Method and device for detecting oil level of compressor

Abstract

This disclosure provides a method and apparatus for detecting the oil level in a compressor. The method includes: acquiring real-time operating parameters of the compressor and an image of the oil level to be measured; matching the timestamps corresponding to the real-time operating parameters and the image of the oil level to be measured; inputting the real-time operating parameters into an oil level height regression prediction model to output a confidence interval for the oil level height of the compressor; determining all candidate oil level heights of the compressor based on the image of the oil level to be measured using an image algorithm; and determining the current oil level height detection result of the compressor based on the confidence interval and the candidate oil level heights. This disclosure utilizes a joint decision-making mechanism based on multimodal data to achieve linked analysis of the oil level state based on time-series data of real-time operating parameters and oil level image data, thereby fully utilizing multimodal data during compressor operation to infer the oil level state and improving the prediction accuracy of compressor oil level detection.

Description

Technical Field

[0001] This disclosure relates to the field of compressor design technology, and in particular to a method and apparatus for detecting the oil level in a compressor. Background Technology

[0002] Industrial compressors in air conditioners, refrigerators, water heaters, and other appliances provide core power support for the stable operation of the entire refrigeration and heat pump system. Localized insufficient or thin oil levels in the compressor's friction pairs are a major cause of abnormal wear in its mechanical parts. Therefore, in-depth research into the occurrence scenarios and underlying mechanisms of insufficient oil conditions in compressors has significant theoretical and engineering value for improving compressor operational reliability. Rolling rotor compressors rely on pressure difference and centrifugal force to deliver lubricating oil to various moving mechanical components of the pump body through the oil suction port at the crankshaft end. As the core source of the lubrication circuit, the miscibility of the lubricating oil and refrigerant in the oil sump and the oil level have a crucial impact on the overall lubrication efficiency and heat dissipation performance of the machine. When the lubricating oil viscosity is too low, the oil level is insufficient, or there is severe vibration in the oil sump, localized insufficient or thin oil levels may occur in the friction pairs, thereby affecting the overall operational reliability of the compressor.

[0003] Compressor oil sump inspection focuses on three areas: the upper motor chamber, the lower oil sump, and the receiver. The focus of inspection differs significantly for each area: the upper motor chamber requires precise assessment to determine if there is excessive oil accumulation; the lower oil sump requires accurate oil level measurement to observe operating status; and the receiver requires close monitoring of internal oil levels and stability to ensure a balanced oil circulation supply to the compressor. Summary of the Invention

[0004] The technical problem to be solved by this disclosure is to overcome the defect of inaccurate oil level detection in compressors in the prior art, and to provide a method and device for detecting the oil level in compressors.

[0005] This disclosure solves the above-mentioned technical problems through the following technical solution:

[0006] This disclosure provides a method for detecting the oil level in a compressor, the method comprising:

[0007] The real-time operating parameters of the compressor and the oil level image of the compressor to be measured are obtained; wherein the timestamps corresponding to the real-time operating parameters and the oil level image to be measured are matched.

[0008] The real-time operating parameters are input into the oil level regression prediction model to output the confidence interval of the oil level of the compressor.

[0009] Based on the image of the oil level to be measured, all candidate oil level heights of the compressor are determined;

[0010] The current oil level detection result of the compressor is determined based on the confidence interval and the candidate oil level height.

[0011] Preferably, the step of determining the current oil level detection result of the compressor based on the confidence interval and the candidate oil level height includes:

[0012] In response to all candidate oil level heights being outside the confidence interval, the oil level height detection result of the previous moment is determined as the current oil level height detection result;

[0013] or,

[0014] In response to the fact that only one candidate oil level height is within the confidence interval, the candidate oil level height is determined as the current oil level height detection result;

[0015] or,

[0016] In response to several candidate oil level heights falling within the confidence interval, the median value of the confidence interval is determined;

[0017] Obtain the absolute value of the difference between each candidate oil level height and the median value;

[0018] The candidate oil level height corresponding to the smallest absolute value is determined as the current oil level height detection result.

[0019] Preferably, the oil level detection method for the compressor further includes:

[0020] The oil level in the upper chamber of the compressor motor, the lower oil sump, and the liquid receiver is detected simultaneously by a multi-window oil level data acquisition device.

[0021] The multi-window oil level data acquisition device includes several camera lenses parallel to the windows of the compressor.

[0022] Preferably, the step of determining all candidate oil level heights of the compressor based on the image of the oil level to be measured includes:

[0023] The image of the oil surface to be tested is binarized;

[0024] The edge detection results of the binarized oil surface image are obtained by using an image edge detection algorithm.

