Anodic oxidation tank monitoring method based on sensing data analysis

Abstract

The invention relates to the technical field of oxidation tank monitoring, and discloses an anodic oxidation tank monitoring method based on sensing data analysis, which comprises the following steps: in a non-working state of an anodic oxidation tank, carrying out damage inspection on the inner surface of the anodic oxidation tank to obtain a damage evaluation value; under the working state of the anodic oxidation tank, sensing data are obtained; processing and analyzing the sensing data to obtain an abnormal evaluation value reflecting the abnormal degree of the working state of the anodic oxidation tank; converting the sensing data into picture data, inputting the picture data into the trained anomaly recognition model, and taking an output result of the anomaly recognition model as an anomaly judgment reference value; the damage evaluation value, the abnormity evaluation value and the abnormity judgment reference value are integrated, the operation risk value of the operation risk of the anodic oxidation tank can be comprehensively and accurately reflected, and therefore automatic monitoring and early warning of the anodic oxidation tank are achieved.

Description

Technical Field

[0001] The present application relates to the technical field of oxidation tank monitoring, and in particular to an anodic oxidation tank monitoring method based on sensor data analysis. Background Art

[0002] Anodizing tank is a key equipment in anodizing process. Its main function is to enhance the hardness, wear resistance and corrosion resistance of metal surface, and also to increase the adhesion and coating performance of metal surface. During the oxidation process, the oxide film produced on the surface of aluminum or aluminum alloy has multiple pores and tiny protrusions. These tiny pores and protrusions are conducive to other treatment processes, such as electrophoretic coating, printing, film lamination, etc.

[0003] In order to ensure the production quality of the anodizing tank, it is necessary to ensure that the oxidation process is carried out under appropriate conditions, so as to obtain an oxide layer with stable quality and excellent performance. At the current stage, the pH value or temperature of the electrolyte is more monitored, and a comprehensive judgment is made in conjunction with the later monitoring of the film performance. For example, the film thickness can be obtained by microscopy or eddy current method, the film hardness can be tested by pencil hardness test or Vickers hardness test, and the corrosion resistance can be tested by salt spray test (such as placing the sample in a salt spray test box to simulate the salt spray environment and observe whether the film has corrosion spots or blistering), acid immersion test (immersing the sample in a specific acid solution for a certain period of time to observe the corrosion of the film), etc., as well as the sealing quality test through fingerprint test, staining spot test, acetic acid-sodium acetate test method, and comprehensive evaluation of the film performance model.

[0004] In addition, the surface inspection of the anodizing tank is also an important part of production safety. It is necessary to check whether there are cracks, corrosion, wear or other damage on the surface of the tank. The usual method is to use visual inspection or tools such as magnifying glasses to conduct a comprehensive inspection of the tank, paying special attention to vulnerable parts such as welds and joints. At the same time, ensure that the inner wall of the tank is clean and free of dirt to ensure the purity and oxidation effect of the electrolyte, and clean it when necessary.

[0005] However, although such multiple tests can accurately evaluate the quality of the product, they require the interruption of the production process and cannot dynamically monitor the operating status of the electrolytic cell while maintaining the operation of the anodizing tank. In addition, the cost is high and needs to be improved. In addition, the method of inspecting the cell body by manual visual inspection or a magnifying glass is subjective and cannot be fully trusted, which also needs to be improved. Summary of the invention

[0006] In order to comprehensively, timely and accurately monitor and evaluate the anodizing tank, the present application provides an anodizing tank monitoring method based on sensor data analysis, which adopts the following technical solutions:

[0007] A method for monitoring an anodizing tank based on sensor data analysis, comprising:

[0008] In a non-working state of the anodizing tank, performing a damage inspection on the inner surface of the anodizing tank to obtain a damage assessment value S;

[0009] Acquire sensor data when the anodizing tank is in operation;

[0010] Processing and analyzing the sensor data to obtain an abnormality evaluation value reflecting the degree of abnormality of the working state of the anodizing tank;

[0011] The sensor data is converted into image data and then input into a trained abnormality recognition model, and the output result of the abnormality recognition model is used as a reference value for abnormality judgment;

[0012] The operation risk value of the anodizing tank is obtained by comprehensively considering the damage assessment value S, the abnormality assessment value and the abnormality judgment reference value, and the state of the anodizing tank is predicted according to the operation risk value;

[0013] Wherein, the abnormality recognition model is a trained neural network model.

