A liquid level wave fitting monitoring method and system based on liquid level monitor

Abstract

The present application provides a liquid level wave fitting monitoring method and system based on a liquid level monitor, which relates to the field of liquid level monitoring technology, including: the liquid level monitor obtains multiple liquid level wave signal sets corresponding to multiple liquid level position points according to a preset trajectory; performs approximation processing on the liquid level wave signal set; divides each liquid level wave signal into segments according to a fixed time window, and outputs the corresponding multiple segments of liquid level wave signals; performs cubic polynomial fitting on each segment of the liquid level wave signal to obtain multiple fitting waveforms corresponding to multiple liquid level wave signal sets; performs overall fitting on multiple fitting waveforms based on the distribution of multiple liquid level position points, outputs an overall fitting model for liquid level fluctuations, and calls the overall fitting model for liquid level fluctuations to monitor liquid level fluctuations. The present application solves the technical problems that traditional liquid level monitors mostly use fixed-point monitoring, have a small coverage range, and have insufficient fluctuation fitting accuracy, and improves the comprehensiveness, accuracy, and reliability of the monitoring results.

Description

Technical Field

[0001] The present application relates to the technical field of liquid level monitoring, and in particular to a liquid level wave fitting monitoring method and system based on a liquid level monitor. Background Art

[0002] Liquid level monitoring is widely used in the fields of chemical industry, energy, environment, etc. It is used to monitor the height, fluctuation and flow state of liquid in real time to ensure the stability and safety of the process. At present, liquid level monitors usually use fixed monitoring devices to monitor the liquid surface through sensors such as ultrasound, radar or floats. However, traditional liquid level monitoring methods have certain limitations. In large-area liquid surfaces or complex fluctuation environments, fixed-point monitoring is difficult to fully and accurately capture the overall fluctuation of the liquid level.

[0003] Existing liquid level monitors usually only obtain liquid level height data at local fixed positions, with a limited monitoring range that cannot effectively cover the entire liquid surface. Due to factors such as liquid level fluctuations, temperature changes, or container shape, the data provided by fixed monitoring points may not accurately reflect the true state of the entire liquid surface. In addition, in terms of data processing, existing technologies mostly use linear or first-order polynomial fitting methods to process liquid level fluctuation data. Although the calculation is simple, the fitting accuracy is low for complex fluctuation characteristics, especially nonlinear fluctuations, and it cannot accurately reflect the actual situation of liquid level fluctuations. Especially in scenarios where complex fluctuations occur frequently, the real-time response capability and accuracy of existing technologies are difficult to meet the needs of high-precision liquid level monitoring and cannot provide a reliable early warning mechanism. Summary of the invention

[0004] The present application provides a liquid level wave fitting monitoring method and system based on a liquid level monitor, which solves the technical problem that traditional liquid level monitors mostly use fixed-point monitoring, have a small coverage range and insufficient fluctuation fitting accuracy, resulting in limited accuracy and comprehensiveness of monitoring results, and achieves the technical effect of comprehensively and dynamically monitoring liquid level fluctuations, thereby improving the accuracy and reliability of monitoring results.

[0005] In view of the above problems, on the one hand, the present application provides a liquid surface wave fitting monitoring method based on a liquid level monitor, the method comprising: obtaining a movable liquid level monitor, the liquid level monitor moves according to a preset trajectory, and obtaining multiple liquid surface wave signal sets corresponding to multiple liquid surface position points; performing approximation processing on the multiple liquid surface wave signal sets to obtain multiple liquid surface wave signal sets after processing; setting a fixed time window, segmenting each liquid surface wave signal according to the fixed time window, and outputting multiple liquid surface wave signals corresponding to each liquid surface wave signal; using polynomial fitting to perform cubic polynomial fitting on each liquid surface wave signal in the multiple liquid surface wave signals to obtain multiple fitting waveforms corresponding to the multiple liquid surface wave signal sets; performing overall fitting on the multiple fitting waveforms based on the distribution of the multiple liquid surface position points, outputting an overall fitting model of liquid surface fluctuation, calling the overall fitting model of liquid surface fluctuation to monitor the liquid surface fluctuation, and triggering an alarm signal if the set fluctuation threshold is exceeded.