[0025] The edge detection results are subjected to Hough linear transform to detect all straight lines in the oil surface image to be tested;

[0026] Select the target line from all lines whose absolute slope value is less than a preset value;

[0027] The target straight line is determined as the candidate oil level height.

[0028] Preferably, the oil level height regression prediction model is obtained by training samples labeled with historical operating parameters and historical liquid level heights;

[0029] And / or,

[0030] The oil level detection method for the compressor also includes:

[0031] Using real-time operating parameters and current oil level detection results as samples, the oil level regression prediction model is iteratively optimized.

[0032] Preferably, the real-time operating parameters include at least one of the following: current intake pressure, current exhaust pressure, current intake temperature, current exhaust temperature, and current operating power.

[0033] This disclosure also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and used to run on the processor, wherein the processor executes the computer program to implement the above-described method for detecting the oil level of a compressor.

[0034] This disclosure also provides an oil level detection device for a compressor, the oil level detection device for the compressor including a multi-window oil level data acquisition device and electronic equipment as described above;

[0035] The multi-window oil surface data acquisition device includes several cameras, camera fixing clips, knob structure, perforated base plate, column, light source assembly, and light source connection plate.

[0036] Several cameras are fixed to the column by camera fixing clips, and the camera lenses are parallel to the viewport of the compressor; the camera fixing clips are integrated with the knob structure; the column is vertically fixed to the perforated base plate; the light source assembly is distributed around the camera periphery by the light source connecting plate;

[0037] The camera fixing clip is used to move the camera up and down along the column; the knob structure drives the camera to rotate; the light source assembly is used to provide a lighting environment.

[0038] This disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method for detecting the oil level of a compressor.

[0039] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method for detecting the oil level of a compressor.

[0040] Based on common knowledge in the field, the above-mentioned preferred conditions can be combined arbitrarily to obtain various preferred embodiments of this disclosure.

[0041] The positive and progressive effects of this disclosure are as follows:

[0042] This disclosure utilizes a joint decision-making mechanism based on multimodal data to achieve linked analysis of oil level status based on real-time operating parameter time-series data and oil level image data. This fully leverages multimodal data during compressor operation to infer oil level status, thereby improving the predictive accuracy of compressor oil level detection. Attached Figure Description

[0043] Figure 1 A first flowchart of a compressor oil level detection method provided as an exemplary embodiment of this disclosure;

[0044] Figure 2 A second flowchart of a compressor oil level detection method provided as an exemplary embodiment of this disclosure;

[0045] Figure 3 A schematic diagram of the structure of an oil level detection device for a compressor provided as an exemplary embodiment of this disclosure;

[0046] Figure 4 A schematic diagram illustrating a specific example of an oil level detection device for a compressor provided in an exemplary embodiment of this disclosure;

[0047] Figure 5 A physical illustration of a specific example of an oil level detection device for a compressor provided as an exemplary embodiment of this disclosure;

[0048] Figure 6 This is a schematic diagram of the structure of an electronic device provided as an exemplary embodiment of the present disclosure. Detailed Implementation

[0049] The present disclosure is further illustrated below by way of embodiments, but the present disclosure is not limited to the scope of the embodiments described herein.

[0050] Example 1

[0051] Currently, compressor oil level detection solutions are mainly divided into two technical categories. The first category involves performing difference image processing on the oil level image and a standard reference image, then using a preset threshold to quantitatively detect the oil level. This solution uses difference image analysis as its core detection method, but its detection accuracy has significant limitations. If the standard image and the image under test experience displacement during the acquisition phase, the accuracy of the final detection result will be significantly degraded. The second category achieves accurate oil level extraction through a series of image processing steps, including image binarization, edge extraction, Hough line transform, and slope filtering. This method combines the advantages of high detection speed and high operational stability, and has been widely used in compressor oil level detection scenarios. This embodiment is based on this detection method with targeted technical improvements.

[0052] However, both of the above methods have significant technical drawbacks: first, neither fully utilizes the operating data collected by sensors during compressor operation, making collaborative detection of multi-source data impossible; second, the screening process for candidate liquid level prediction lines is susceptible to significant external variable factors, requiring dynamic adjustment of screening conditions, ultimately leading to a need for further improvement in the robustness and generalization of the methods. Furthermore, no oil level detection method or corresponding supporting device for multiple viewing windows has yet been developed in the current technical field, leaving a significant gap in the relevant technological system.