[0014] Before the anodizing tank is put into operation, the damage assessment value of the inner surface of the anodizing tank is first calculated and obtained, and then the sensor data of the anodizing tank is analyzed and calculated, and the abnormal judgment reference value obtained by the neural network model is corrected. Finally, the comprehensive damage assessment value is used to obtain the operation risk value that can comprehensively and accurately reflect the operation risk of the anodizing tank, thereby realizing automatic monitoring and early warning of the anodizing tank and preventing problems before they occur.

[0015] Optionally, according to the method for monitoring an anodizing tank based on sensor data analysis according to claim 1, it is characterized in that the process of obtaining the damage assessment value S comprises:

[0016] A sliding guide rail is arranged in the length direction of the anodizing tank, a sliding block is arranged on the sliding guide rail, a telescopic rod elastically abutting against the anodizing tank is arranged on the sliding block, a spherical rolling portion is arranged on one end of the telescopic rod, and a gyroscope is arranged on the other end;

[0017] Presetting a measurement path on the surface to be measured of the anodizing tank;

[0018] The spherical rolling part is brought into contact with one end of the measuring path, and the spherical rolling part is driven to move at a constant speed along the measuring path to the other end of the measuring path;

[0019] Acquiring in real time the acceleration component data generated by the gyroscope in a direction perpendicular to the surface of the anodizing tank during the movement of the spherical rolling portion;

[0020] The damage assessment value S is obtained according to the acceleration component data and the corresponding historical acceleration component data.

[0021] Optionally, the process of obtaining the damage assessment value S according to the acceleration component data and the corresponding historical acceleration component data includes:

[0022]

[0023] Wherein, S is the damage assessment value, Si is the corresponding single damage assessment value obtained in the i-th measurement, Xi is the total length value of the i-th measurement path, Δxi is the acceleration displacement value, Ai(x) represents the relationship between the acceleration component data obtained in the i-th measurement and the corresponding uniform motion distance length x, Ais(x) represents the relationship between the historical acceleration component data corresponding to the i-th measurement and the corresponding uniform motion distance length x, Amai is the maximum value of the acceleration component data in the i-th measurement, and Amsi is the maximum value in the historical acceleration component data corresponding to the i-th measurement. is the average value of the acceleration component data in the i-th measurement, and n is the total number of the measurement paths.

[0024] Optionally, the method for obtaining the abnormal evaluation value includes:

[0025]

[0026] Wherein, F is the abnormal evaluation value, Fk(t) is the abnormal evaluation value of the kth monitoring period at time t, i is the serial number of the major category of the sensor data, j is the serial number of the minor category under the major category, mi is the number of minor categories under the major category, αi is the preset weight coefficient of the ith major category, and D ij is the abnormal estimation score of the corresponding small category item, UD k-1 is the deviation estimate of the k-1th monitoring period; dopt ij (t) is the optimal standard value of the corresponding parameter of the jth sub-category in the ith sub-category, d ij (t) is the measured value of the parameter item corresponding to the jth sub-category in the ith sub-category, dmin ij is the lowest allowable lower limit of the parameter corresponding to the jth sub-category in the ith major category, dmax ij It is the highest allowed upper limit of the parameter corresponding to the jth subcategory within the ith major category.

[0027] Optionally, the deviation estimate UD of the k-1th monitoring period k-1 The methods for obtaining include:

[0028]

[0029] Among them, t a ,t b are the starting point and end point of the k-1th monitoring period, d k-1 min、d k-1 max are the minimum and maximum values ​​measured in the k-1th monitoring period, respectively.

[0030] Optionally, the process of obtaining the operation risk value of the anodizing tank by integrating the damage assessment value S, the abnormality assessment value and the abnormality judgment reference value includes:

[0031]

[0032] Among them, GDF(t) is the operation risk value, and FAI(t) is the abnormality judgment reference value.

[0033] Optionally, the process of predicting the state of the anodizing tank according to the operation risk value includes:

[0034] Comparing the operation risk value with a preset operation assessment interval [GDFA, GDFB];

[0035] If GDF(t)∈[GDFA, GDFB], it is judged that there is no failure risk in the anodizing tank monitoring;

[0036] Otherwise, it is judged that there is a risk of failure in the anodizing tank monitoring.

[0037] Optionally, the sensor data includes electrolyte parameters and electrical parameters and tank liquid circulation state parameters;

[0038] The electrolyte parameters include key ion concentration, pH value, and temperature value of the electrolyte;

[0039] The electrical parameters include pure anodizing voltage and hard anodizing voltage and current density value;

[0040] The tank liquid circulation state parameters include tank liquid circulation flow rate value and pressure value.

