Calendering gradient control system

Abstract

The invention provides a calendering gradient control system, and relates to the technical field of photovoltaic glass production, the system collects temperature data of a monitoring area in real time through a special infrared thermal imager deployed in an industrial site, the thermal imager continuously obtains temperature field information at a frame frequency greater than or equal to 30Hz, each monitoring point can be provided with a plurality of POIs, and the POIs are connected with the monitoring point. Meanwhile, characteristic parameters such as the highest temperature, the average temperature and the temperature gradient in the region are calculated; collected temperature data are transmitted to a local control device through an industrial router, and on the edge side, the system executes the following core processing: real-time calculation of regional temperature distribution characteristics, execution of an adaptive dynamic threshold algorithm, and preliminary alarm judgment and data classification. The system has the advantages of high integration level and the like.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic glass production technology, and in particular to a rolling gradient control system. Background Technology

[0002] Temperature is a crucial parameter for production safety and equipment monitoring in industries such as industrial manufacturing and energy. Currently, temperature monitoring systems on the market generally suffer from technical deficiencies such as low system integration, poor environmental adaptability, simplistic alarm mechanisms, low data processing efficiency, and a lack of dedicated hardware optimization. Specifically, hardware and upper-level software systems often come from different vendors, with inconsistent interfaces hindering efficient collaboration; fixed alarm thresholds fail to adapt to the dynamic changes in complex industrial environments, easily leading to false alarms or missed alarms; and fragmented data storage schemes result in weak real-time data synchronization capabilities between local and remote cloud environments, hindering subsequent big data analysis. Therefore, there is an urgent need in this field for a highly integrated, intelligent, adaptive, stable, and reliable integrated hardware and software temperature measurement solution. Summary of the Invention

[0003] This invention aims to provide a highly integrated, intelligent, adaptive, stable, and reliable intelligent industrial temperature monitoring system and method through deep hardware and software co-design. It addresses the problems of low system integration, simplistic alarm mechanisms, inadequate data processing, and lack of dedicated hardware optimization in existing technologies. By achieving high-precision temperature detection, enhancing the system's adaptability and intelligence, and providing powerful data analysis capabilities, it offers strong support for predictive maintenance and process optimization.

[0004] The system consists of a hardware layer, a software layer, and a data layer, forming an organic whole.

[0005] Hardware layer:

[0006] Dedicated infrared thermal imager: with a resolution of ≥640*480 pixels, temperature measurement accuracy of ±2°C, frame rate of ≥30Hz, and supports multiple POI (point of interest) temperature measurement and regional temperature statistics (maximum temperature, average temperature, temperature gradient).

[0007] Industrial routers: IP67 protection rating, operating temperature -40 to 75°C, network self-healing time <20ms, and support for multiple industrial protocols such as Modbus TCP / IP and PROFINET.

[0008] Local control equipment: It adopts an embedded industrial computer, equipped with a high-performance processor, memory and solid-state drive, and is responsible for real-time data processing and edge computing.

[0009] Remote server cluster: Includes high-frequency CPU computing nodes, time-series databases, relational databases, and object storage, used for storing and deeply analyzing massive amounts of data.

[0010] Centralized control platform terminal: adopts a high-resolution industrial-grade display system for visual monitoring and operation.

[0011] Software layer:

[0012] Edge-side software: runs on the local control device and includes a temperature monitoring engine, a data management module, and a communication interface module.

[0013] Cloud-based software: It adopts a cloud-native microservice architecture based on Docker and Kubernetes, and uses technologies such as Java Spring Cloud to provide data reception, storage, analysis and API services.

[0014] Application layer software: Adopting a web front-end and back-end separation architecture, it provides real-time monitoring, alarm management, historical data analysis and system configuration functions.

[0015] The method is executed by the system and includes the following steps:

[0016] 1. Data Acquisition and Feature Calculation: Temperature data of the monitored area is acquired using a dedicated infrared thermal imager; the temperature distribution characteristics of the area are calculated in real time, including the average temperature of the area, the highest / lowest temperature points, and the temperature gradient.

[0017] 2. Adaptive Threshold Calculation and Alarm Judgment: A dynamic threshold algorithm is adopted. The alarm threshold ΔT Dynamic is dynamically adjusted based on the temperature distribution characteristics of the current area and historical data (e.g., Dynamic threshold = base threshold + k1 * current temperature gradient + k2 * historical fluctuation standard deviation (where k1 and k2 are adjustable weight coefficients)). The alarm condition is: when the difference between the temperature TPOI of any monitoring point (POI) and the average temperature TRegion_Avg of the region exceeds the dynamic threshold ΔT Dynamic, that is, when TPOI - TRegion_Avg > ΔT Dynamic, an alarm is triggered.

[0018] 3. Data classification and storage: The system classifies data into regular sampling data and alarm data; a differentiated storage strategy is adopted, with alarm data (including thermal images) being synchronized to the remote server first.

[0019] 4. Multi-dimensional data analysis and display: The centralized control platform provides multi-dimensional data analysis functions based on timelines, such as alarm image sequence playback, temperature change trend charts, and regional temperature distribution heat maps.

[0020] Working principle

[0021] The system uses dedicated infrared thermal imagers deployed at industrial sites to collect temperature data of the monitored area in real time. The thermal imagers continuously acquire temperature field information at a frame rate of ≥30Hz, and multiple Points of Interest (POIs) can be set at each monitoring point. Simultaneously, it calculates characteristic parameters such as the highest temperature, average temperature, and temperature gradient within the area.

