and the change value is obtained. The change value is iterated and compared with zero. If the proportion of the decrease or flat situation exceeds the proportion of the increase situation, it indicates that the dust concentration has recovered to the normal range; if the change value is greater than zero, it indicates that the dust concentration has not recovered to the normal range.

[0012] An AI-based intelligent debugging system for radar level gauges includes the following modules: a communication and data acquisition module, a dust interference identification module, an AI intelligent diagnosis and cloud iteration module, and a dust monitoring and maintenance prompt module. The communication and data acquisition module establishes a communication connection with the radar via Bluetooth, acquires raw radar echo data, and performs preprocessing. The dust interference identification module analyzes the preprocessed radar echo data to obtain dust interference values. Based on these values, it analyzes whether dust interference is detected. If dust interference is detected, an interference warning is issued, and the dust monitoring and maintenance prompt module is activated. If no dust interference is detected, the system detects dust interference. If a dust interference signal is detected, the AI ​​intelligent diagnosis and cloud iteration module will be executed. The AI ​​intelligent diagnosis and cloud iteration module takes in on-site operating condition information in text format, performs parameter matching and diagnostic analysis on the on-site operating condition information and radar echo data based on a large AI model, outputs the analysis results of radar measurement stability, uploads them to the cloud for iterative analysis of the analysis strategy, and returns to the communication and data acquisition module or terminates the process. The dust monitoring and maintenance prompt module continuously monitors changes in dust concentration after an interference warning. If the dust concentration does not return to the normal range, a maintenance prompt is issued, and the process terminates. If the dust concentration returns to the normal range, the process returns to the AI ​​intelligent diagnosis and cloud iteration module.

[0013] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: 1. By extracting three core feature parameters—the number of peaks, the mean noise floor, and the peak signal-to-noise ratio—the dust interference value is calculated comprehensively and compared with the adaptively set dust interference threshold. This achieves accurate quantitative identification of dust interference. When dust interference is detected, the system automatically issues an interference warning, prompting the user to intervene in the feeding time or pay attention to the on-site status. This avoids measurement inaccuracies caused by the deterioration of the echo signal due to dust and improves the environmental adaptability of the radar level gauge under extreme working conditions.

[0014] 2. By statistically analyzing the fluctuation range of dust interference values ​​under normal operating conditions and calculating its average value as the dust interference threshold, the threshold can be adaptively adjusted according to different tanks, media and environmental conditions, avoiding the problems of misjudgment or missed judgment caused by fixed thresholds. This method makes full use of historical operating data, improves the accuracy and robustness of dust identification, and ensures the stable operation of the system in complex industrial sites.

[0015] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0016] Figure 1 This is a flowchart of the intelligent debugging method for radar level gauges according to the present invention.

[0017] Figure 2 This is a structural diagram of the intelligent debugging system for radar level gauges according to the present invention. Detailed Implementation

[0018] 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.

[0019] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.

[0020] Example 1: like Figure 1As shown, this embodiment of the invention provides an intelligent debugging method for radar level gauges based on AI technology, including the following specific steps: Step 1: Due to the complex industrial environment, traditional wired connections are inconvenient. Bluetooth wireless connections enable contactless data interaction, avoiding the need for opening covers or wiring on-site, improving debugging safety. Then, a communication connection is established with the 80GHz high-frequency millimeter-wave radar via Bluetooth to obtain the raw radar echo data. The raw radar echo data is then cleaned to remove redundant values ​​and improve the quality of the raw radar echo data.

[0021] Step Two: By analyzing the raw radar echo data, the number of peaks, the mean noise floor, and the peak signal-to-noise ratio (SNR) are obtained. These parameters are then standardized to eliminate dimensional differences and convert values ​​of different orders of magnitude into a unified range. A comprehensive calculation is then performed to obtain the dust interference value. Based on this value, it is determined whether dust interference is present. If dust interference is detected, a signal is sent to the mobile app's interface via Bluetooth. Upon receiving the signal, the app plays a notification sound, informing the user of the presence of dust interference and prompting them to adjust the feeding time. Step Four is then executed. If no dust interference is detected, Step Three is executed.