[0053] Furthermore, most oil level detection devices currently on the market are limited to detecting the oil level in a single oil chamber, making it difficult to comprehensively characterize the oil level status of multiple related areas inside the compressor. Because the various areas inside the compressor are physically interconnected, the oil level status in different areas is interrelated and collectively affects the compressor's performance. Detection methods based on a single viewing window cannot meet the needs of a comprehensive assessment of the compressor's operating status.

[0054] During compressor operation, the accompanying operating parameters contain rich temporal information, providing crucial prior information for predicting oil level conditions. Core operating parameters reflect the dynamic balance of the lubrication system through different physical mechanisms. For example, temperature and pressure, as core parameters characterizing the thermodynamic state of the system, directly affect the lubrication system's performance through their synergistic effect. When either temperature or pressure abnormally increases, it not only leads to a significant decrease in lubricating oil viscosity and a reduction in oil film carrying capacity, but also exacerbates the oxidation and deterioration process of the lubricating medium, thereby affecting the lubrication reliability and service life of the system.

[0055] In summary, on the one hand, the internal structure of a compressor is complex and contains multiple connected regions. Joint observation of multiple regions can more accurately determine its internal oil level status. However, existing compressor oil level detection devices are mostly limited to observing the liquid level height of a single oil chamber, without considering the correlation between multiple oil chambers, resulting in a lack of global perspective in the observation results. On the other hand, the operating parameters generated during compressor operation are of significant reference value for oil level status diagnosis. However, current oil level detection algorithms based on artificial intelligence technology mostly rely solely on visual modality oil level imaging data, failing to effectively integrate the multi-dimensional operating parameters collected by sensors during compressor operation.

[0056] Therefore, there is an urgent need to propose a detection method and matching device that combines global observation capability with multimodal data fusion characteristics, so as to realize the global observation of the oil level status in each region inside the compressor, and to make full use of the time-series and visual multimodal data generated during the operation of the compressor to realize the real-time detection of the oil level status.

[0057] This embodiment proposes a multi-window, multi-modal compressor oil level detection method and device. On the one hand, through targeted hardware design, it achieves simultaneous detection of multiple windows in various regions inside the compressor, effectively overcoming the shortcomings of existing technologies that cannot perform global observation of the oil level in various regions inside the compressor. On the other hand, by leveraging a joint decision-making mechanism based on multi-modal data, it realizes linked analysis of oil level status based on time-series data of operating parameters and oil level imaging data, solving the problem that existing technologies cannot fully utilize multi-modal data during compressor operation for oil level status inference.

[0058] Figure 1 A flowchart illustrating a method for detecting the oil level in a compressor, as provided in an exemplary embodiment of this disclosure, is shown below. Figure 1 The methods for detecting the oil level in a compressor include:

[0059] S1. Obtain the real-time operating parameters of the compressor and the image of the oil level to be measured on the compressor.

[0060] Among them, the real-time operating parameters and the timestamps corresponding to the oil surface images to be measured are matched.

[0061] It should be noted that the real-time operating parameters and the measured oil level height in the oil level image are synchronized using the same absolute time base. For example, the real-time operating parameters are collected every three seconds, while the measured oil level image is collected every one second. Therefore, the timestamp of the real-time operating parameters is included in the timestamp of the measured oil level image. The timestamps are not completely identical, but they are matched. This can be understood as the oil level images at seconds 1, 2, and 3 being predicted using the confidence interval of the predicted oil level height based on the real-time operating parameters at second 1.

[0062] Real-time operating parameters include current intake pressure, current exhaust pressure, current intake temperature, current exhaust temperature, and current operating power.

[0063] S2. Input the real-time operating parameters into the oil level regression prediction model to output the confidence interval of the compressor's oil level.

[0064] S3. Based on the image of the oil level to be tested, determine all candidate oil level heights of the compressor.

[0065] S4. Determine the current oil level detection result of the compressor based on the confidence interval and candidate oil level heights.

[0066] In this embodiment, a joint decision-making mechanism based on multimodal data is used to realize the linkage analysis of oil level status based on real-time operating parameter time series data and oil level image data. This fully utilizes multimodal data during compressor operation to infer the oil level status and improves the prediction accuracy of compressor oil level detection.

[0067] In an optional implementation, step S4 includes:

[0068] If all candidate oil level heights are outside the confidence interval, the oil level height detection result from the previous moment is determined as the current oil level height detection result.

[0069] In an optional implementation, step S4 may further include:

[0070] If only one candidate oil level height is within the confidence interval, the candidate oil level height is determined as the current oil level height detection result.