[0006] On the other hand, the present application also provides a liquid surface wave fitting monitoring system based on a liquid level monitor, the system comprising: a liquid surface wave signal acquisition module, the liquid surface wave signal acquisition module is used to acquire a movable liquid level monitor, the liquid level monitor moves according to a preset trajectory, and acquires multiple liquid surface wave signal sets corresponding to multiple liquid surface position points; a liquid surface signal processing module, the liquid surface signal processing module is used to perform approximation processing on the multiple liquid surface wave signal sets, and acquire multiple processed liquid surface wave signal sets; a liquid surface signal division module, the liquid surface signal division module is used to set a fixed time window, and divide each liquid surface according to the fixed time window. The wave signal is divided into segments, and multiple segments of liquid surface wave signals corresponding to each liquid surface wave signal are output; a waveform segmented fitting module, the waveform segmented fitting module is used to use polynomial fitting to perform cubic polynomial fitting on each segment of the multiple liquid surface wave signals, so as to obtain multiple fitting waveforms corresponding to the multiple liquid surface wave signal sets; a fluctuation overall fitting module, the fluctuation overall fitting module is used to perform overall fitting on the multiple fitting waveforms based on the distribution of the multiple liquid surface position points, output a liquid surface fluctuation overall fitting model, call the liquid surface fluctuation overall fitting model to monitor the liquid surface fluctuation, and trigger an alarm signal if the set fluctuation threshold is exceeded.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] A movable liquid level monitor is obtained, and the liquid level monitor moves according to a preset trajectory to obtain multiple liquid level wave signal sets corresponding to multiple liquid level position points. Compared with fixed point monitoring, mobile monitoring can cover more liquid level positions and more comprehensively reflect the characteristics of liquid level fluctuations, thereby improving the accuracy of liquid level fluctuation monitoring. Approximation processing is performed on the multiple liquid level wave signal sets to obtain multiple liquid level wave signal sets after processing; through approximation processing, noise and interference in the signal can be removed to improve the accuracy and reliability of the data. A fixed time window is set, and each liquid level wave signal is segmented according to the fixed time window, and multiple liquid level wave signals corresponding to each liquid level wave signal are output to facilitate subsequent analysis and fitting. The introduction of a fixed time window can better cope with situations where the fluctuation frequency is fast or the amplitude is large, refine the monitoring of liquid level fluctuations, improve the time resolution of signal processing, and thus improve the real-time performance and response speed of the overall monitoring. Polynomial fitting is used to perform cubic polynomial fitting on each liquid level wave signal in the multiple liquid level wave signals to obtain multiple fitting waveforms corresponding to the multiple liquid level wave signal sets. Compared with simple linear or quadratic fitting, cubic polynomials can fit the nonlinear characteristics of liquid level fluctuations more accurately, thereby more accurately reflecting the complex changes of liquid level fluctuations and improving the accuracy of liquid level fluctuation simulation. Based on the distribution of the multiple liquid level position points, the multiple fitting waveforms are fitted as a whole, and the overall fitting model of liquid level fluctuations is output, which ensures the global consistency and continuity of the model and improves the accuracy and stability of monitoring. The overall fitting model of liquid level fluctuations is called to monitor liquid level fluctuations. If the set fluctuation threshold is exceeded, an alarm signal is triggered to achieve timely warning of abnormal fluctuations.

[0009] To summarize, this application realizes comprehensive monitoring of large-area liquid levels by introducing mobile monitoring, approximation processing, multi-segment cubic polynomial fitting and the construction of an overall fitting model, and can accurately fit the liquid level fluctuation characteristics and capture the liquid level fluctuation in real time and accurately, thereby significantly improving the comprehensiveness and accuracy of the liquid level monitoring results, and providing reliable data support for real-time monitoring and abnormal alarm mechanisms.

[0010] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 A schematic diagram of a flow chart of a liquid level wave fitting monitoring method based on a liquid level monitor provided in an embodiment of the present application;

[0012] Figure 2A schematic diagram of a flow chart of obtaining a preset trajectory in a liquid level wave fitting monitoring method based on a liquid level monitor provided in an embodiment of the present application;

[0013] Figure 3 A schematic diagram of a flow chart of outputting an overall fitting model of liquid level fluctuations in a liquid level wave fitting monitoring method based on a liquid level monitor provided in an embodiment of the present application;

[0014] Figure 4 A schematic structural diagram of a liquid level wave fitting monitoring system based on a liquid level monitor provided in an embodiment of the present application.

[0015] Explanation of reference numerals: liquid surface wave signal acquisition module 10 , liquid surface signal processing module 20 , liquid surface signal division module 30 , waveform segmentation fitting module 40 , wave overall fitting module 50 . DETAILED DESCRIPTION

[0016] The embodiments of the present application provide a liquid level wave fitting monitoring method and system based on a liquid level monitor, thereby solving the technical problem that traditional liquid level monitors mostly use fixed-point monitoring, have a small coverage range, and have insufficient fluctuation fitting accuracy, resulting in limited accuracy and comprehensiveness of monitoring results. This achieves the technical effect of comprehensively and dynamically monitoring liquid level fluctuations, thereby improving the accuracy and reliability of monitoring results.

[0017] Embodiment 1, as Figure 1 As shown, the embodiment of the present application provides a liquid level wave fitting monitoring method based on a liquid level monitor, the method comprising:

[0018] Step S100: obtaining a movable liquid level monitor, wherein the liquid level monitor moves along a preset trajectory to obtain a plurality of liquid level wave signal sets corresponding to a plurality of liquid level position points.

[0019] Specifically, a movable liquid level monitor is a monitoring device that can change position and can move within a container according to a preset path or trajectory. This monitor is equipped with a sensor that can measure the liquid level, such as a pressure sensor, a capacitive sensor, or an ultrasonic sensor. Using a movable liquid level monitor to observe at different positions on the liquid surface can obtain more extensive and diverse data.

[0020] Before monitoring, one or more movement paths of the monitor are pre-set, i.e., preset trajectories. The preset trajectory contains multiple liquid level positions, each of which represents a specific coordinate. The liquid level monitor collects liquid level fluctuation signals at these positions. After the monitoring starts, the liquid level monitor moves on the liquid surface according to the preset trajectory, and collects corresponding liquid fluctuation signal data at each liquid level position point passed. These signal data include information such as fluctuation height, period, amplitude, etc., to form a liquid level wave signal set corresponding to each liquid level position point.

[0021] For example, to monitor the fluctuation of liquid in a pool, the liquid level monitor moves slowly along a preset circular trajectory. Every time a liquid level position point is passed, the liquid level monitor uses its configured ultrasonic sensor to send a signal and receive the reflected data to obtain a liquid level wave signal set of the current liquid level position point. By traversing the entire preset trajectory, the monitor covers all preset liquid level positions, thereby obtaining multiple liquid level wave signal sets.

[0022] Through mobile monitoring, the liquid level monitor can collect comprehensive liquid level fluctuation data over a large area, providing sufficient data support for subsequent signal processing, thereby improving the accuracy of the monitoring results.

[0023] Step S200: performing approximation processing on the plurality of liquid surface wave signal sets to obtain a plurality of processed liquid surface wave signal sets.

[0024] Specifically, the multiple liquid surface wave signal sets obtained by the liquid level monitor may contain interference introduced by factors such as sensor errors and environmental noise. Therefore, it is necessary to perform approximation processing on these liquid surface wave signal sets to eliminate the noise or interference in the signals and make the signals closer to the actual liquid surface fluctuations. Approximation processing can use signal processing technologies such as Fourier transform, wavelet transform, and filtering algorithms.

