Data processing and communication method of heat seal detection system

Abstract

The invention relates to the technical field of industrial detection and Internet of Things communication, and discloses a data processing and communication method of a heat seal detection system, which comprises the following steps: synchronously triggering temperature acquisition through a production line beat signal, eliminating environmental noise interference by adopting a dynamic filtering algorithm, automatically selecting a proper algorithm based on a mathematical model, and calculating model characteristics; the temperature change is predicted through the model, so that accurate judgment is obtained; threshold remote configuration and feature data uploading are realized by using a network protocol; the system comprises a temperature acquisition module, an edge computing unit and a cloud management platform. The problems that an existing detection system is high in data transmission delay and inflexible in parameter configuration are solved, millisecond-level abnormal response and production line cooperative control is achieved, and the system is suitable for intelligent quality monitoring of food and medicine packaging.

Description

Technical Field

[0001] The present invention relates to the technical fields of industrial inspection and Internet of Things communication, and particularly to a data processing and communication method for a heat sealing detection system. Background Art

[0002] Heat sealing detection is a detection method used to evaluate the sealing performance of heat-sealing materials (such as plastic bags, packaging materials, etc.). Through this detection, it can be judged whether the sealing of the heat-sealed part is good, so as to ensure the quality and safety of products, and observe whether there are obvious defects in the heat-sealed part, such as air bubbles, wrinkles, fractures, etc.

[0003] The existing technologies for heat sealing detection systems have the following defects: Unidirectional data transmission: Traditional detection devices only support local alarms and cannot achieve remote diagnosis and parameter optimization; Rigid algorithms: The thresholds are fixed and lack self-learning ability, and cannot adapt to changes in the thermal conductivity of materials; Complex communication: Traditional wired networks require a huge amount of work to re-lay wires on site. Summary of the Invention

[0004] Based on the technical problems existing in the background art, the present invention proposes a data processing and communication method for a heat sealing detection system.

[0005] The data processing and communication method for a heat sealing detection system proposed by the present invention includes the following steps: Synchronous triggering: Receive the beat pulse signal sent by the production line PLC, and wake up the infrared temperature measurement module according to the rising edge of the pulse;

[0006] Temperature acquisition and filtering: Continuously acquire the temperature analog signal of the sealing area at a sampling rate greater than 100Hz; Use the sliding average filtering algorithm to discard a specific number of data at the beginning and end, and take the effective data in the middle to calculate the characteristic temperature value;

[0007] Abnormality determination: Compare the characteristic temperature value with the comprehensive result of the predicted temperature, slip correction, and threshold. If the difference between any predicted value and the current value exceeds the threshold, trigger a first-level alarm; Synchronously calculate the standard deviation of the current temperature distribution. If the standard deviation is greater than 15% of the historical data average, trigger a second-level alarm; If the difference between the current average value and the historical average value is too large, trigger a second-level alarm;

[0008] Data communication: Compress and package the alarm event and the characteristic temperature value through the network protocol, and upload them to the cloud management platform; Receive the dynamic threshold configuration instruction sent by the cloud and update the local storage parameters.

[0009] In step S2, the moving average filtering algorithm is specifically as follows: Given an original data sequence T1, T2... Tn, the first m and the last m data are removed, and the arithmetic mean of the remaining data is calculated, where m = floor(n × 10%); In step S4, differential coding technology is used for data compression, and only the difference between the current characteristic temperature and the temperature of the previous cycle is transmitted.

[0010] During the network protocol communication process, the cloud management platform assigns an independent Topic to each detection device, and the format is: / {messag_Class} / {deviceID};

[0011] The alarm event data packet contains fields such as timestamp, device ID, temperature value, and alarm level, and the data packet structure is in JSON format.

[0012] Temperature acquisition module: It includes an infrared sensor array and a signal conditioning circuit, and is rigidly connected to the production line robotic arm;

[0013] Edge computing unit: Integrates a microcontroller and a 4G / WiFi dual-mode communication chip, deploys filtering algorithms, automatic algorithm selection, algorithm fitting, and anomaly determination programs;

[0014] Cloud management platform: Provides a visual threshold configuration interface and generates a temperature distribution statistical report based on historical data.

[0015] The edge computing unit is connected to the production line PLC through an RS485 interface or an IO port to receive the beat pulse signal and device status information;

[0016] The cloud management platform sets up multi-level permission management. The production line operator can only view the alarm records, and the engineer can modify the threshold parameters.

[0017] The dynamic filtering algorithm includes:

[0018] Mean filtering based on a sliding window; Weighted moving average filtering; Kalman filtering;

[0019] Wavelet threshold denoising;

[0020] Among them, when the temperature change rate > 10°C / ms, the Kalman filtering mode is enabled, and the historical state prediction value and the real-time measurement value are fused for recursive calculation. The prediction model equation is:

[0021]

[0022] Where State estimate value (after prediction or update), K: Weight coefficient, balancing prediction and measurement, z: Actual sensor data, H: Mapping matrix from state to measurement, x^k∣k: Posterior state estimate at the current time k (optimal estimate combining prediction and observation), The prior state prediction at the current moment k, zk: the actual observed value at moment k, Kk: the Kalman Gain, which determines the weight distribution of prediction and observation. The innovation, which represents the difference between the actual observation and the predicted observation.