[0071] In an optional implementation, step S4 may further include:

[0072] In response to several candidate oil level heights falling within a confidence interval, the median value of the confidence interval is determined.

[0073] Obtain the absolute value of the difference between the height of each candidate oil surface and the median value.

[0074] The candidate oil level height corresponding to the smallest absolute value is determined as the current oil level height detection result.

[0075] In this embodiment, based on the oil surface detection algorithm using image-based algorithms, a confidence interval from machine learning algorithms and statistical principles is introduced to model the time-series data generated by the sensor, thereby obtaining the confidence interval for the oil surface height. The optimal solution for the oil surface height detection result is obtained by jointly considering the relationship between the image detection algorithm and the confidence interval, further improving the prediction accuracy of oil surface detection.

[0076] In an optional implementation, the compressor oil level detection method further includes:

[0077] S5. The oil level in the upper chamber of the compressor motor, the lower oil sump, and the liquid receiver is simultaneously detected by the multi-window oil level data acquisition device.

[0078] The multi-window oil level data acquisition device includes several camera lenses parallel to the compressor's windows.

[0079] In this embodiment, by using multi-window detection of the compressor oil level, the shortcomings of existing oil level detection schemes, which can only detect a single window, are overcome, and the effect of global monitoring of the compressor's operating status is achieved.

[0080] In an optional implementation, step S3 includes:

[0081] S31. Frame-sampling is performed on the video data of the oil surface to be tested, with a frame-sampling frequency of intervals. Each frame is skipped once.

[0082] S32. Each frame of the oil surface image to be tested is cropped, rotated, and standardized preprocessed to improve image clarity.

[0083] S33. Perform filtering and noise reduction processing on the preprocessed oil surface image to remove random interference noise.

[0084] S34. Perform binarization processing on the noise-reduced oil surface image to be tested.

[0085] Binarization of the oil surface image can simplify the image information, separate the oil surface from the background, and lay the foundation for subsequent image processing.

[0086] S35. Obtain the edge detection results of the binarized oil surface image using an image edge detection algorithm.

[0087] The goal of image edge detection algorithms is to detect as many real edges as possible, to locate edges as accurately as possible, and to ensure that detected edges are as unique as possible, i.e., to find the optimal edge detection result. One image edge detection algorithm is the Canny Edge Detection algorithm.

[0088] S36. Perform Hough linear transform on the edge detection results to detect all straight lines in the oil surface image to be tested.

[0089] The Hough line transform can detect straight lines in an image and is highly robust to noise and discontinuous edges.

[0090] S37. Select the target line from all lines whose absolute slope value is less than the preset value.

[0091] The preset value can be set according to the actual situation, for example, it can be set to 0.1.

[0092] Since the oil surface is a horizontal straight line, the Hough line transform is performed on the edge detection results to detect all straight lines in the oil surface image to be tested, and the lines with an absolute slope of less than 0.1 are selected.

[0093] S38. Merge similar target lines to remove duplicates.

[0094] S39. Determine the deduplicated target straight line as the candidate oil surface height, and denote the set of all candidate oil surface heights for each oil surface image to be tested as H:

[0095] ;

[0096] in, An index for the candidate oil surface height. The total number of candidate oil surface heights. Indicates the first Candidate oil level height.

[0097] In this embodiment, image algorithms such as binarization, image edge detection, and Hough line transform are used to identify all candidate oil surface heights in the oil surface image to be tested.

[0098] In an alternative implementation, the oil level regression prediction model is trained using samples labeled with historical operating parameters and historical liquid level.

[0099] The specific training process is as follows:

[0100] I. Based on window size A sliding window (i.e., covering the current time and the previous time) (Data from each time point), constructing a high-dimensional input feature vector, specifically including: real-time operating parameters: current inhalation pressure. Exhaust pressure Intake temperature Exhaust temperature Operating power Historical values ​​of operating parameters within the sliding window: Previous Inhalation pressure at any given moment Exhaust pressure Intake temperature Exhaust temperature Operating power Historical oil level: Previous Liquid level at that moment .

[0101] Feature total Dimensions (5-dimensional real-time operating parameters, 5(K-1)-dimensional historical operating parameters, K-1-dimensional historical oil level). Based on the current moment... liquid level height As label values. To ensure the predictive stability of the model in the initial state, N data points are used for preliminary training of the regression model.

[0102] II. Formula (1) constructs a ridge regression model based on time series data, with The feature vector is the input and the current oil level height. Build a regression model for the output labels.

