Intelligent control system and method for soft bag package seal integrity leak detector

Abstract

The application discloses an intelligent control system and method of a soft bag package seal integrity leak detector, and belongs to the technical field of automatic control. The intelligent control system comprises an autonomous learning module and a controller. The autonomous learning module generates a reward signal according to a measured vacuum degree of a seal cavity when measuring the seal integrity of a soft bag package to be measured and a vacuum degree of the seal cavity when measuring the seal integrity of a soft bag package of a measurement standard, provided by a vacuum degree sensor. The autonomous learning module generates a state vector at time t according to the reward signal, and generates a control parameter vector of the controller according to the state vector. The controller controls the working state of a vacuum pump according to the control parameter vector. The application has a high intelligent level and can autonomously learn the control parameter vector of the controller.

Description

Technical Field

[0001] This invention relates to an intelligent control system and method for a leak detector for the seal integrity of soft bag packaging, belonging to the field of automatic control technology. Background Technology

[0002] Chinese invention patent application CN119043600A discloses a leak detector and method for the seal integrity of flexible bag packaging. The detector includes a sealing chamber, a pressure sensor, and a first processor. The flexible bag to be tested is placed in the sealing chamber, and the chamber is evacuated using a vacuum pump. Evacuation stops when the vacuum level in the sealing chamber falls below a set value. The pressure sensor detects the pressure generated by the expansion of the flexible bag and converts the pressure value into pressure electrical information. The first processor includes a first airtightness detection module, which outputs the first airtightness of the flexible bag seal based on the pressure electrical information. The first airtightness detection module is trained by a first neural network, which includes a first... The system consists of an input layer and a first two-dimensional neuron intermediate layer. During training, the pressure electrical information of a standard qualified soft bag package under a set vacuum level and the measured airtightness are input into the first input layer. The pressure electrical information of the standard qualified soft bag package and the measured airtightness are learned into the first two-dimensional neuron intermediate layer. During measurement, the pressure information of the soft bag package to be tested under a set vacuum level is input into the first input layer of the first airtightness detection module. The pressure electrical information of the soft bag package to be tested is clustered with the pressure information of the standard qualified soft bag package in the first two-dimensional neuron intermediate layer. The airtightness of the standard qualified soft bag package in the first two-dimensional neuron intermediate layer with the smallest Eulerian distance is corrected by a correction factor and becomes the first airtightness of the soft bag package to be tested.

[0003] However, its control over the vacuum pump is relatively low in terms of intelligence. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent control system and method for a leak detector for the seal integrity of soft bag packaging, which can autonomously learn the control parameters of the vacuum pump controller and has a high degree of intelligence.

[0005] To achieve the aforementioned objective, this invention provides an intelligent control system for a leak detector for the seal integrity of flexible bag packaging. The detector includes a sealed cavity, a vacuum sensor, and a vacuum pump. The vacuum pump evacuates the sealed cavity according to instructions from the intelligent control system. The vacuum sensor measures the vacuum level. The intelligent control system includes a self-learning module and a controller. The self-learning module generates a reward signal based on the measured vacuum level of the sealed cavity when measuring the seal integrity of the flexible bag packaging under test, provided by the vacuum sensor, and the vacuum level of the sealed cavity when measuring the seal integrity of a standard flexible bag packaging. The self-learning module is based on reward signals Generate the state vector at time t According to the state vector Control strategy for generating controller , This is the control parameter vector for the controller; the controller uses the control parameter vector... Control the operating status of the vacuum pump.

[0006] To achieve the aforementioned objective, this invention also provides an intelligent control method for a leak detector for the seal integrity of soft bag packaging, comprising the following steps:

[0007] The vacuum level inside the sealed cavity is measured using a vacuum sensor.

[0008] The self-learning module generates a reward signal based on the measured vacuum level of the sealed cavity when measuring the seal integrity of the soft bag packaging under test, provided by the vacuum sensor, and the vacuum level of the sealed cavity when measuring the seal integrity of the standard soft bag packaging. ;

[0009] Based on the sequence reward signal through the self-learning module Generate the state vector at time t According to the state vector The target control strategy of the generator controller , It is the control parameter vector of the controller;

[0010] The controller uses the control parameter vector Control the operating status of the vacuum pump.

[0011] To achieve the aforementioned objective, the present invention also provides an intelligent control method for a leak detector for the seal integrity of soft bag packaging, comprising:

[0012] The vacuum level inside the sealed cavity is measured using a vacuum sensor.