[0022] Step 3: Input on-site operating condition information in text form via the APP, such as: medium type, tank structure, and process status. The lightweight natural language processing model built into the APP parses this text in real time, extracting key entities such as medium type, tank structure, and process status through word segmentation and entity recognition technology, and mapping them into high-dimensional semantic vectors. At the same time, it calculates the deviation of the number of peaks, the noise floor mean, and the peak signal-to-noise ratio, that is, subtracting the historical mean of each of the current number of peaks, the noise floor mean, and the peak signal-to-noise ratio, and then dividing by their respective historical standard deviations, thus obtaining a standard deviation that measures the deviation of the current value from the normal fluctuation range. The standardized values ​​of the surrounding data are used to form an echo feature vector. The high-dimensional semantic vector is concatenated with the echo feature vector and then input into a Transformer-based multimodal fusion encoder. A comprehensive state vector is generated through a cross-modal attention mechanism. Then, a locally deployed lightweight large model performs stability diagnosis in parallel based on this vector, outputting a measurement stability score and the cause of the anomaly. Finally, the structured results are returned to the APP interface. At the same time, the debugging data is encrypted and uploaded to the cloud. The basic model is iteratively trained through LoRA fine-tuning and the knowledge graph is updated to achieve continuous evolution of the diagnostic strategy. Then, depending on the user's choice, the process can return to step one to continue monitoring or end the process.

[0023] Step 4: Analyze the trend of dust interference value. If the dust concentration has not returned to the normal range, a maintenance prompt will be issued, indicating that there is a physical equipment malfunction, such as dust collector failure or poor sealing of the silo, and the process will end. If the dust concentration returns to the normal range, return to Step 3.

[0024] Example 2 differs from Example 1 in that: The method for analyzing whether dust interference has occurred based on dust interference values ​​is as follows: Set a dust interference threshold and compare the dust interference value with the dust interference threshold. If the dust interference value is greater than the dust interference threshold, it means that the dust is interfering; if the dust interference value is less than or equal to the dust interference threshold, it means that the dust is not interfering.

[0025] The specific method for setting the dust interference threshold is as follows: Under normal circumstances, the fluctuation range of dust interference values ​​is detected and statistically analyzed to obtain a set of normal dust interference values. The average value of each normal dust interference value in the set of normal dust interference values ​​is calculated to obtain the dust interference threshold.

[0026] The dust interference value is obtained as follows: The dust interference value is obtained by standardizing the number of peaks, the noise floor mean, and the peak signal-to-noise ratio and then calculating them together. ; in, Indicates the dust interference value. Indicates the number of peaks. Indicates the noise floor mean. Indicates peak signal-to-noise ratio. Representing positive real numbers, to avoid meaningless dust interference values, the larger the number of peaks, the higher the noise floor, the lower the signal-to-noise ratio, the more severe the dust interference.

[0027] The noise floor mean is obtained as follows: As dust concentration accumulates and the energy from countless scatterings is superimposed, the overall background noise gradually increases. Therefore, by selecting areas without target interference in the original radar echo data, such as the distance segment below the end of the range or the bottom of the warehouse, the amplitude values ​​of all sampling points in this area are extracted and the arithmetic mean is calculated to obtain the noise floor mean. This value is a quantitative indicator that reflects the level of environmental background noise.

[0028] The specific method for obtaining the number of peaks is as follows: Because dust generates numerous scattering sources instantly, many stray peaks immediately appear on the echo curve. Therefore, the echo amplitude of each sampling point is detected one by one across the entire range, from the radar antenna to the bottom of the warehouse, and all local maxima higher than the amplitudes of the two adjacent points are filtered out. Then, a preset amplitude threshold is superimposed on the current noise floor average, which is usually 3dB to 6dB according to engineering experience. If it is less than 3dB, the random fluctuations of the noise itself are easily included, leading to more false alarms. If it is greater than 6dB, some weak but real dust reflections will be missed, leading to more missed alarms. Therefore, in order to ensure that the number of peaks can truly reflect the density of stray reflection sources caused by dust in the environment, only valid peaks whose peak amplitude exceeds the sum of the noise floor average and the set threshold are retained, thus eliminating false peaks caused by noise fluctuations. Finally, the number of valid peaks is counted to obtain the peak count, which reflects the density of stray reflection sources caused by dust in the current environment.

[0029] The specific method for obtaining the peak signal-to-noise ratio is as follows: First, among all the detected local maxima, the position of the main wave corresponding to the material surface during the last measurement is used as a reference point. A search interval is set near the current position to accurately locate the main wave peak corresponding to the current material surface and extract its peak amplitude. Then, the logarithmic difference between this peak amplitude and the noise floor mean is calculated to obtain the peak signal-to-noise ratio (PSNR), which is in decibels. This PSNR represents the absorption and attenuation effect of dust on the radar signal. This peak amplitude decreases as the dust concentration increases. At the same time, dust scattering causes the noise floor mean to rise, and the logarithmic difference between this peak amplitude and the noise floor mean decreases. Therefore, the smaller the PSNR, the weaker the material surface echo or the stronger the background noise, and the more serious the dust interference.