[0103]

[0104] in, Indicates the dependent variable, oil level. Represents the characteristic matrix, This represents the regression coefficient.

[0105] Formula (2) introduces a method to improve the generalization ability of the model into the loss function of the linear model. Regularization. A norm is a measure of the "size" of a vector in a vector space; it is calculated as the square root of the sum of the squares of the vector elements. The loss function used is Mean Squared Error (MSE), and the regularization coefficient is set to... To balance the model's fitting ability and generalization performance.

[0106]

[0107] in, Represents the number of samples.

[0108] Formula (3) minimizes the above objective function Get parameters The optimal estimate.

[0109]

[0110] in, Represents the transpose of a matrix. Represents the identity matrix.

[0111] Formula (4) approximates the error variance of the model using the training set residuals. .

[0112]

[0113] Formula (5) assumes that the prediction error approximately follows a Student distribution (also known as a Gorsett distribution) for the test sample. Its confidence level is The prediction interval is:

[0114]

[0115] in, This indicates a degree of freedom of N-6K and a significance level of 1. quantiles of the student distribution; when When N and K are known, their values ​​can be determined directly by consulting a student distribution table or by using statistical software. This represents the transpose of a matrix.

[0116] To achieve standardized recording and characterization, a lower limit for the oil surface height confidence interval is defined. and upper limit As shown in formulas (6) and (7) respectively.

[0117]

[0118]

[0119] Let the confidence interval be... Predict the center of the interval This is also the predicted value of the regression model, denoted as... .

[0120] In this embodiment, a time-series data model (oil level height regression prediction model) is performed on the time-series data generated by the sensor using machine learning algorithms and confidence intervals based on statistical principles, thereby obtaining the confidence interval for the oil level height.

[0121] Based on this, the multimodal joint decision-making for oil surface height is as follows:

[0122] Confidence intervals output by time series model As prior knowledge, remember The liquid level detection result at that moment is The candidate oil surface heights output by the image algorithm are filtered using the following decision logic:

[0123] Case 1: All candidate oil surface heights are outside the confidence interval: Image detection is deemed faulty at the current time. The oil level height was reused at the moment before The test results, i.e. .

[0124] Scenario 2: Only one candidate oil level height Within the confidence interval: Output the height corresponding to the candidate line. .

[0125] Scenario 3: Multiple candidate oil level heights Within the confidence interval: Calculate the height of each candidate oil level relative to the median of the confidence interval. The distance is calculated, and the candidate oil surface height with the smallest distance from the median of the confidence interval is selected as the final prediction result, as shown in formula (8):

[0126]

[0127] In an optional implementation, the compressor oil level detection method further includes:

[0128] S6. Using real-time operating parameters and current oil level detection results as samples, iteratively optimize the oil level regression prediction model.

[0129] Specifically:

[0130] The final oil level output from the joint decision-making process As the true labels, they are fed back into the oil surface height regression prediction model, which performs incremental learning based on the batch samples generated in real time. Let the size of the new batch be... The newly added batch feature matrix is The new batch label is Record the trained model. , Then the updated covariance matrix is ​​given by formulas (9) and (10):

[0131]

[0132]

[0133] in, The forgetting factor can make oil surface height regression prediction models pay more attention to recent samples. Represents the covariance matrix. This represents the cross-product vector.

[0134] Formula (11) is used to update the model parameters:

[0135] In this embodiment, the oil level regression prediction model is iteratively optimized by using the latest real-time operating parameters and oil level detection results, thereby improving the accuracy of the oil level regression prediction model and thus improving the precision of the oil level prediction results.

[0136] Figure 2 This is a schematic diagram of the specific process of this embodiment. Figure 2 The candidate line is the candidate oil level height in this embodiment, and the time series regression model is the oil level height regression prediction model in this embodiment.

[0137] The following is a specific example illustrating the oil level detection method for the compressor in this embodiment. The specific steps of this example are as follows:

[0138] Step 1: Acquisition and Alignment of Time-Series Data and Video Data

[0139] The sensor collects timing parameters of the compressor during operation, including suction pressure. Exhaust pressure Intake temperature Exhaust temperature Operating power Set the sampling frequency of timing condition parameters Synchronously record absolute timestamps.

[0140] Acquire video data from inside the compressor's viewing cavity and set the video data frame acquisition frequency. That is, 25 frames of images are captured per second.