[0013] The autonomous learning module generates a reward signal based on the measured vacuum level of the sealed cavity when measuring the seal integrity of the soft bag packaging under test, provided by the vacuum sensor, and the vacuum level of the sealed cavity when measuring the seal integrity of the standard soft bag packaging. ;

[0014] Based on the sequence reward signal through the self-learning module Generate the state vector at time t According to the state vector The target control strategy of the generator controller , It is the control parameter vector of the controller;

[0015] The controller uses the control parameter vector Control the operating status of the vacuum pump.

[0016] Compared with existing technologies, the intelligent control system and method of the flexible bag packaging seal integrity leak detector provided by this invention generate reward signals based on the measured vacuum degree of the sealing cavity when measuring the seal integrity of the flexible bag packaging under test, provided by the vacuum degree sensor, and the vacuum degree of the sealing cavity when measuring the seal integrity of the standard flexible bag packaging. ; through the self-learning module based on sequence reward signals Generate the state vector at time t According to the state vector The target control strategy of the generator controller , It is the controller's control parameter vector; the controller uses the control parameter vector... It controls the working status of the vacuum pump, and can autonomously learn the control parameters of the vacuum pump controller, demonstrating a high degree of intelligence. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the components of the soft bag packaging seal integrity leak detector provided in the first embodiment of the present invention.

[0018] Figure 2 This is a block diagram of the intelligent system for detecting leaks in the seal integrity of soft bag packaging provided in the first embodiment of the present invention. Detailed Implementation

[0019] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0020] In this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0021] First Embodiment

[0022] Figure 1 This is a schematic diagram of the composition of the soft bag packaging seal integrity leak detector provided in the first embodiment of the present invention, as shown below. Figure 1 As shown, the detector includes a sealed cavity 1, a vacuum sensor, and a vacuum pump. The vacuum pump pumps air into the sealed cavity according to the instructions of the intelligent control system. The probe of the vacuum sensor is placed inside the sealed cavity to measure the vacuum level inside the sealed cavity.

[0023] In the first embodiment, the soft bag packaging seal integrity leakage detector further includes a processing device, which includes an image processing module, a first sealing determination unit, a force processing module, a second sealing determination unit, and an intelligent control system.

[0024] In the first embodiment, the flexible bag packaging seal integrity leak detector also includes a camera device. When a standard flexible bag is placed in the sealed cavity, the sealed cavity is evacuated by a vacuum pump. When the standard flexible bag expands, the evacuation stops. At this time, the vacuum degree of the sealed cavity is obtained by a vacuum sensor, and then the camera device is used to take a picture of the standard flexible bag, and the image of the standard flexible bag is transmitted to the image processing module. During measurement, the flexible bag to be tested is placed in the sealed cavity, the sealed cavity is evacuated by a vacuum pump, and the vacuum degree of the sealed cavity is measured by a vacuum sensor. The intelligent control system controls the vacuum pump and uses the camera device to take a picture of the flexible bag to be tested, and the image of the flexible bag to be tested is also transmitted to the image processing module.

[0025] The image processing module includes an image data acquisition unit and a feature extraction unit, wherein the image data acquisition unit acquires image data acquired by the camera device.

[0026] The feature extraction unit performs image processing on the image data acquired by the image data acquisition unit to extract the feature quantities of the flexible bag packaging from the image. The feature extraction unit can also perform image processing such as noise removal and edge extraction, determine the outline of the flexible bag packaging in the image, and extract the feature quantities of the outline of the flexible bag packaging from the determined outline portion of the image.

[0027] The first sealing performance determination unit compares the feature values ​​of the outline of the standard soft bag packaging image extracted by the feature extraction unit with the feature values ​​of the outline of the soft bag packaging image to be tested, in order to determine whether the sealing performance of the soft bag packaging to be tested is good or bad.

[0028] The first sealing performance determination unit calculates the similarity between the feature quantity of the outline of the standard soft bag packaging image extracted by the feature extraction unit and the feature quantity of the outline of the soft bag packaging image to be tested. If the similarity is not within the range of the first threshold, the soft bag packaging to be tested is determined to be an unqualified product with poor sealing integrity; conversely, if it is within the range of the first threshold, the soft bag packaging to be tested is determined to be a qualified product with good sealing integrity.

[0029] The display unit can display the result of the judgment of the first sealing judgment unit, including whether the soft bag to be tested is good or bad.