[0025] Exemplarily, wavelet transform technology is used to approximate multiple liquid surface wave signal sets. Wavelet transform can decompose complex signals into components of different scales. First, select appropriate wavelet basis functions (such as Haar wavelet, Daubechies wavelet, etc.) to perform multi-level decomposition of the signals in the liquid surface wave signal set to identify the noise information carried. Next, threshold processing is performed on the noise components. Reasonable threshold parameters are set, and high-frequency components exceeding the threshold are identified as noise and suppressed, while retaining low-frequency components that reflect the real liquid surface fluctuations. Finally, the processed low-frequency components are reconstructed by inverse wavelet transform to obtain the denoised liquid surface wave signal.

[0026] Through approximation processing, the noise or interference in the original fluctuation signal is removed, and the real information of the liquid level fluctuation is retained, providing a more accurate data basis for subsequent polynomial fitting and overall liquid level fluctuation monitoring.

[0027] Step S300: setting a fixed time window, segmenting each liquid surface wave signal according to the fixed time window, and outputting multiple segments of liquid surface wave signals corresponding to each liquid surface wave signal.

[0028] Specifically, in liquid level fluctuation monitoring, the acquired liquid surface wave signal is a continuous wave signal. In order to conduct a more accurate analysis and capture the fluctuation of the liquid surface in different time periods, it is necessary to segment each liquid surface wave signal and split the continuous data into shorter intervals.

[0029] First, according to the characteristics of liquid level fluctuations and monitoring requirements, a specific time length is pre-set as the time unit for analyzing liquid surface wave signals, namely, the time window. This time window is fixed. According to the set time window, each liquid surface wave signal is divided into multiple small segments, and multiple liquid surface wave signals corresponding to each liquid surface wave signal are generated. Each segment of the signal represents the liquid surface fluctuation within a time window, so that each segment of data can be processed independently in the future. For example, if the set time window is 5 minutes, then a liquid surface wave signal lasting 20 minutes will be divided into 4 segments, each lasting 5 minutes.

[0030] Step S400: performing cubic polynomial fitting on each segment of the multiple segments of liquid surface wave signals using polynomial fitting to obtain multiple fitting waveforms corresponding to the multiple liquid surface wave signal sets.

[0031] Specifically, cubic polynomial fitting is used to process each segmented liquid surface wave signal to capture the complex changes in the liquid surface fluctuation, such as peaks, troughs and changing curvature. The fitting process uses the least squares method to continuously adjust the parameters in the polynomial to find an optimal cubic polynomial function so that the curve of the function has the smallest deviation from the original liquid surface wave signal. After the fitting is completed, each liquid surface wave signal will correspond to a cubic polynomial function, and then a fitting waveform will be obtained. The fitting waveform can smoothly approximate the liquid surface fluctuation characteristics in each signal segment, thereby reflecting the overall trend and characteristics of the liquid surface fluctuation in each time window.

[0032] Step S500: Based on the distribution of the multiple liquid level position points, the multiple fitting waveforms are overall fitted, and an overall fitting model of liquid level fluctuation is output. The overall fitting model of liquid level fluctuation is called to monitor liquid level fluctuation, and an alarm signal is triggered if the set fluctuation threshold is exceeded.

[0033] Specifically, in order to accurately capture the fluctuation characteristics of the liquid level in the entire monitoring area and provide comprehensive data support for liquid level monitoring, it is necessary to integrate the multiple fitting waveforms generated above into a mathematical model that can comprehensively reflect the overall fluctuation characteristics according to the distribution of multiple liquid level position points, namely, the overall fitting model of liquid level fluctuation.

[0034] According to the distribution of multiple liquid level position points and the corresponding multiple fitting waveforms, statistical methods (such as principal component analysis PCA, mixed Gaussian model GMM, etc.) are used to comprehensively analyze the common characteristics of the fluctuations, and to construct an overall fitting model of liquid level fluctuations that can summarize the fluctuation patterns of all position points. During the liquid level fluctuation monitoring process, the real-time collected liquid level fluctuation data is input into the model to calculate the current liquid level fluctuation. If the calculated fluctuation amplitude exceeds the preset fluctuation threshold, an alarm signal will be triggered to alert the operator to possible abnormal conditions. Among them, the fluctuation threshold is a specific value set to determine whether the liquid level fluctuation is within the normal range.

[0035] Through the overall fitting model and alarm mechanism of liquid level fluctuation, the liquid level fluctuation can be monitored in real time, so as to promptly discover potential abnormal situations, effectively improving the safety and reliability of liquid level monitoring.

[0036] Further, such as Figure 2 As shown, the preset trajectory described in the embodiment of the present application is obtained by identifying the liquid surface scene, and the method includes:

[0037] Step S110: discretizing the position points of the liquid surface scene, and outputting N position points, where N is a positive integer, and each position point corresponds to an area on the liquid surface.

[0038] Step S120: Identify the first type of liquid level position points and the second type of liquid level position points of the N position points, wherein the first type of liquid level position points are position points whose liquid level fluctuation intensity is greater than or equal to a preset fluctuation intensity, and the second type of liquid level position points are position points whose liquid level fluctuation intensity is less than the preset fluctuation intensity.

[0039] Step S130: Select from the first type of liquid level position points and the second type of liquid level position points to obtain M position points that meet a preset coverage uniformity, where M is a positive integer less than N.

[0040] Step S140: constructing the preset trajectory with the M position points, and the liquid level monitor moves according to the preset trajectory.

[0041] Specifically, by identifying the liquid surface scene, different liquid surface position points are determined, and then a preset trajectory is constructed.