[0023] The beneficial effects of the present invention are as follows:

[0024] Improved real-time performance: The latency from data collection to cloud storage is < 200 ms; False alarm rate reduction: The misjudgment rate is reduced from 12% to below 3% through dynamic filtering; Reduction in operation and maintenance costs: Remote configuration reduces on-site debugging man-hours by approximately 70%. Description of the Drawings

[0025] Figure 1 It is a diagram of the network packet fields (downlink) of the data processing and communication method of the heat-sealing detection system proposed by the present invention;

[0026] Figure 2 It is a diagram of the network packet fields (uplink) of the data processing and communication method of the heat-sealing detection system proposed by the present invention;

[0027] Figure 3 It is a hardware system architecture diagram of the data processing and communication method of the heat-sealing detection system proposed by the present invention;

[0028] Figure 4 It is a software flowchart of the data processing and communication method of the heat-sealing detection system proposed by the present invention;

[0029] Figure 5 It is a temperature curve diagram of the data processing and communication method of the heat-sealing detection system proposed by the present invention;

[0030] Figure 6 It is a comparison diagram of the temperature before and after correction of the data processing and communication method of the heat-sealing detection system proposed by the present invention. Detailed Embodiment

[0031] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments.

[0032] The data processing and communication method of the heat-sealing detection system includes the following steps: Synchronous triggering: Receive the beat pulse signal sent by the production line PLC, and wake up the infrared temperature measurement module according to the rising edge of the pulse;

[0033] Temperature acquisition and filtering: Continuously acquire the temperature analog signal in the sealing area at a sampling rate greater than 100Hz; adopt the moving average filtering algorithm, discard a specific number of data at the beginning and end, and calculate the characteristic temperature value from the valid data in the middle;

[0034] Abnormality determination: Compare the characteristic temperature value with the comprehensive result of the predicted temperature, slip correction, and threshold. If the difference between any predicted value and the current value exceeds the threshold, a first-level alarm is triggered; synchronously calculate the standard deviation of the current temperature distribution. If the standard deviation is greater than 15% of the average value of historical data, a second-level alarm is triggered; if the difference between the current average value and the historical average value is too large, a second-level alarm is triggered;

[0035] Data communication: Compress and package the alarm event and characteristic temperature value through the network protocol and upload them to the cloud management platform; receive the dynamic threshold configuration instruction sent by the cloud and update the local storage parameters;

[0036] Synchronous trigger mechanism: Use the PLC pulse signal to accurately control the data acquisition timing, with an error ≤ 30ms from the production line beat.

[0037] The moving average filtering algorithm in step S2 is specifically as follows: Let the original data sequence be T1, T2... Tn, remove the first m and the last m data, and calculate the arithmetic average of the remaining data, where m = floor(n×10%); in step S4, differential coding technology is used for data compression, and only the difference between the current characteristic temperature and the temperature in the previous cycle is transmitted;

[0038] Dynamic filtering algorithm: For the characteristics of rapid temperature rise and fall during the heat sealing process, such as Figure 3 Temperature curve, such as Figure 5 Temperature curve, design a segmented filtering strategy: When the temperature change rate > 10℃ / ms, enable the emergency mode and only retain the peak temperature data; under normal conditions, use the moving average filtering to eliminate environmental interference.

[0039] During the network protocol communication process, the cloud management platform assigns an independent Topic to each detection device, and the format is: / {messag_Class} / {deviceID};

[0040] The alarm event data packet contains fields such as timestamp, device ID, temperature value, and alarm level, and the data packet structure is in JSON format;

[0041] Intelligent communication protocol: Define the network message format network message fields (downlink) for a dedicated network, supporting resume of data transmission after network interruption and data reissuance.

[0042] Temperature acquisition module: Includes an infrared sensor array and a signal conditioning circuit, and is rigidly connected to the production line robotic arm;

[0043] Edge computing unit: Integrates a microcontroller and a 4G / WiFi dual-mode communication chip, deploys filtering algorithms, automatic algorithm selection, algorithm fitting, and anomaly determination programs;

[0044] Cloud management platform: Provides a visual threshold configuration interface and generates a temperature distribution statistical report based on historical data.

[0045] The edge computing unit is connected to the production line PLC through the RS485 interface or IO port to receive beat pulse signals and device status information;

[0046] The cloud management platform sets up multi-level permission management. Production line operators can only view alarm records, and engineers can modify threshold parameters.