[0030] The method for analyzing the changing trend of dust interference values ​​is as follows: A detection period is set. When the dust interference value is greater than the dust interference threshold, the detection period is triggered. Within the detection period, the difference between the dust interference value and the dust interference threshold at each moment is calculated, and the change value is obtained. The change values ​​are iterated and compared with zero. If the proportion of decrease or flat values ​​exceeds the proportion of increase values, it means that the dust concentration has returned to the normal range. If all change values ​​are greater than zero, it means that the dust concentration has not returned to the normal range.

[0031] Example 3: like Figure 2 As shown: An intelligent debugging system for radar level gauges based on AI technology includes the following specific modules: Communication and data acquisition module: Establishes a communication connection with the radar via Bluetooth, acquires raw radar echo data, and performs preprocessing. Dust interference identification module: Based on preprocessed radar echo data, dust interference is analyzed to obtain dust interference values. The module analyzes whether dust interference is detected. If a dust interference signal is detected, an interference warning is issued, and the dust monitoring and maintenance prompt module is executed. If no dust interference signal is detected, the AI ​​intelligent diagnosis and cloud iteration module is executed. AI Intelligent Diagnosis and Cloud Iteration Module: Input on-site working condition information in text form, and perform parameter matching and diagnostic analysis on the on-site working condition information and radar echo data based on the AI ​​large model. Output the analysis results of radar measurement stability, upload to the cloud to iterate the analysis strategy, and return to the communication and data acquisition module or end. Dust monitoring and maintenance prompt module: continuously monitors the change in dust concentration after interference warning. If the dust concentration does not return to the normal range, a maintenance prompt is issued and the process ends. If the dust concentration returns to the normal range, the process returns to the AI ​​intelligent diagnosis and cloud iteration module.

[0032] The preferred embodiments disclosed in this invention are merely illustrative examples of feasible implementation methods and are not intended to exhaust all technical details of the invention, nor do they constitute a limitation on the scope of protection of the invention. Those skilled in the art should understand that the core concept of this invention lies in acquiring raw radar echo data via Bluetooth, extracting three core feature parameters—the number of peaks, the noise floor mean, and the peak signal-to-noise ratio—and comprehensively calculating the dust interference value. This value is then compared with an adaptively set threshold to achieve accurate quantitative identification and graded early warning of dust interference. Simultaneously, combined with user-input textual operating condition information, multimodal fusion and large-model diagnosis are used to output measurement stability analysis results, and the diagnostic strategy is continuously optimized through cloud-based iteration. Finally, trend analysis distinguishes between process dust and equipment faults, achieving closed-loop control and maintenance prompts. This overall technical concept covers the complete chain from data acquisition, feature extraction, intelligent diagnosis to closed-loop control, demonstrating a high degree of systematicity and innovation.

[0033] In practical applications, those skilled in the art can make appropriate adjustments, combinations, or substitutions to the methods or systems described in this embodiment based on specific production conditions, equipment configurations, and process requirements, without departing from the core concept of this invention. For example, regarding the acquisition method, in addition to Bluetooth, wireless or wired communication methods such as Wi-Fi, 4G / 5G, and industrial Ethernet can also be used; regarding the data processing algorithm, wavelet transform and Hilbert-Huang transform can be used to replace the local maximum method for peak detection, quantile statistics or adaptive threshold algorithms can be used to replace the simple arithmetic mean for noise basis extraction, and the peak signal-to-noise ratio can be calculated by taking the logarithm of the linear domain ratio instead of direct difference calculation; regarding the control threshold, in addition to being based on the historical mean, the dust interference threshold can also be set dynamically by combining the standard deviation, or adaptively generated using a machine learning classifier; regarding the specific implementation of the execution unit, early warning prompts can be made using various methods such as APP push, SMS alarm, and central control room pop-ups, and maintenance prompts can be linked to the factory operation and maintenance system to automatically generate work orders. These adjustments are all reasonable modifications under the core concept of this invention and should not be considered as departing from the protection scope of this invention.