[0030] In the first embodiment, the first sealing performance determination unit can be generated by training a known first machine learning machine. When training the first machine learning machine to become the first sealing performance determination unit, the feature extraction unit inputs the feature values ​​of the contour lines of a standard flexible bag packaging image and the vacuum level of the sealing cavity into the first machine learning machine, using the standard flexible bag packaging sealing performance as teacher data for supervised learning. Based on the results of this supervised learning, the first sealing performance determination unit determines whether the flexible bag packaging to be tested is good or bad, based on the feature values ​​of the contour lines of the flexible bag packaging image to be tested and the vacuum level of the sealing cavity obtained by the feature extraction unit.

[0031] When the first sealing determination unit is configured as a machine learning device for supervised learning, the feature extraction unit can extract the contour length, divide the contour into a set number of parts, and use the tangent angle at the midpoint of each arc segment and the contour angle as features of the contour of the soft bag packaging.

[0032] Optionally, the testing instrument includes a force gauge. When a standard flexible bag is placed in the sealed cavity, the force gauge contacts the upper surface of the standard flexible bag. Then, a vacuum pump evacuates the sealed cavity. When the standard flexible bag inflates, the evacuation stops. At this time, a vacuum sensor obtains the vacuum level of the sealed cavity, and the expansion force of the standard flexible bag measured by the force gauge is transmitted to the force processing module. During measurement, the force gauge contacts the upper surface of the flexible bag to be tested. Then, the vacuum pump evacuates the sealed cavity according to the control of the intelligent control system. The vacuum sensor measures the vacuum level of the sealed cavity, and the expansion force of the flexible bag to be tested measured by the force gauge is transmitted to the force processing module.

[0033] The force processing module includes an expansion force data acquisition unit, which is used to acquire expansion force data generated by the force gauge when the soft bag packaging expands due to the vacuuming of the sealed cavity.

[0034] The second sealing performance assessment unit compares the expansion force of the standard flexible bag packaging extracted by the expansion force data acquisition unit with the expansion force of the flexible bag packaging to be tested to determine the sealing performance of the flexible bag packaging. The second sealing performance assessment unit calculates the difference between the expansion force of the standard flexible bag packaging extracted by the expansion force data acquisition unit and the expansion force of the flexible bag packaging to be tested. If the difference is not within the range of a second threshold, the flexible bag packaging to be tested is judged as a non-conforming product with poor sealing integrity; conversely, if the difference is within the range of the second threshold, the flexible bag packaging to be tested is judged as qualified with good sealing integrity.

[0035] The display unit can display the determination result of whether the soft bag packaging is good or bad based on the determination result of the second sealing determination unit.

[0036] The second sealing performance determination unit can also be generated by training a known second machine learning machine. When training the known second machine learning machine into the second sealing performance determination unit, the expansion force data of a standard flexible bag package and the vacuum degree of the sealing cavity, obtained by the expansion force data acquisition unit, are input into the second machine learning machine. The sealing performance of the standard flexible bag package is used as teacher data for supervised learning. Based on the results of this supervised learning, the second sealing performance determination unit determines whether the flexible bag package under test is good or bad, based on the expansion force and vacuum degree of the flexible bag package under test obtained by the expansion force data acquisition unit.

[0037] Optionally, the present invention utilizes The first and second machine learning machines are trained, where, Indicates splicing, X represents the feature quantity of the standard flexible pouch packaging extracted by the feature extraction unit, where X is the vacuum degree of the sealed cavity. The expansion force of standard flexible pouch packaging acquired by the expansion force data acquisition unit specifically includes:

[0038] S1: Feature quantities of standard soft pouch packaging extracted by the feature extraction unit. The vacuum level X is input to the input layer of the first machine learning machine, and the output layer outputs the airtightness of the standard soft bag packaging. ;

[0039] S2: Based on the airtightness of standard soft bag packaging The loss function was calculated based on the airtightness of the standard soft bag packaging. ;

[0040] S3: Obtain the expansion force of standard soft pouch packaging from the expansion force data acquisition unit. The vacuum level X is input to the input layer of the second machine learning device, and the output layer outputs the airtightness of the standard soft bag packaging. ;

[0041] S4: Sealing according to standard flexible bag packaging The loss function was calculated based on the airtightness of the standard soft bag packaging. ;

[0042] S5: Calculate the airtightness of standard flexible bag packaging according to the following formula. Sealing properties of standard soft bag packaging Consistency regularization loss :

[0043] In the formula, This indicates that the gradient has stopped; Represents the divergence function;

[0044] S6: Calculate the total loss according to the following formula:

[0045] In the formula, , and It's a hyperparameter;

[0046] S7: Determine the total loss If the loss is not minimized, adjust the parameters of the first and second machine learning machines using gradient descent based on the total loss; if the loss is minimized, output the parameters of the first and second machine learning machines.