[0141] Step 2: Time Series Data Regression Modeling

[0142] Step 2.1: Model Pre-training

[0143] The ridge regression model was trained using 2000 manually labeled data points. The 2000 data points contained operating parameters (pressure, temperature, operating power) and their corresponding liquid levels at 2000 time points. To ensure coverage of various compressor operating states, the 2000 data points were uniformly sampled data from 2 hours from startup to shutdown, including various operating states of the compressor from static start-up to high-speed operation to slow-down shutdown.

[0144] Set the sliding window size to K=5, and construct the input feature vector, including the operating parameters (inhalation pressure) at the current time t. Exhaust pressure Intake temperature Exhaust temperature Operating power ), and the eigenvalues ​​of the first four time points. ,…, ,…, , ,…, , ,…, ; ,…, ; Oil level at the first four moments The model's label is the current liquid level height. Manual annotation is used to ensure the accuracy of the initial model.

[0145] In summary, a total of 1996 training data points were obtained for a specific test oil cavity.

[0146] Step 2.2 Dynamic Adjustment of the Model

[0147] Oil level is predicted based on data acquired by sensors and video capture data, and the predicted value is calculated using a regression model. With joint decision correction value The mean squared error (MSE) is used to dynamically update the regression model parameters every 50 frames until all samples have been predicted.

[0148] Step 3: Image Algorithm for Oil Surface Height Detection

[0149] Step 3.1 Video frame dropping: Set the frame dropping frequency That is, a frame is skipped every 25 frames.

[0150] Step 3.2 Image preprocessing: The input image is cropped and rotated to obtain a clear and standardized cross-sectional view of the compressor oil level window.

[0151] Step 3.3 Oil Surface Height Detection: The preprocessed image is binarized, Canny edge extraction is performed, and Hough transform is used to detect straight lines. Only lines with a slope within 0.1 are retained, resulting in several candidate oil surface heights. These are then analyzed using a set... express.

[0152] Step 4: Multimodal Joint Decision-Making for Oil Surface Height

[0153] With a confidence level of 95%, for the oil surface height confidence interval obtained from the oil surface height regression prediction model, the relationship between all candidate oil surface heights and the confidence interval is determined. If no candidate oil surface height falls within the confidence interval, the oil surface height prediction result for the current frame is retained as the oil surface height prediction result for the previous frame. If exactly one candidate oil surface height falls within the confidence interval, that candidate oil surface height is determined as the oil surface height detection result for the current frame. If multiple candidate oil surface heights fall within the confidence interval, the candidate oil surface height closest to the median confidence level is selected as the final detection result.

[0154] Step 5: Output of test results

[0155] Each window corresponds to a detection result of the oil level. The oil level detection results of each window are aligned according to the timestamp. This data can be used to detect the operating status and health of the compressor.

[0156] Example 2

[0157] Figure 3This is a schematic diagram of the structure of an oil level detection device for a compressor provided as an exemplary embodiment of the present disclosure. The oil level detection device for the compressor includes a multi-window oil level data acquisition device and an electronic device provided in Embodiment 3 below.

[0158] The multi-window oil surface data acquisition device includes several cameras 1, camera fixing clips 2, knob structure 3, perforated base plate 4, column 5, light source assembly 6, and light source connecting plate 7.

[0159] Several cameras 1 are fixed to the column 5 by camera mounting clips 2, with the lenses of the cameras 1 parallel to the compressor's viewing window. The camera mounting clips 2 are integrated with the knob structure 3. The column 5 is vertically fixed to the perforated base plate 4. The light source assembly 6 is distributed around the outer periphery of the cameras 1 via light source connecting plates 7.

[0160] The camera mounting clip 2 is used to move the camera 1 up and down along the column 5. The knob structure 3 rotates the camera 1. The light source assembly 6 provides the lighting environment.

[0161] The following is a specific example applied to the oil level detection environment of an air conditioning compressor, which details the multi-window oil level data acquisition device. Figure 4 This is a schematic diagram of the multi-window oil level data acquisition device in this example. Figure 5 This is a physical simulation image of the multi-window oil level data acquisition device in this example, located next to an air conditioner.

[0162] The core components of this multi-window oil surface data acquisition device include a perforated base plate 4, a column 5, a light source connecting plate 7, a matching adapter block 401 and guide rod 402, a camera fixing clip 2, and a matching knob structure. This device has two core functions: firstly, it enables synchronous acquisition of multi-window data, primarily achieved through the array of perforations in the perforated base plate, the column capable of being fixed in multiple positions, and the coordinated use of a ring light source; secondly, it enables two-step adjustment of the imaging pose. First, coarse adjustment is achieved through pre-calibration of the column and base plate assembly positions, followed by fine adjustment through the translation adjustment of the camera fixing clip and the rotation of the knob structure. Ultimately, the device meets the ideal imaging conditions—the camera lens is parallel to the compressor window, and the light source around the lens is uniform and stable.