[0034] Furthermore, the technical concepts disclosed in this invention possess universal scalability and adaptability. They are not only applicable to powder measurement scenarios such as cement silos and flour silos in the embodiments described, but can also be applied in similar technical fields or related industrial processes through analogy, transplantation, or improvement. For example, dust interference problems also exist in other powder storage scenarios such as coal powder silos, plastic particle silos, and feed silos, and the technical solution of this invention can be directly transplanted. In liquid storage tanks, although there is no dust interference, interference sources such as foam and steam can be identified using a similar approach; only the characteristic parameters need to be adjusted to be sensitive indicators for foam or steam. In high-dust environments such as mining and metallurgy, the dust identification and adaptive adjustment method of this invention also has application value. Any technical solution formed by making logically equivalent substitutions, reasonable adjustments to the order of steps, or recombination of module functions based on the principles, ideas, or framework disclosed in this specification should be considered to fall within the spirit and scope of this invention.

[0035] It should be further clarified that the specific descriptions and drawings in the patent documents are only for assisting in understanding the present invention, and their details should not be interpreted as limitations on the claims. For example, the specific values ​​given in the embodiments, such as the amplitude threshold of 3dB to 6dB, the LoRA fine-tuning method, and the Transformer encoder structure, are preferred examples in engineering practice, not the only implementation methods. In actual implementation, those skilled in the art can optimize these parameters or choose alternative technologies according to specific scenarios. As long as the functions and effects achieved are substantially the same as those of the present invention, they should be considered equivalent implementations of the present invention. The true scope of protection of the present invention should be determined by the content of the claims recorded in the authorized text, and should cover all equivalent technical solutions that comply with the provisions of the patent law under these claims. Equivalent technical solutions include, but are not limited to: technical solutions that achieve substantially the same functions and effects by substantially the same means, and that can be conceived by those skilled in the art after reading the specification of the present invention without creative effort; and technical solutions formed by simply replacing, decomposing, or merging the technical features of the present invention.

[0036] All implementation methods that achieve the same or similar functions and effects through reasonable changes in technical means under the guidance of the concept of this invention fall within the scope of protection sought by this invention. For example, changing the calculation of dust interference value in step two from addition and reciprocal form to weighted product or neural network output, as long as its essence is still a comprehensive measurement of the number of peaks, the noise floor mean, and the peak signal-to-noise ratio, should be considered an equivalent solution of this invention; changing the multimodal fusion in step three from Transformer to attention mechanism or other fusion strategies, as long as it still achieves joint encoding of text semantics and echo features, does not depart from the protection scope of this invention; changing the trend analysis in step four from difference ratio to slope judgment or machine learning prediction, as long as it is still used to distinguish whether dust has subsided, should be covered.

[0037] Therefore, the descriptions in this specification are merely illustrative. Any adjustments to implementation methods, equivalent substitutions of technical features, or further applications based on the concept of this invention, as long as they do not depart from the overall technical approach described in this invention, should be included within the scope of protection of this invention. We encourage those skilled in the art to innovate and optimize based on their understanding of the core of this invention and in conjunction with specific practices, so as to jointly promote the progress and development of related technologies. The disclosure of this invention aims to promote technical exchange and innovation, rather than to limit subsequent research and development directions. Any improvements or extensions based on this invention, as long as their core still relies on the dust interference quantitative identification, multimodal fusion diagnosis, and closed-loop control mechanism disclosed in this invention, should be authorized or licensed through legal means, while respecting originality, in order to maintain a healthy innovation ecosystem and intellectual property order.

Claims

1. A method for intelligent debugging of radar level gauges based on AI technology, characterized in that: The specific steps include the following: Step 1: Establish a communication connection with the radar via Bluetooth, acquire the raw radar echo data, and perform preprocessing. Step 2: Perform dust interference analysis based on the preprocessed radar echo data to obtain the dust interference value. Analyze whether dust interference has occurred based on the dust interference value. If a dust interference signal is detected, issue an interference warning and proceed to Step 4; otherwise, proceed to Step 3. Step 3: Input the on-site working condition information in text form, and perform parameter matching and diagnostic analysis on the on-site working condition information and radar echo data based on the AI ​​big model. Output the analysis results of radar measurement stability, upload them to the cloud to iterate the analysis strategy, and return to Step 1 or end. Step 4: Continuously monitor the change in dust concentration after the interference warning. If the dust concentration does not return to the normal range, issue a maintenance prompt and end the process; if the dust concentration returns to the normal range, return to Step 3.