[0047] Figure 2 This is a block diagram of the intelligent system for detecting leaks in the seal integrity of soft bag packaging provided in the first embodiment of the present invention, as shown below. Figure 2 As shown, the vacuum sensor is used to measure the vacuum level inside the sealed cavity. The intelligent control system includes a self-learning module and a controller. The self-learning module generates a reward signal based on the measured vacuum level of the sealed cavity when measuring the seal integrity of the soft bag packaging under test, provided by the vacuum sensor, and the vacuum level of the sealed cavity when measuring the seal integrity of the standard soft bag packaging. The self-learning module is based on reward signals Generate the state vector at time t According to the state vector Control strategy for generating controller , This is the control parameter vector for the controller; the controller uses the control parameter vector... Control the operating status of the vacuum pump.

[0048] Still Figure 2 As shown, the controller output is When measuring the seal integrity of standard soft bag packaging, the vacuum degree of the sealed cavity obtained by the vacuum sensor is: When measuring the seal integrity of the soft bag packaging to be tested, the vacuum degree of the sealed cavity measured by the vacuum sensor at time t is: The difference between them is the reward signal. Then we have:

[0049]

[0050] Written in matrix form:

[0051]

[0052] In the formula, , These are the three control quantities of the controller at time t;

[0053] ;

[0054] This invention utilizes the advantage function Update the three control variables.

[0055] This invention is illustrated using three control quantity examples. The controller can have N control quantities. When there are N control quantities, , These are N control variables of the controller at time t. This invention utilizes a dominance function. Update N control variables.

[0056] In the first embodiment, It can be written as The functional relationship, that is

[0057] ,

[0058] In the formula, express and The functional relationship; yes The parameters at time t, n=1,2,…,N; are obtained through the dominance function. Update according to the following formula :

[0059] ,

[0060] In the formula, ,

[0061] , These are the state-value functions at time t+1 and time t, respectively. For adjustment coefficients; This is the discount factor; express right gradient; The reward obtained at time t; This is the set of all control parameter vectors.

[0062] In the first embodiment, the autonomous learning module generates the state value function of the state at time t according to the following formula. :

[0063] ,

[0064] In the formula, It is a state-value function; It is a state-value function The parameters at time t are updated according to the following formula. :

[0065] ,

[0066] In the formula, For adjustment coefficients; express right The gradient.

[0067] The intelligent control system of the flexible bag packaging seal integrity leak detector provided in the first embodiment of the present invention measures the vacuum degree inside the sealed cavity through a vacuum sensor; and generates a reward signal through an autonomous learning module based on the measured vacuum degree of the sealed cavity when measuring the seal integrity of the flexible bag packaging under test and the vacuum degree of the sealed cavity when measuring the seal integrity of the standard flexible bag packaging provided by the vacuum sensor. ; through the self-learning module based on sequence reward signals Generate the state vector at time t According to the state vector The target control strategy of the generator controller , It is the controller's control parameter vector; the controller uses the control parameter vector... It controls the working state of the vacuum pump, and can autonomously learn the control parameters of the vacuum pump controller, demonstrating a high degree of intelligence.

[0068] Second Embodiment

[0069] The second embodiment of the present invention only describes the contents that are different from those of the first embodiment; the contents that are the same will not be described again.

[0070] The second embodiment of the present invention also provides an intelligent control method for a leak detector for the seal integrity of soft bag packaging, which includes:

[0071] The vacuum level inside the sealed cavity is measured using a vacuum sensor.

[0072] The self-learning module generates a reward signal based on the measured vacuum level of the sealed cavity when measuring the seal integrity of the soft bag packaging under test, provided by the vacuum sensor, and the real-time vacuum level of the sealed cavity when measuring the seal integrity of the standard soft bag packaging. ;

[0073] Based on the sequence reward signal through the self-learning module Generate the state vector at time t According to the state vector The target control strategy of the generator controller , It is the control parameter vector of the controller;

[0074] The controller uses the control parameter vector Control the operating status of the vacuum pump.