[0042] First, the continuous area in the liquid surface scene is divided into many small areas, each of which has a corresponding position point, and N position points are obtained. Among them, N represents the total number of position points obtained after segmentation, and N is a positive integer. Next, a reference value of liquid surface fluctuation is set, that is, the preset fluctuation intensity. According to the relative size of the liquid surface fluctuation intensity of each position point and the preset fluctuation intensity, these N position points are classified, and the points with fluctuation intensity greater than or equal to the preset fluctuation intensity are classified as the first type of liquid surface position points, and the points with fluctuation intensity less than the preset value are classified as the second type of liquid surface position points. Then, the first type of liquid surface position points and the second type of liquid surface position points are screened to select M position points. These M position points should evenly cover all liquid surface areas to effectively represent the overall situation of the entire liquid surface fluctuation and avoid excessive concentration or omission of the monitoring area. Finally, these M position points are connected to generate one or more preset tracks. The liquid level monitor moves on the liquid surface according to the track, evenly covering areas with different fluctuation intensities to ensure that comprehensive and effective data are collected.

[0043] For example, in a large reservoir, it is necessary to monitor the liquid level fluctuation. First, the entire liquid surface of the reservoir is discretized by the sensor, and the surface of the reservoir is divided into 100 positions (ie, N=100), which are evenly distributed on the water surface. Then, by analyzing the fluctuation intensity of each position point, the points with a fluctuation intensity greater than or equal to 20cm are marked as the first type of liquid surface position points (high fluctuation area), and the rest are marked as the second type of liquid surface position points (low fluctuation area). 20 (M=20) position points are selected from these 100 position points, which are evenly distributed throughout the reservoir area to ensure that the monitoring covers areas with strong fluctuations and relatively stable areas. Based on these 20 points, a preset monitoring trajectory is generated to ensure that when the liquid level monitor moves on the surface of the reservoir along this trajectory, it can access these 20 position points in sequence to conduct detailed monitoring of liquid level fluctuations.

[0044] Furthermore, the preset trajectory in the embodiment of the present application is obtained by identifying the liquid surface scene, and the method further includes:

[0045] Step S150: performing edge fluctuation recognition on the liquid surface scene to obtain an edge fluctuation index.

[0046] Step S160: If the edge fluctuation index is greater than a preset edge fluctuation index, a third type of liquid surface position point is generated, wherein the third type of liquid surface position point is an edge liquid surface position point based on a container shape in the liquid surface scene.

[0047] Step S170: Select from the first type of liquid level position points, the second type of liquid level position points and the third type of liquid level position points to update the preset trajectory.

[0048] Specifically, the fluctuation characteristics of the edge area are usually different from those in the middle of the liquid surface. Large fluctuations may occur due to factors such as the shape and structure of the container. When generating a preset trajectory, the fluctuation characteristics of the edge area must be considered and liquid surface position points must be set in the edge area to achieve comprehensive fluctuation monitoring.

[0049] First, the edge area of ​​the liquid container is monitored and analyzed to determine the fluctuation of the liquid surface at the edge. Since the shape of the container has a significant impact on the fluctuation of the liquid surface (such as corners, curved edges, etc.), the fluctuation in this area often has special rules. Through edge fluctuation recognition technology, a parameter reflecting the intensity of edge fluctuation is calculated, namely the edge fluctuation index, which is used to quantify the fluctuation at the edge.

[0050] The calculated edge fluctuation index is compared with the preset edge fluctuation index. If the edge fluctuation intensity exceeds the preset threshold, it means that the liquid surface edge fluctuates violently and additional monitoring points are needed. According to the container shape and edge fluctuation index, a third type of liquid surface position points are added. These points are located in the container edge area and are specifically used to monitor specific situations of edge fluctuation.

[0051] Then, the first, second and third types of liquid level position points are combined and screened among these position points to ensure that the selected monitoring points can evenly cover the entire liquid surface, especially the areas with more violent fluctuations and edge positions. Connect the new monitoring point distribution, update the original preset trajectory, generate a new preset trajectory, ensure that the liquid level monitor can move along the most reasonable trajectory, and further improve the comprehensiveness of the monitoring data.

[0052] Further, such as Figure 3 As shown, in the embodiment of the present application, the multiple fitting waveforms are fitted as a whole based on the distribution of the multiple liquid level position points, and the overall fitting model of the liquid level fluctuation is output. The method includes:

[0053] Step S510: performing neighborhood fitting waveform identification on each fitting waveform based on the distribution of the multiple liquid level position points, and obtaining a neighborhood fitting waveform corresponding to each fitting waveform, wherein the neighborhood fitting waveform of each fitting waveform is a fitting waveform corresponding to liquid level position points belonging to the same category and the distance between positions is less than a preset distance.

[0054] Step S520: Optimizing each fitting waveform according to the neighborhood fitting waveform corresponding to each fitting waveform, and outputting a plurality of optimized fitting waveforms.

[0055] Step S530: performing overall fitting on the multiple optimized fitting waveforms based on the distribution of the multiple liquid surface position points, and outputting an overall fitting model of the liquid surface fluctuation.

[0056] Specifically, for the overall fitting of multiple fitting waveforms, the first step is to identify the neighborhood fitting waveforms according to the distribution of each liquid surface position point, that is, for each fitting waveform, identify other fitting waveforms adjacent to it. These neighborhood waveforms usually come from the same category of liquid surface position points, and the distance between each other is less than a preset threshold to ensure similar fluctuation characteristics. First, the distance between the liquid surface position points corresponding to each fitting waveform is calculated, and the fitting waveform corresponding to the liquid surface position point whose distance is less than the preset distance is identified as the neighborhood fitting waveform. For each fitting waveform, the fitting waveforms adjacent to it are collected to form a neighborhood waveform set. Each waveform in this neighborhood waveform set has similarity. For example, if the fitting waveform of a certain position point represents an area with large fluctuations, that is, the first type of liquid surface position point, the adjacent fitting waveforms belonging to the same type of position point can be determined by neighborhood waveform recognition.