[0041] In summary, the present application includes at least one of the following beneficial technical effects:

[0042] Before the anodizing tank is put into operation, the damage assessment value of the inner surface of the anodizing tank is first calculated and obtained, and then the sensor data of the anodizing tank is analyzed and calculated, and the abnormal judgment reference value obtained by the neural network model is corrected. Finally, the comprehensive damage assessment value is used to obtain the operation risk value that can comprehensively and accurately reflect the operation risk of the anodizing tank, thereby realizing automatic monitoring and early warning of the anodizing tank and preventing problems before they occur. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic diagram of the anodizing tank monitoring method of the present invention. DETAILED DESCRIPTION

[0044] Embodiments of the present application are described in detail below, examples of which are shown in the accompanying drawings.

[0045] In the description of this specification, the description with reference to the terms "certain embodiments", "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiments or examples are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0046] The present application embodiment discloses an anodizing tank monitoring method based on sensor data analysis, comprising:

[0047] In a non-working state of the anodizing tank, performing a damage inspection on the inner surface of the anodizing tank to obtain a damage assessment value S;

[0048] Acquire sensor data when the anodizing tank is in operation;

[0049] Processing and analyzing the sensor data to obtain an abnormality evaluation value reflecting the degree of abnormality of the working state of the anodizing tank;

[0050] The sensor data is converted into image data and then input into a trained abnormality recognition model, and the output result of the abnormality recognition model is used as a reference value for abnormality judgment;

[0051] The operation risk value of the anodizing tank is obtained by comprehensively considering the damage assessment value S, the abnormality assessment value and the abnormality judgment reference value, and the state of the anodizing tank is predicted according to the operation risk value;

[0052] Wherein, the abnormality recognition model is a trained neural network model.

[0053] In this embodiment, before the anodizing tank works, the damage assessment value of the inner surface of the anodizing tank is first calculated and obtained, and then the sensor data of the anodizing tank is analyzed and calculated, and the abnormal judgment reference value obtained by the neural network model is corrected. Finally, the comprehensive damage assessment value is obtained to obtain an operation risk value that can comprehensively and accurately reflect the operation risk of the anodizing tank, thereby realizing automatic monitoring and early warning of the anodizing tank and preventing problems before they occur.

[0054] The neural network model can be CNN, which stands for Convolutional Neural Network. It is a deep learning algorithm mainly used in image recognition, image classification and other fields. During the training stage, the change curves of each corresponding data in the sensor data over time can be merged into the same blank picture, and different change curves can be divided into different types by different colors, and then the corresponding values ​​​​can be quantified by the depth of the color. For example, the temperature can be selected as blue, 20 degrees corresponds to 15% dark blue, and 30 degrees corresponds to 20% dark blue.

[0055] Optionally, according to the method for monitoring an anodizing tank based on sensor data analysis according to claim 1, it is characterized in that the process of obtaining the damage assessment value S comprises:

[0056] A sliding guide rail is arranged in the length direction of the anodizing tank, a sliding block is arranged on the sliding guide rail, a telescopic rod elastically abutting against the anodizing tank is arranged on the sliding block, a spherical rolling part is arranged on one end of the telescopic rod, and a gyroscope can be arranged on the other end;

[0057] Presetting a measurement path on the surface to be measured of the anodizing tank;

[0058] The spherical rolling part is brought into contact with one end of the measuring path, and the spherical rolling part is driven to move at a constant speed along the measuring path to the other end of the measuring path;

[0059] Acquiring in real time the acceleration component data generated by the gyroscope in a direction perpendicular to the surface of the anodizing tank during the movement of the spherical rolling portion;

[0060] The damage assessment value S is obtained according to the acceleration component data and the corresponding historical acceleration component data.

[0061] In this embodiment of the present invention, a fixed sliding guide rail can be set at the top of the notch of the anodizing tank, and the sliding block is provided with a driving wheel structure that can drive itself to slide on the sliding guide rail. The driving wheel structure can be driven remotely, and the telescopic rod can be hinged on the sliding block through a spring. The spherical rolling part can use a ball with a diameter of 0.5 cm (similar to the tip of a ballpoint pen), so that when the spherical rolling part moves and contacts with the crack on the measuring path, the acceleration component data can be obtained through the gyroscope, so as to estimate the crack condition of the surface to be measured.