[0075] The intelligent control method of the flexible bag packaging seal integrity leak detector provided in the third embodiment of the present invention measures the vacuum degree in the sealing cavity through a vacuum sensor using a self-learning module; the self-learning module generates a reward signal based on the measured vacuum degree of the sealing cavity when measuring the seal integrity of the flexible bag packaging under test and the vacuum degree of the sealing cavity when measuring the seal integrity of the standard flexible bag packaging, provided by the vacuum sensor. ; through the self-learning module based on sequence reward signals Generate the state vector at time t According to the state vector The target control strategy of the generator controller , It is the controller's control parameter vector; the controller uses the control parameter vector... It controls the working state of the vacuum pump, and can autonomously learn the control parameters of the vacuum pump controller, demonstrating a high degree of intelligence.

[0076] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A smart control system for a soft bag package seal integrity leak detector, wherein, The detector comprises a sealed cavity, a vacuum degree sensor and a vacuum pump, the vacuum pump performs air extraction on the sealed cavity according to the instruction of the intelligent control system; the vacuum degree sensor is used for measuring the vacuum degree in the sealed cavity, and the intelligent control system comprises an autonomous learning module and a controller, the autonomous learning module generates a reward signal according to the measured vacuum degree of the sealed cavity when measuring the sealing integrity of the soft bag package to be measured and the vacuum degree of the sealed cavity when measuring the sealing integrity of the soft bag package according to the measurement standard , the autonomous learning module generates a state vector at time t according to the reward signal , generates a control strategy of the controller according to the state vector , and generates a control parameter vector of the controller , , , the controller controls the working state of the vacuum pump according to the control parameter vector . , are N control quantities of the controller at time t, respectively; the N control quantities are updated by a dominance function ​ ， wherein represents a function relationship with ; is a parameter at time t, n = 1, 2, … N; is updated according to the following formula : , In the formula, ， , Vt+1and Vtare the state-value functions at time t+1and time t, respectively; is an adjustment factor; is a discount factor; denotes the gradient of ; is the set of all control parameter vectors.

2. The intelligent control system of a soft bag package seal integrity leak detector according to claim 1, wherein, The autonomous learning module generates a state value function for a state at time t according to the following equation : , wherein is a state value function; is a parameter at time t; updated according to : , In the formula, is an adjustment coefficient; denotes the gradient of the gradient of 3. The intelligent control system of a soft bag package seal integrity leak detector according to claim 2, wherein, , wherein for measuring the sealing integrity of the soft pouch package under test, the vacuum level sensor measures the vacuum level in the sealing cavity at time t; for measuring the sealing integrity of the standard soft pouch package, the vacuum level sensor acquires the vacuum level in the sealing cavity.

4. A method of intelligent control of a soft bag package seal integrity leak detector, characterized by, comprising the steps of: measuring the vacuum degree in the sealed cavity by a vacuum degree sensor; generating a reward signal by the autonomous learning module according to the measured vacuum degree of the sealing cavity when measuring the sealing integrity of the soft bag package to be measured provided by the vacuum degree sensor and the vacuum degree of the sealing cavity when measuring the sealing integrity of the soft bag package of the measurement standard ; by the autonomous learning module according to the sequence reward signal a state vector at time t , generating a target control policy for the controller according to the state vector , is a control parameter vector of the controller;​ by the controller according to the control parameter vector controlling the operating state of the vacuum pump; , are N control quantities of the controller at time t, respectively; update N control quantities by advantage function ​ ， wherein represents and a function relationship; is a parameter at time t, n = 1, 2,... N; is updated according to : , In the formula, ， , Vt+1and Vtare the state-value functions at time t+1and time t, respectively; is an adjustment factor; is a discount factor; denotes the gradient of with respect to ; and is the set of all control parameter vectors.

5. The intelligent control method of a soft bag package seal integrity leak detector according to claim 4, characterized in that, The autonomous learning module generates a state value function for a state at time t according to the following equation : , wherein is a state value function; is a parameter at time t; updated according to : , wherein is an adjustment factor; denotes the gradient of .

6. The intelligent control method of a soft bag package seal integrity leak detector according to claim 5, wherein, , wherein for measuring the sealing integrity of the soft bag package under test, the vacuum level sensor measures the vacuum level in the sealing chamber at time t; for measuring the sealing integrity of the standard soft bag package, the vacuum level sensor acquires the vacuum level in the sealing chamber.

Images