[0163] The core function of this example multi-window oil level data acquisition device is described as follows:

[0164] The perforated base plate has multiple perforations arranged in an array. These perforations are used to fit screws to achieve detachable fixing of the column. By selecting perforations at different positions, the horizontal fixing position of the column on the base plate can be flexibly adjusted. Furthermore, a single base plate can accommodate the synchronous fixing of multiple columns, thereby realizing multi-window detection function at multiple horizontal positions.

[0165] The column is vertically fixed on the perforated base plate, and multiple cameras can be fixed in layers along the vertical direction of the column to achieve multi-window detection in the vertical space dimension; in addition, the column achieves a stable assembly with the light source through the mechanical connection structure of the adapter block and the guide rod.

[0166] To clearly illustrate the spatial relationships of the various hardware components, a rectangular coordinate system is established with the intersection of the column and the perforated base plate as the origin. The column axis is defined as the z-axis, and the two mutually perpendicular base edges on the perforated base plate are set as the x-axis and y-axis, respectively. The coordinate system is constructed following the right-hand screw rule. By determining the assembly positions of the column and the base plate, the initial matching of the camera lens and the oil cavity window is achieved.

[0167] The light source connection plate adopts a ring-shaped outer frame planar structure, with the light source components distributed around multiple cameras on the same vertical line, providing a uniform and stable lighting environment for the camera imaging process and effectively reducing the impact of ambient light interference on image quality.

[0168] The camera fixing clip can move up and down along the column and be positioned and fixed by the locking device; the knob structure is integrated with the clip, and its core function is to magnetically fix the camera. The knob can also drive the camera to rotate around the x-axis and z-axis, maximizing the freedom of adjustment of the camera's pose, and ultimately achieving high-precision matching between the camera lens and the oil cavity window.

[0169] If it is necessary to simultaneously detect the oil status of the upper cavity of the motor, the lower oil sump, and the reservoir, two columns can be fixed on the base plate of the detection system, and an asymmetrical camera layout scheme can be adopted: one column integrates dual vision acquisition units (cameras), while the other column is only equipped with a single vision acquisition unit (camera) to complete the collaborative detection of multiple areas.

[0170] In addition, the specific hardware installation steps for the multi-window oil surface data acquisition device in this example are as follows:

[0171] Fix the column to the base plate and determine the approximate position of the base plate. Fix multiple cameras to the column using fixing clips. Fix the adapter block to the top of the column and connect the light source. Adjust the position of the base plate and the camera so that the camera's field of view can completely include the target viewing cavity. The camera lens needs to be surrounded inside the ring light source. Based on the above steps, roughly position an imaging device.

[0172] Adjust the knob on the clamp to fine-tune the camera's position, ensuring that the camera lens is parallel to the oil chamber sight glass, and try to ensure that the camera's oil chamber imaging is not tilted or distorted.

[0173] Example 3

[0174] Figure 6This is a schematic diagram of the structure of an electronic device according to an example embodiment of the present disclosure. The electronic device includes a memory, a processor, and a computer program stored in the memory and used to run on the processor. When the processor executes the computer program, it implements the oil level detection method of the compressor in any of the above embodiments. Figure 6 The electronic device 90 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0175] like Figure 6 As shown, the electronic device 90 can be manifested as a general-purpose computing device, such as a server device. The components of the electronic device 90 may include, but are not limited to: at least one processor 91, at least one memory 92, and a bus 93 connecting different device components (including memory 92 and processor 91).

[0176] Bus 93 includes a data bus, an address bus, and a control bus.

[0177] The memory 92 may include volatile memory, such as random access memory (RAM) 921 and / or cache memory 922, and may further include read-only memory (ROM) 923.

[0178] The memory 92 may also include a program tool 925 (or utility) having a set (at least one) program module 924, such program module 924 including but not limited to: operating device, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0179] The processor 91 executes various functional applications and data processing by running computer programs stored in the memory 92, such as the compressor oil level detection method provided in any of the above embodiments.

[0180] Electronic device 90 can also communicate with one or more external devices 94 (e.g., keyboard, pointing device, etc.). This communication can be performed through input / output (I / O) interface 95. Furthermore, electronic device 90 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public network, such as the Internet) via network adapter 96. As shown, network adapter 96 communicates with other modules of electronic device 90 via bus 93. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with electronic device 90, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) devices, tape drives, and data backup storage devices.