[0062] Optionally, the process of obtaining the damage assessment value S according to the acceleration component data and the corresponding historical acceleration component data includes:

[0063]

[0064] Wherein, S is the damage assessment value, Si is the corresponding single damage assessment value obtained in the i-th measurement, Xi is the total length value of the i-th measurement path, Δxi is the acceleration displacement value, Ai(x) represents the relationship between the acceleration component data obtained in the i-th measurement and the corresponding uniform motion distance length x, Ais(x) represents the relationship between the historical acceleration component data corresponding to the i-th measurement and the corresponding uniform motion distance length x, Amai is the maximum value of the acceleration component data in the i-th measurement, and Amsi is the maximum value in the historical acceleration component data corresponding to the i-th measurement. is the average value of the acceleration component data in the i-th measurement, and n is the total number of the measurement paths.

[0065] In this embodiment of the present invention, the movement speed of the sliding block needs to be accelerated from 0 to the specified speed and then move at a uniform speed, in order to improve the accuracy of damage assessment, and the distance Δxi representing the acceleration or deceleration section is eliminated from the measurement path; in addition, when calculating Ai(x)-Ais(x), the independent variable x in the antecedent equation can be selected, and then the independent variable x in the consequent equation can be increased or decreased to make Ai(x)-Ais(x) the minimum value, which is equivalent to translating the curves corresponding to the two equations on the horizontal axis. If the translation is performed, in the calculation, The Xi-2Δxi in is modified to Xi-2Δxi-2ΔY, where ΔY is the relative translation value. This can eliminate the vibration caused by the movement of the sliding block or the telescopic rod itself as much as possible, thereby achieving a more accurate evaluation effect.

[0066] Optionally, the method for obtaining the abnormal evaluation value includes:

[0067]

[0068] Wherein, F is the abnormal evaluation value, Fk(t) is the abnormal evaluation value of the kth monitoring period at time t, i is the serial number of the major category of the sensor data, j is the serial number of the minor category under the major category, mi is the number of minor categories under the major category, αi is the preset weight coefficient of the ith major category, and D ij is the abnormal estimation score of the corresponding small category item, UD k-1 is the deviation estimate of the k-1th monitoring period; dopt ij (t) is the optimal standard value of the corresponding parameter of the jth sub-category in the ith sub-category, d ij (t) is the measured value of the parameter item corresponding to the jth sub-category in the ith sub-category, dmin ij is the lowest allowable lower limit of the parameter corresponding to the jth sub-category in the ith major category, dmax ij It is the highest allowed upper limit of the parameter corresponding to the jth subcategory within the ith major category.

[0069] In this embodiment of the present invention, the sensor data includes electrolyte parameters and electrical parameters and tank liquid circulation state parameters;

[0070] The electrolyte parameters include the key ion concentration, pH value and temperature value of the electrolyte. Taking the electrolyte parameters as an example, they can be the first major category, and the pH value is the second minor category of the first major category.

[0071] Electrical parameters include pure anodizing voltage and hard anodizing voltage as well as current density values;

[0072] The tank liquid circulation state parameters include the tank liquid circulation flow rate value and pressure value.

[0073] Optionally, the deviation estimate UD of the k-1th monitoring period k-1 The methods for obtaining include:

[0074]

[0075] Among them, t a ,t b are the starting point and end point of the k-1th monitoring period, d k-1 min、d k-1 max are the minimum and maximum values ​​measured in the k-1th monitoring period, respectively.

[0076] Optionally, the process of obtaining the operation risk value of the anodizing tank by integrating the damage assessment value S, the abnormality assessment value and the abnormality judgment reference value includes:

[0077]

[0078] Wherein, GDF(t) is the operation risk value, and FAI(t) is the abnormality judgment reference value. In this embodiment of the present invention, the damage assessment value S is multiplied to affect the operation risk value, which can maximize the impact of mechanical trauma on the safe operation of the anodizing tank.

[0079] Optionally, the process of predicting the state of the anodizing tank according to the operation risk value includes:

[0080] Comparing the operation risk value with a preset operation assessment interval [GDFA, GDFB];

[0081] If GDF(t)∈[GDFA, GDFB], it is judged that there is no failure risk in the anodizing tank monitoring;

[0082] Otherwise, it is judged that there is a risk of failure in the anodizing tank monitoring.

[0083] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for monitoring an anodizing tank based on sensor data analysis, characterized in that: include: In a non-working state of the anodizing tank, performing a damage inspection on the inner surface of the anodizing tank to obtain a damage assessment value S; Acquire sensor data when the anodizing tank is in operation; Processing and analyzing the sensor data to obtain an abnormality evaluation value reflecting the degree of abnormality of the working state of the anodizing tank; The sensor data is converted into image data and then input into a trained abnormality recognition model, and the output result of the abnormality recognition model is used as a reference value for abnormality judgment; The operation risk value of the anodizing tank is obtained by combining the damage assessment value S, the abnormality assessment value and the abnormality judgment reference value, and the state of the anodizing tank is predicted according to the operation risk value; Wherein, the abnormality recognition model is a trained neural network model.