[0057] After obtaining the neighborhood fitting waveform, the parameters of the fitting waveform are fine-tuned or recalculated based on the information of each fitting waveform and its neighborhood waveform to generate an optimized fitting waveform, i.e., an optimized fitting waveform, to improve the accuracy and consistency of the fitting waveform. These optimized fitting waveforms are theoretically closer to the actual fluctuation conditions. For example, a fitting waveform can be smoothed by weighted averaging of adjacent waveforms to reduce the impact of noise and local fluctuations on the overall fluctuation trend. Finally, all optimized fitting waveforms are globally integrated and analyzed to construct an overall fitting model of liquid level fluctuations that covers the fluctuation characteristics of all position points. This model not only reflects the fluctuation conditions of each liquid level position point, but also captures the overall fluctuation characteristics.

[0058] Further, each fitting waveform is optimized according to the neighborhood fitting waveform corresponding to each fitting waveform, and further includes:

[0059] Step S521: Obtain the fluctuation trend difference characteristics between each fitting waveform and the neighboring fitting waveforms of each fitting waveform, including fluctuation amplitude, fluctuation frequency and fluctuation phase.

[0060] Step S522: construct an error function according to the fluctuation trend difference characteristics, use the error function to optimize the derivative continuity of each fitting waveform, and output multiple optimized fitting waveforms.

[0061] Specifically, each fitting waveform is optimized according to the neighborhood fitting waveform corresponding to each fitting waveform. First, the difference in fluctuation amplitude, fluctuation frequency and fluctuation phase between each fitting waveform and its corresponding neighborhood fitting waveform must be calculated. Among them, the fluctuation amplitude refers to the maximum height difference of the liquid surface fluctuation, the fluctuation frequency refers to the speed of the periodic change of the fluctuation, and the fluctuation phase refers to the starting point of the fluctuation in time. These difference characteristics reflect the inconsistency of the local fluctuation pattern and provide a basis for further optimization.

[0062] Then, a function that can comprehensively consider the fluctuation difference is defined, namely the error function, which takes the difference in fluctuation amplitude, fluctuation frequency and fluctuation phase as input and outputs a numerical value representing the degree of difference. The error function can be constructed based on methods such as variance or cosine similarity to quantify the deviation of each fitted waveform relative to its neighboring waveform.

[0063] Each fitted waveform is optimized using the constructed error function. By minimizing the error function, the parameter values ​​of the fluctuation amplitude, fluctuation frequency and fluctuation phase are optimized to ensure that the derivative (i.e., the rate of change) of the fitted waveform remains continuous between adjacent waveforms, eliminate the sudden changes and uneven parts between the waveforms, and enable the fitted waveform to form a smooth transition between different monitoring points, so as to better capture the true situation of the liquid level fluctuation. The optimization process can minimize the error function through methods such as gradient descent, Newton's method, and Levenberg-Marquardt algorithm, thereby gradually optimizing the shape of each fitted waveform and finally obtaining multiple optimized fitted waveforms.

[0064] The optimized fitting waveform generated through the above waveform optimization process is more natural and smooth in terms of fluctuation amplitude, frequency and phase. It not only enhances the accuracy of local fluctuation details, but also takes into account the consistency of the overall fluctuation trend, and can better reflect the actual situation of liquid level fluctuation.

[0065] Furthermore, based on the distribution of the plurality of liquid level position points, the plurality of optimized fitting waveforms are fitted as a whole, and an overall fitting model of liquid level fluctuation is output, further comprising:

[0066] Step S531: performing overall fitting on the multiple optimized fitting waveforms based on the distribution of the multiple liquid surface position points, and outputting an overall fitting model of the initial liquid surface fluctuation.

[0067] Step S532: Perform global anomaly correction on the initial liquid level fluctuation overall fitting model, and output the corrected liquid level fluctuation overall fitting model, wherein the global anomaly correction includes frequency consistency anomaly correction of the global liquid level fluctuation and amplitude anomaly correction of the global liquid level fluctuation.

[0068] Specifically, the generation of the overall fitting model of the liquid level fluctuation is first based on the distribution of multiple liquid surface position points. The optimized fitting waveforms scattered at different position points are integrated into a complete liquid surface fluctuation model to generate an initial overall fitting model of the liquid surface fluctuation, so as to preliminarily reflect the overall fluctuation of the liquid surface.

[0069] In the liquid surface fluctuation model, the frequency or amplitude of the fluctuation should remain consistent or change reasonably. If the fluctuation frequency or amplitude of certain locations deviates from the overall trend, this may represent an abnormal situation. For example, due to local disturbances or sensor errors, the fluctuation frequency of a certain location on the liquid surface is very high, far exceeding the fluctuation frequency of surrounding locations. At this time, the initial model needs to be globally corrected to eliminate local anomalies in the initial model. The correction process includes frequency consistency anomaly correction and amplitude anomaly correction.

[0070] Frequency consistency anomaly correction first needs to check the distribution of fluctuation frequencies of each point in the initial model and identify the anomalies that are significantly different from the global frequency trend. Signal processing techniques such as filtering algorithms, Fourier transforms, and wavelet transforms are used to adjust the frequency of the anomalies so that they tend to the overall frequency trend of the model to ensure the consistency of the global frequency.

[0071] Similarly, for amplitude anomaly correction, we first need to conduct a comprehensive check on the fluctuation amplitudes of all locations in the initial model. We need to find monitoring points where the fluctuation amplitudes due to emergencies or measurement errors are significantly beyond the global range. We can use smoothing techniques such as sliding window averaging or adaptive filtering to correct the abnormal amplitudes and return them to the normal fluctuation amplitude range.