2. The intelligent debugging method for radar level gauges based on AI technology according to claim 1, characterized in that: The method for analyzing whether dust interference has occurred based on dust interference values ​​is as follows: Set a dust interference threshold and compare the dust interference value with the dust interference threshold. If the dust interference value is greater than the dust interference threshold, it means that the dust is interfering; if the dust interference value is less than or equal to the dust interference threshold, it means that the dust is not interfering.

3. The intelligent debugging method for radar level gauges based on AI technology according to claim 2, characterized in that: The specific method for setting the dust interference threshold is as follows: Under normal circumstances, the fluctuation range of dust interference values ​​is detected and statistically analyzed to obtain a set of normal dust interference values. The average value of each normal dust interference value in the set of normal dust interference values ​​is calculated to obtain the dust interference threshold.

4. The intelligent debugging method for radar level gauges based on AI technology according to claim 3, characterized in that: The dust interference value is obtained as follows: By analyzing the raw radar echo data, the number of peaks, the mean noise floor, and the peak signal-to-noise ratio are obtained. The number of peaks, the mean noise floor, and the peak signal-to-noise ratio are then standardized and calculated to obtain the dust interference value. ； in, Indicates the dust interference value. Indicates the number of peaks. Indicates the noise floor mean. Indicates peak signal-to-noise ratio. It represents a positive real number.

5. The intelligent debugging method for radar level gauges based on AI technology according to claim 4, characterized in that: The noise floor mean is obtained as follows: The amplitude values ​​of the detection sampling points in the original radar echo data are extracted and the arithmetic mean is calculated to obtain the noise floor mean.

6. The intelligent debugging method for radar level gauges based on AI technology according to claim 5, characterized in that: The specific method for obtaining the number of peaks is as follows: The echo amplitude of each sampling point is detected one by one, and all local maxima points with amplitudes higher than the two adjacent points are selected. Then, based on the current noise floor mean, a preset amplitude threshold is superimposed, and only valid peaks with peak amplitudes exceeding the sum of the noise floor mean and the set threshold are retained. Finally, the number of valid peaks is counted to obtain the peak count.

7. The intelligent debugging method for radar level gauges based on AI technology according to claim 6, characterized in that: The specific method for obtaining the peak signal-to-noise ratio is as follows: First, among all the detected local maxima, the position of the main wave corresponding to the material surface during the last measurement is used as a reference point to lock the main wave peak corresponding to the current material surface and extract its peak amplitude; then, the logarithmic difference between this peak amplitude and the mean of the noise floor is calculated to obtain the peak signal-to-noise ratio.

8. The intelligent debugging method for radar level gauges based on AI technology according to claim 7, characterized in that: The analysis method for the changing trend of the dust interference value is as follows: A detection period is set. When the dust interference value is greater than the dust interference threshold, the detection period is triggered. Within the detection period, the difference between the dust interference value and the dust interference threshold at each moment is calculated, and the change value is obtained. The change values ​​are iterated and compared with zero. If the proportion of decrease or flat values ​​exceeds the proportion of increase values, it means that the dust concentration has returned to the normal range. If all change values ​​are greater than zero, it means that the dust concentration has not returned to the normal range.

9. An AI-based intelligent debugging system for radar level gauges, used to implement the AI-based intelligent debugging method for radar level gauges as described in any one of claims 1-8, characterized in that, The AI-based intelligent debugging system for radar level gauges includes: a communication and data acquisition module, a dust interference identification module, an AI intelligent diagnosis and cloud iteration module, and a dust monitoring and maintenance prompt module. Communication and data acquisition module: Establishes a communication connection with the radar via Bluetooth, acquires raw radar echo data, and performs preprocessing. Dust interference identification module: Based on preprocessed radar echo data, dust interference is analyzed to obtain dust interference values. The module analyzes whether dust interference is detected. If a dust interference signal is detected, an interference warning is issued, and the dust monitoring and maintenance prompt module is executed. If no dust interference signal is detected, the AI ​​intelligent diagnosis and cloud iteration module is executed. AI Intelligent Diagnosis and Cloud Iteration Module: Input on-site working condition information in text form, and perform parameter matching and diagnostic analysis on the on-site working condition information and radar echo data based on the AI ​​large model. Output the analysis results of radar measurement stability, upload to the cloud to iterate the analysis strategy, and return to the communication and data acquisition module or end. Dust monitoring and maintenance prompt module: continuously monitors the change in dust concentration after interference warning. If the dust concentration does not return to the normal range, a maintenance prompt is issued and the process ends. If the dust concentration returns to the normal range, the process returns to the AI ​​intelligent diagnosis and cloud iteration module.

Images