[0047] The dynamic filtering algorithms include:

[0048] Mean filtering based on a sliding window; Weighted moving average filtering; Kalman filtering;

[0049] Wavelet threshold denoising;

[0050] Among them, when the temperature change rate > 10°C / ms, the Kalman filtering mode is enabled, and the historical state prediction value and the real-time measurement value are fused for recursive calculation. The prediction model equation is:

[0051]

[0052] Where State estimate value (predicted or updated), K: Weight coefficient, balancing prediction and measurement, z: Actual sensor data, H: Mapping matrix from state to measurement, x^k∣k: Posterior state estimate at the current time k (optimal estimate combined with prediction and observation), Prior state prediction at the current time k, zk: Actual observation value at time k, Kk: Kalman Gain, determining the weight allocation of prediction and observation, Observation residual (Innovation), representing the difference between the actual observation and the predicted observation.

[0053] In the present invention, through hardware deployment: an infrared temperature measurement module is installed 100 mm behind the heat sealing station and fixed by a bracket to avoid the influence of mechanical vibration. The edge computing unit is integrated into the equipment control cabinet and connected to the PLC through shielded twisted pair wires; software configuration: set the initial threshold on the cloud management platform, the first-level alarm threshold: the difference between the actual value and the predicted value exceeds 5 degrees Celsius, the standard deviation warning value: 1.3 times the average standard deviation of historical data, and the second-level alarm is activated. After the adaptive mode is activated, the platform automatically analyzes the data every week and recommends a threshold optimization plan. The average value warning threshold: the difference between the average value and the historical average value is too large, and the second-level alarm is activated; abnormal handling process. When the first-level alarm is triggered, the edge unit immediately sends an emergency stop instruction to the PLC and prompts through an audible and visual alarm. After the second-level alarm is triggered, the system automatically saves the temperature data of the last 10 minutes for offline analysis.

[0054] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. A data processing and communication method for a heat seal detection system, characterized in that: The following steps are involved: S1: Synchronous trigger: Receive the beat pulse signal sent by the production line PLC and wake up the infrared temperature measurement module according to the rising edge of the pulse; S2: Temperature collection and filtering: Continuously collect the temperature analog signal of the sealing area at a sampling rate greater than 100 Hz; Adopt the sliding average filtering algorithm, discard a certain amount of data at the beginning and end, and take the middle valid data to calculate the characteristic temperature value; S3: Abnormal determination: Compare the characteristic temperature value with the predicted temperature, slip correction and threshold comprehensive results. If the difference between the predicted value and the current value at any point exceeds the threshold, a first-level alarm is triggered. The standard deviation of the current temperature distribution is calculated synchronously. If the standard deviation is greater than 15% of the average value of historical data, a secondary alarm is triggered; The difference between the current average value and the historical average value is too large, triggering a secondary alarm; S4: Data Communication: Compress and package alarm events and characteristic temperature values ​​through network protocols and upload them to the cloud management platform; Receive dynamic threshold configuration instructions sent by the cloud and update local storage parameters.

2. The data processing and communication method of the heat seal detection system according to claim 1, characterized in that: The sliding average filtering algorithm in step S2 is specifically: Assume that the original data sequence is T1, T2...Tn, remove the first m and last m data, and calculate the arithmetic mean of the remaining data, where m = floor (n × 10%); The data compression in step S4 adopts the difference encoding technology, and only transmits the difference between the current characteristic temperature and the temperature of the previous cycle.

3. The data processing and communication method of the heat seal detection system according to claim 1, characterized in that: During the network protocol communication process, the cloud management platform allocates an independent Topic to each detection device in the following format: / {messag_Class} / {deviceID}; The alarm event data packet contains timestamp, device ID, temperature value, and alarm level fields, and the data packet structure is in JSON format.

4. The data processing and communication method of the heat seal detection system according to claims 1-3, characterized in that: include: Temperature acquisition module: It contains an infrared sensor array and signal conditioning circuit, which is rigidly connected to the production line robot arm; Edge computing unit: Integrate microcontroller and 4G / WiFi dual-mode communication chip, deploy filtering algorithm, automatic algorithm selection, algorithm fitting and abnormality judgment program; Cloud management platform: Provides a visual threshold configuration interface and generates temperature distribution statistics reports based on historical data.

5. The data processing and communication method of the heat seal detection system according to claim 1, characterized in that: The edge computing unit is connected to the production line PLC via an RS485 interface or an IO port to receive a beat pulse signal and equipment status information; The cloud management platform is equipped with multi-level authority management, where production line operators can only view alarm records, and engineers can modify threshold parameters.

6. The method according to claim 4, characterized in that: The dynamic filtering algorithm includes: Mean filtering based on sliding window; Weighted moving average filter; Kalman filter; Wavelet threshold denoising; Among them, when the temperature change rate is greater than 10℃ / ms, the Kalman filter mode is enabled, and the historical state prediction value and the real-time measurement value are integrated for recursive calculation. The prediction model equation is: in State estimate (prediction or update), K: weight coefficient, balancing prediction and measurement, z: actual sensor data, H: mapping matrix from state to measurement, x^k|k: posterior state estimate at the current time k (combining the optimal estimate after prediction and observation), The prior state prediction at the current time k, z k : The actual observation value at time k, Kk: Kalman gain, which determines the weight distribution between prediction and observation, Observation residual (Innovation), which represents the difference between the actual observation and the predicted observation.

Images