[0181] It should be noted that although several units / modules or sub-units / modules of the electronic device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0182] Example 4

[0183] This disclosure also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the oil level detection method for a compressor provided in any of the above embodiments.

[0184] The readable storage medium may be more specifically adopted, including but not limited to: portable disk, hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0185] Example 5

[0186] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the oil level detection method for a compressor as described above.

[0187] The program code for executing the computer program product disclosed herein can be written in any combination of one or more programming languages. The program code can be executed entirely on a user device, partially on a user device, as a stand-alone software package, partially on a user device and partially on a remote device, or entirely on a remote device.

[0188] While specific embodiments of this disclosure have been described above, those skilled in the art should understand that these are merely illustrative examples, and the scope of protection of this disclosure is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principles and essence of this disclosure, but all such changes and modifications fall within the scope of protection of this disclosure.

Claims

1. A method for detecting the oil level in a compressor, characterized in that, The oil level detection method for the compressor includes: The real-time operating parameters of the compressor and the oil level image of the compressor to be measured are obtained; wherein the timestamps corresponding to the real-time operating parameters and the oil level image to be measured are matched. The real-time operating parameters are input into the oil level regression prediction model to output the confidence interval of the oil level of the compressor. Based on the image of the oil level to be measured, all candidate oil level heights of the compressor are determined; The current oil level detection result of the compressor is determined based on the confidence interval and the candidate oil level height.

2. The method for detecting the oil level in a compressor as described in claim 1, characterized in that, The step of determining the current oil level detection result of the compressor based on the confidence interval and the candidate oil level includes: In response to all candidate oil level heights being outside the confidence interval, the oil level height detection result of the previous moment is determined as the current oil level height detection result; or, In response to the fact that only one candidate oil level height is within the confidence interval, the candidate oil level height is determined as the current oil level height detection result; or, In response to several candidate oil level heights falling within the confidence interval, the median value of the confidence interval is determined; Obtain the absolute value of the difference between each candidate oil level height and the median value; The candidate oil level height corresponding to the smallest absolute value is determined as the current oil level height detection result.

3. The method for detecting the oil level in a compressor as described in claim 1, characterized in that, The oil level detection method for the compressor also includes: The oil level in the upper chamber of the compressor motor, the lower oil sump, and the liquid receiver is detected simultaneously by a multi-window oil level data acquisition device. The multi-window oil level data acquisition device includes several camera lenses parallel to the windows of the compressor.

4. The method for detecting the oil level in a compressor as described in claim 1, characterized in that, The step of determining all candidate oil level heights of the compressor based on the image of the oil level to be tested includes: The image of the oil surface to be tested is binarized; The edge detection results of the binarized oil surface image are obtained by using an image edge detection algorithm. The edge detection results are subjected to Hough linear transform to detect all straight lines in the oil surface image to be tested; Select the target line from all lines whose absolute slope value is less than a preset value; The target straight line is determined as the candidate oil level height.

5. The method for detecting the oil level in a compressor as described in claim 1, characterized in that, The oil level height regression prediction model is obtained by training samples labeled with historical operating parameters and historical liquid level heights. And / or, The oil level detection method for the compressor also includes: Using real-time operating parameters and current oil level detection results as samples, the oil level regression prediction model is iteratively optimized.

6. The method for detecting the oil level in a compressor as described in claim 1, characterized in that, The real-time operating parameters include at least one of the following: current intake pressure, current exhaust pressure, current intake temperature, current exhaust temperature, and current operating power.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and for running on the processor, characterized in that, When the processor executes the computer program, it implements the oil level detection method for the compressor according to any one of claims 1 to 6.

8. An oil level detection device for a compressor, characterized in that, The compressor oil level detection device includes a multi-window oil level data acquisition device and the electronic device as described in claim 7; The multi-window oil surface data acquisition device includes several cameras, camera fixing clips, knob structure, perforated base plate, column, light source assembly, and light source connection plate. Several cameras are fixed to the column by camera fixing clips, and the camera lenses are parallel to the viewport of the compressor; the camera fixing clips are integrated with the knob structure; the column is vertically fixed to the perforated base plate; the light source assembly is distributed around the camera periphery by the light source connecting plate; The camera fixing clip is used to move the camera up and down along the column; the knob structure drives the camera to rotate; the light source assembly is used to provide a lighting environment.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the oil level detection method for the compressor according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the oil level detection method for the compressor as described in any one of claims 1-6.

Images