2. The method for monitoring an anodizing tank based on sensor data analysis according to claim 1, characterized in that: The process of obtaining the damage assessment value S includes: A sliding guide rail is arranged in the length direction of the anodizing tank, a sliding block is arranged on the sliding guide rail, a telescopic rod elastically abutting against the anodizing tank is arranged on the sliding block, a spherical rolling portion is arranged on one end of the telescopic rod, and a gyroscope is arranged on the other end; Presetting a measurement path on the surface to be measured of the anodizing tank; The spherical rolling part is brought into contact with one end of the measuring path, and the spherical rolling part is driven to move at a constant speed along the measuring path to the other end of the measuring path; Acquiring in real time the acceleration component data generated by the gyroscope in a direction perpendicular to the surface of the anodizing tank during the movement of the spherical rolling portion; The damage assessment value S is obtained according to the acceleration component data and the corresponding historical acceleration component data.

3. The anodizing tank monitoring method based on sensor data analysis according to claim 2 is characterized in that: The process of obtaining the damage assessment value S according to the acceleration component data and the corresponding historical acceleration component data includes: Wherein, S is the damage assessment value, Si is the corresponding single damage assessment value obtained in the i-th measurement, Xi is the total length value of the i-th measurement path, Δxi is the acceleration displacement value, Ai(x) represents the relationship between the acceleration component data obtained in the i-th measurement and the corresponding uniform motion distance length x, Ais(x) represents the relationship between the historical acceleration component data corresponding to the i-th measurement and the corresponding uniform motion distance length x, Amai is the maximum value of the acceleration component data in the i-th measurement, and Amsi is the maximum value in the historical acceleration component data corresponding to the i-th measurement. is the average value of the acceleration component data in the i-th measurement, and n is the total number of the measurement paths.

4. The method for monitoring an anodizing tank based on sensor data analysis according to claim 1, characterized in that: The method for obtaining the abnormal evaluation value includes: Wherein, F is the abnormal evaluation value, Fk(t) is the abnormal evaluation value of the kth monitoring period at time t, i is the serial number of the major category of the sensor data, j is the serial number of the minor category under the major category, mi is the number of minor categories under the major category, αi is the preset weight coefficient of the ith major category, and D ij is the abnormal estimation score of the corresponding small category item, UD k-1 is the deviation estimate of the k-1th monitoring period; dopt ij (t) is the optimal standard value of the corresponding parameter of the jth sub-category in the ith sub-category, d ij (t) is the measured value of the parameter item corresponding to the jth sub-category in the ith sub-category, dmin ij is the lowest allowable lower limit of the parameter corresponding to the jth sub-category in the ith major category, dmax ij It is the highest allowed upper limit of the parameter corresponding to the jth subcategory within the ith major category.

5. The method for monitoring an anodizing tank based on sensor data analysis according to claim 4, characterized in that: The deviation estimate UD of the k-1th monitoring cycle k-1 The methods for obtaining include: Among them, t a ,t b are the starting point and end point of the k-1th monitoring period, d k-1 min、d k-1 max are the minimum and maximum values ​​measured in the k-1th monitoring period, respectively.

6. The method for monitoring an anodizing tank based on sensor data analysis according to claim 4, characterized in that: The process of obtaining the operation risk value of the anodizing tank by integrating the damage assessment value S, the abnormality assessment value and the abnormality judgment reference value comprises: Among them, GDF(t) is the operation risk value, and FAI(t) is the abnormality judgment reference value.

7. The method for monitoring an anodizing tank based on sensor data analysis according to claim 6, characterized in that: The process of predicting the state of the anodizing tank according to the operation risk value includes: Comparing the operation risk value with a preset operation assessment interval [GDFA, GDFB]; If GDF(t)∈[GDFA, GDFB], it is judged that there is no failure risk in the anodizing tank monitoring; Otherwise, it is judged that there is a risk of failure in the anodizing tank monitoring.

8. The method for monitoring an anodizing tank based on sensor data analysis according to claim 1, characterized in that: The sensor data include electrolyte parameters and electrical parameters and tank liquid circulation state parameters; The electrolyte parameters include key ion concentration, pH value, and temperature value of the electrolyte; The electrical parameters include pure anodizing voltage and hard anodizing voltage and current density value; The tank liquid circulation state parameters include tank liquid circulation flow rate value and pressure value.

Images