[0072] Through frequency consistency anomaly correction and amplitude anomaly correction, the anomalies in the initial liquid level fluctuation overall fitting model are corrected, the abnormal noise is eliminated, and a more accurate and reliable liquid level fluctuation overall fitting model is generated. This model can more accurately reflect the general law of liquid level fluctuation.

[0073] Furthermore, the neighborhood fitting waveform corresponding to each fitting waveform is obtained, and the method further includes:

[0074] Step S511: Obtain the spatial distribution density of the multiple liquid surface position points.

[0075] Step S512: if the spatial distribution density is greater than or equal to the preset spatial distribution density, reduce the size of the preset distance; if the spatial distribution density is less than the preset spatial distribution density, increase the size of the preset distance.

[0076] Specifically, in the process of neighborhood fitting waveform recognition, the size of the preset distance has an important influence on the recognition result. The preset distance can be adjusted by the spatial distribution density of the liquid surface position points to ensure that an appropriate number of domain waveforms are obtained.

[0077] First, count the distribution of liquid surface position points in the entire monitoring area, calculate the number of position points per unit area or unit volume, and obtain the spatial distribution density of liquid surface position points. Adjust the preset distance based on the calculated spatial distribution density. If the spatial distribution density is greater than or equal to the preset spatial distribution density, it means that these position points are relatively dense and the distance between the position points is relatively close. In this case, it is necessary to reduce the preset distance to ensure that each fitted waveform is only associated with those very close position points, thereby improving accuracy, preventing too many fitted waveforms from participating in neighborhood calculations, and avoiding interference. Among them, the preset spatial distribution density is a reference density threshold that is pre-set according to the actual application environment and data distribution characteristics when fitting the waveform, as a benchmark for adjusting the neighborhood size.

[0078] On the contrary, if the spatial distribution density is less than the preset spatial distribution density, that is, the liquid surface position points are sparsely distributed, the preset distance needs to be increased to expand the neighborhood range, so as to ensure that each fitted waveform can have enough adjacent waveforms to participate in the neighborhood calculation. In areas with fewer position points, the overall trend of liquid surface fluctuations can be better captured by expanding the neighborhood of the fitted waveform.

[0079] By calculating the spatial distribution density of the liquid surface position points, the range of the neighborhood fitting waveform is dynamically adjusted. In dense areas, the preset distance is reduced to improve accuracy; in sparse areas, the preset distance is increased to ensure that each fitting waveform has enough neighborhood data for fitting. This process can effectively optimize the accuracy and robustness of the fitting waveform, ensuring that the optimized fitting waveform can accurately reflect the actual fluctuation of the liquid surface in both dense and sparse areas.

[0080] In summary, the liquid level wave fitting monitoring method based on the liquid level monitor provided in the embodiment of the present application has the following technical effects:

[0081] A movable liquid level monitor is obtained, and the liquid level monitor moves according to a preset trajectory to obtain multiple liquid level wave signal sets corresponding to multiple liquid level position points. Compared with fixed point monitoring, mobile monitoring can cover more liquid level positions and more comprehensively reflect the characteristics of liquid level fluctuations, thereby improving the accuracy of liquid level fluctuation monitoring. Approximation processing is performed on the multiple liquid level wave signal sets to obtain multiple liquid level wave signal sets after processing; through approximation processing, noise and interference in the signal can be removed to improve the accuracy and reliability of the data. A fixed time window is set, and each liquid level wave signal is segmented according to the fixed time window, and multiple liquid level wave signals corresponding to each liquid level wave signal are output to facilitate subsequent analysis and fitting. The introduction of a fixed time window can better cope with situations where the fluctuation frequency is fast or the amplitude is large, refine the monitoring of liquid level fluctuations, improve the time resolution of signal processing, and thus improve the real-time performance and response speed of the overall monitoring. Polynomial fitting is used to perform cubic polynomial fitting on each liquid level wave signal in the multiple liquid level wave signals to obtain multiple fitting waveforms corresponding to the multiple liquid level wave signal sets. Compared with simple linear or quadratic fitting, cubic polynomial can more accurately fit the nonlinear characteristics of liquid level fluctuations, thereby more accurately reflecting the complex changes of liquid level fluctuations and improving the accuracy of liquid level fluctuation simulation.

[0082] Based on the distribution of the multiple liquid level position points, each fitting waveform is identified by a neighborhood fitting waveform, and the neighborhood fitting waveform corresponding to each fitting waveform is obtained; each fitting waveform is optimized according to the neighborhood fitting waveform corresponding to each fitting waveform, and multiple optimized fitting waveforms are output; based on the distribution of the multiple liquid level position points, the multiple optimized fitting waveforms are fitted as a whole, and the overall fitting model of liquid level fluctuation is output. By optimizing and fitting each fitting waveform as a whole, the global consistency and continuity of the model are ensured, and the accuracy and stability of monitoring are improved. The overall fitting model of liquid level fluctuation is called to monitor liquid level fluctuations. If the set fluctuation threshold is exceeded, an alarm signal is triggered, thereby realizing timely early warning of abnormal fluctuations.

[0083] Overall, the embodiments of the present application achieve comprehensive monitoring of large-area liquid levels by introducing mobile monitoring, approximation processing, multi-segment cubic polynomial fitting and the construction of an overall fitting model, and can accurately fit the liquid level fluctuation characteristics and capture the liquid level fluctuation in real time and accurately, thereby significantly improving the comprehensiveness and accuracy of the liquid level monitoring results, and providing reliable data support for real-time monitoring and abnormal alarm mechanisms.

[0084] Embodiment 2, as Figure 4 As shown, an embodiment of the present application provides a liquid level wave fitting monitoring system based on a liquid level monitor, the system comprising:

[0085] The liquid surface wave signal acquisition module 10 is used to acquire a movable liquid level monitor, and the liquid level monitor moves according to a preset trajectory to acquire multiple liquid surface wave signal sets corresponding to multiple liquid surface position points.

[0086] The liquid level signal processing module 20 is used to perform approximation processing on the multiple liquid level wave signal sets to obtain multiple processed liquid level wave signal sets.

[0087] The liquid level signal division module 30 is used to set a fixed time window, divide each liquid level wave signal into segments according to the fixed time window, and output multiple segments of liquid level wave signals corresponding to each liquid level wave signal.

[0088] The waveform segmented fitting module 40 is used to perform cubic polynomial fitting on each segment of the multiple liquid surface wave signals by using polynomial fitting to obtain multiple fitting waveforms corresponding to the multiple liquid surface wave signal sets.

[0089] The fluctuation overall fitting module 50 is used to perform overall fitting on the multiple fitting waveforms based on the distribution of the multiple liquid surface position points, output an overall fitting model of the liquid surface fluctuation, call the overall fitting model of the liquid surface fluctuation to monitor the liquid surface fluctuation, and trigger an alarm signal if the set fluctuation threshold is exceeded.

[0090] Furthermore, the system of the embodiment of the present application obtains the preset trajectory by identifying the liquid surface scene, and the execution steps include:

[0091] The liquid surface scene is discretized into position points to output N position points, where N is a positive integer and each position point corresponds to an area on the liquid surface; the first type of liquid surface position points and the second type of liquid surface position points of the N position points are identified, where the first type of liquid surface position points are position points where the liquid surface fluctuation intensity is greater than or equal to a preset fluctuation intensity, and the second type of liquid surface position points are position points where the liquid surface fluctuation intensity is less than the preset fluctuation intensity; M position points that meet a preset coverage uniformity are selected from the first type of liquid surface position points and the second type of liquid surface position points, where M is a positive integer less than N; the preset trajectory is constructed using the M position points, and the liquid level monitor moves according to the preset trajectory.

[0092] Furthermore, the system of the embodiment of the present application obtains the preset trajectory by identifying the liquid surface scene, and the execution step also includes:

[0093] Perform edge fluctuation identification on the liquid surface scene to obtain an edge fluctuation index; if the edge fluctuation index is greater than a preset edge fluctuation index, generate a third type of liquid surface position point, wherein the third type of liquid surface position point is an edge liquid surface position point based on a container shape in the liquid surface scene; select from the first type of liquid surface position points, the second type of liquid surface position points and the third type of liquid surface position points to update the preset trajectory.

[0094] Furthermore, the steps executed by the overall fluctuation fitting module 50 in the embodiment of the present application include:

[0095] Based on the distribution of the multiple liquid level position points, a neighborhood fitting waveform is identified for each fitting waveform to obtain a neighborhood fitting waveform corresponding to each fitting waveform, wherein the neighborhood fitting waveform of each fitting waveform is a fitting waveform corresponding to liquid level position points belonging to the same category and the distance between positions is less than a preset distance; each fitting waveform is optimized according to the neighborhood fitting waveform corresponding to each fitting waveform, and a plurality of optimized fitting waveforms are output; based on the distribution of the multiple liquid level position points, the multiple optimized fitting waveforms are fitted as a whole, and an overall fitting model of liquid level fluctuation is output.

[0096] Furthermore, the overall wave fitting module 50 of the embodiment of the present application optimizes each fitting waveform according to the neighborhood fitting waveform corresponding to each fitting waveform, and the execution steps include:

[0097] Obtain the fluctuation trend difference characteristics between each fitting waveform and the neighboring fitting waveforms of each fitting waveform, including fluctuation amplitude, fluctuation frequency and fluctuation phase; construct an error function according to the fluctuation trend difference characteristics, use the error function to optimize the derivative continuity of each fitting waveform, and output multiple optimized fitting waveforms.

[0098] Furthermore, the overall fitting module 50 of the embodiment of the present application performs overall fitting on the multiple optimized fitting waveforms based on the distribution of the multiple liquid level position points and outputs an overall fitting model of the liquid level fluctuation, and the execution steps include:

[0099] Based on the distribution of the multiple liquid level position points, the multiple optimized fitting waveforms are fitted as a whole, and an overall fitting model of the initial liquid level fluctuation is output; the overall fitting model of the initial liquid level fluctuation is subjected to global anomaly correction, and the overall fitting model of the liquid level fluctuation that has passed the correction is output, wherein the global anomaly correction includes frequency consistency anomaly correction of the global liquid level fluctuation and amplitude anomaly correction of the global liquid level fluctuation.

[0100] Furthermore, the overall wave fitting module 50 of the embodiment of the present application obtains the neighborhood fitting waveform corresponding to each fitting waveform, and the execution step further includes:

[0101] Obtain the spatial distribution density of the multiple liquid surface position points; if the spatial distribution density is greater than or equal to the preset spatial distribution density, reduce the size of the preset distance; if the spatial distribution density is less than the preset spatial distribution density, increase the size of the preset distance.

[0102] Through the above-mentioned detailed description of a liquid level wave fitting monitoring method based on a liquid level monitor in this specification, technical personnel in this field can clearly understand that this embodiment is a liquid level wave fitting monitoring system based on a liquid level monitor. For the system disclosed in Example 2, since it corresponds to the method disclosed in Example 1 and has corresponding functional modules and beneficial effects, the relevant parts can be referred to the description of the method part.

[0103] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be 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 present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A liquid level wave fitting monitoring method based on a liquid level monitor, characterized in that: The method comprises: Acquire a movable liquid level monitor, wherein the liquid level monitor moves along a preset trajectory to acquire a plurality of liquid level wave signal sets corresponding to a plurality of liquid level position points; performing approximation processing on the plurality of liquid surface wave signal sets to obtain a plurality of processed liquid surface wave signal sets; Setting a fixed time window, segmenting each liquid surface wave signal according to the fixed time window, and outputting multiple segments of liquid surface wave signals corresponding to each liquid surface wave signal; Performing cubic polynomial fitting on each segment of the multiple segments of liquid surface wave signals by using polynomial fitting to obtain multiple fitting waveforms corresponding to the multiple liquid surface wave signal sets; Based on the distribution of the multiple liquid level position points, the multiple fitting waveforms are fitted as a whole, an overall fitting model of liquid level fluctuation is output, the overall fitting model of liquid level fluctuation is called to monitor liquid level fluctuation, and an alarm signal is triggered if the set fluctuation threshold is exceeded; The preset trajectory is obtained by identifying the liquid surface scene, and the method includes: Performing position point discretization processing on the liquid surface scene, and outputting N position points, where N is a positive integer, and each position point corresponds to an area on the liquid surface; Identify a first type of liquid level position point and a second type of liquid level position point of the N position points, wherein the first type of liquid level position point is a position point where the liquid level fluctuation intensity is greater than or equal to a preset fluctuation intensity, and the second type of liquid level position point is a position point where the liquid level fluctuation intensity is less than the preset fluctuation intensity; Selecting from the first type of liquid level position points and the second type of liquid level position points to obtain M position points that meet a preset coverage uniformity, where M is a positive integer less than N; The preset trajectory is constructed by the M position points, and the liquid level monitor moves according to the preset trajectory; Wherein, the preset trajectory is obtained by identifying the liquid surface scene, and the method further includes: Performing edge fluctuation recognition on the liquid surface scene to obtain an edge fluctuation index; If the edge fluctuation index is greater than the preset edge fluctuation index, a third type of liquid surface position point is generated, wherein the third type of liquid surface position point is an edge liquid surface position point based on a container shape in the liquid surface scene; Select from the first type of liquid level position points, the second type of liquid level position points and the third type of liquid level position points to update the preset trajectory.

2. A liquid level wave fitting monitoring method based on a liquid level monitor as claimed in claim 1, characterized in that: Based on the distribution of the plurality of liquid surface position points, the plurality of fitting waveforms are fitted as a whole, and an overall fitting model of liquid surface fluctuation is outputted. The method comprises: Performing neighborhood fitting waveform recognition on each fitting waveform based on the distribution of the multiple liquid level position points, and obtaining a neighborhood fitting waveform corresponding to each fitting waveform, wherein the neighborhood fitting waveform of each fitting waveform is a fitting waveform corresponding to liquid level position points belonging to the same category and the distance between positions of which is less than a preset distance; Optimizing each fitting waveform according to the neighborhood fitting waveform corresponding to each fitting waveform, and outputting a plurality of optimized fitting waveforms; Based on the distribution of the multiple liquid level position points, the multiple optimized fitting waveforms are overall fitted to output an overall fitting model of the liquid level fluctuation.

3. A liquid level wave fitting monitoring method based on a liquid level monitor as claimed in claim 2, characterized in that: Each fitting waveform is optimized according to the neighborhood fitting waveform corresponding to each fitting waveform, and the method includes: Obtaining fluctuation trend difference characteristics between each fitting waveform and the neighboring fitting waveforms of each fitting waveform, including fluctuation amplitude, fluctuation frequency and fluctuation phase; According to the fluctuation trend difference characteristics, an error function is constructed, and the derivative continuity of each fitting waveform is optimized using the error function to output multiple optimized fitting waveforms.

4. A liquid level wave fitting monitoring method based on a liquid level monitor as claimed in claim 2, characterized in that: Based on the distribution of the multiple liquid surface position points, the multiple optimized fitting waves are fitted as a whole, and an initial liquid surface fluctuation overall fitting model is output; The initial liquid level fluctuation overall fitting model is subjected to global anomaly correction, and the overall liquid level fluctuation fitting model that has passed the correction is output, wherein the global anomaly correction includes frequency consistency anomaly correction of the global liquid level fluctuation and amplitude anomaly correction of the global liquid level fluctuation.

5. The liquid level wave fitting monitoring method based on the liquid level monitor according to claim 2, characterized in that: Obtaining a neighborhood fitting waveform corresponding to each fitting waveform, the method further includes: Obtaining the spatial distribution density of the plurality of liquid surface position points; If the spatial distribution density is greater than or equal to the preset spatial distribution density, the size of the preset distance is reduced; if the spatial distribution density is less than the preset spatial distribution density, the size of the preset distance is increased.

6. A liquid level wave fitting monitoring system based on a liquid level monitor, characterized in that: The system is used to execute a liquid level wave fitting monitoring method based on a liquid level monitor according to any one of claims 1 to 5, comprising: A liquid surface wave signal acquisition module, the liquid surface wave signal acquisition module is used to acquire a movable liquid level monitor, the liquid level monitor moves according to a preset trajectory, and acquires a plurality of liquid surface wave signal sets corresponding to a plurality of liquid surface position points; A liquid level signal processing module, the liquid level signal processing module is used to perform approximation processing on the multiple liquid level wave signal sets to obtain multiple processed liquid level wave signal sets; A liquid level signal division module, the liquid level signal division module is used to set a fixed time window, divide each liquid level wave signal into segments according to the fixed time window, and output multiple segments of liquid level wave signals corresponding to each liquid level wave signal; A waveform segmented fitting module, the waveform segmented fitting module is used to perform cubic polynomial fitting on each segment of the multiple liquid surface wave signals by using polynomial fitting to obtain multiple fitting waveforms corresponding to the multiple liquid surface wave signal sets; The fluctuation overall fitting module is used to perform overall fitting on the multiple fitting waveforms based on the distribution of the multiple liquid surface position points, output an overall fitting model of the liquid surface fluctuation, call the overall fitting model of the liquid surface fluctuation to monitor the liquid surface fluctuation, and trigger an alarm signal if the set fluctuation threshold is exceeded.

Images