[0022] The collected temperature data is transmitted to the local control device (embedded industrial computer) via an industrial router. At the edge, the system performs the following core processes: real-time calculation of regional temperature distribution characteristics; execution of an adaptive dynamic threshold algorithm; and preliminary alarm judgment and data classification.

[0023] The system employs an intelligent data classification strategy. Regular data is collected at a preset frequency and transmitted using compressed data; alarm data is transmitted in real-time with priority, containing complete data packets. The reliability of data transmission is ensured through the breakpoint resume function of the industrial router.

[0024] The remote server cluster receives and stores various types of data, providing the following core functions: real-time data visualization; historical data backtracking analysis; multi-dimensional data comparison; and intelligent report generation.

[0025] In traditional temperature monitoring, we typically focus on whether the temperature at a certain point exceeds a threshold (e.g., the temperature of a calender roll exceeds 500°C). However, the temperature gradient in this invention describes the rate and direction of temperature change in space within a specific region. It is a vector that contains both magnitude and direction information.

[0026] Temperature gradient itself is a diagnostic indicator. When the temperature gradient is small, it means the temperature across the entire area is uniform, the equipment is operating smoothly, and it is in a "quiet" state. In this case, the system considers the operating condition to be good, and any minor temperature anomaly warrants attention. Therefore, the alarm device will lower the dynamic threshold (ΔT Dynamic) to make the system more sensitive, enabling it to capture those minute anomaly signals that stand out against a stable background, thus effectively reducing the risk of missed alarms. When the temperature gradient is large, it means there is significant uneven heating or cooling within the area, and the equipment itself may be in a state of heating up, cooling down, or exhibiting inherent thermal inhomogeneity. In this context, the normal fluctuation range of individual measuring points may also be large. In this case, the alarm device will raise the dynamic threshold to make the system more lenient, avoiding misinterpreting normal operating condition fluctuations as fault alarms, thus effectively reducing the false alarm rate.

[0027] Compared with existing technologies, this solution has the following technical effects.

[0028] 1. The deep integration of hardware and software avoids compatibility issues between devices from different vendors, achieves optimal utilization of hardware resources and efficient execution of instructions, and improves the overall system response speed and stability.

[0029] 2. Adaptive area monitoring and dynamic threshold algorithm effectively avoid false alarms caused by normal fluctuations in ambient temperature, and sensitively capture slow temperature rise anomalies that may be ignored by fixed thresholds, making the system "intelligent" and adaptable to complex and ever-changing industrial environments.

[0030] 3. Based on POI and regional characteristics, differentiated alarms are implemented to achieve differentiated monitoring of areas with different importance and thermal characteristics, making alarms more targeted, enhancing early warning value, and improving the precision and professionalism of monitoring.

[0031] 4. Intelligent data classification and differentiated storage synchronization strategy: Data is divided into regular sampling data and alarm data, and a differentiated synchronization strategy is adopted: alarm data is synchronized in real time with priority. Industrial routers are used to realize breakpoint resume and encrypted compression transmission, which reduces the pressure on network bandwidth and server storage.

[0032] 5. Adopting industrial-grade hardware architecture selection and design ensures long-term stable operation of the system under high-temperature interference environment, extends equipment life and reduces maintenance costs. Attached Figure Description

[0033] Figure 1 Image of the monitoring terminal screen;

[0034] Figure 2 This is a schematic diagram of the overall architecture of the system;

[0035] Figure 3 This is a flowchart of the system's workflow;

[0036] Figure 4 This is a schematic diagram illustrating the principle of the adaptive dynamic threshold algorithm.

[0037] Figure 5 This is a schematic diagram illustrating how the system is deployed in a photovoltaic rolled glass production line as shown in the embodiment. Detailed Implementation

[0038] The following are specific embodiments of the present invention, which are described in conjunction with the accompanying drawings. However, the present invention is not limited to these embodiments.

[0039] I. An intelligent industrial temperature detection system

[0040] The system consists of a hardware layer, a software layer, and a data layer, forming an organic whole. Its architecture is as follows: Figure 2 As shown.

[0041] 1. Hardware layer:

[0042] Dedicated infrared thermal imager: with a resolution of ≥640*480 pixels, temperature measurement accuracy of ±2°C, frame rate of ≥30Hz, and supports multiple POI (point of interest) temperature measurement and regional temperature statistics (maximum temperature, average temperature, temperature gradient).

[0043] Industrial routers: IP67 protection rating, operating temperature -40 to 75°C, network self-healing time <20ms, and support for multiple industrial protocols such as Modbus TCP / IP and PROFINET.

[0044] Local control equipment: It adopts an embedded industrial computer, equipped with a high-performance processor, memory and solid-state drive, and is responsible for real-time data processing and edge computing.

[0045] Remote server cluster: Includes high-frequency CPU computing nodes, time-series databases, relational databases, and object storage, used for storing and deeply analyzing massive amounts of data.

[0046] Centralized control platform terminal: adopts a high-resolution industrial-grade display system for visual monitoring and operation.

[0047] 2. Software layer:

[0048] Edge-side software: runs on the local control device and includes a temperature monitoring engine, a data management module, and a communication interface module.

[0049] Cloud-based software: It adopts a cloud-native microservice architecture based on Docker and Kubernetes, and uses technologies such as Java Spring Cloud to provide data reception, storage, analysis and API services.

[0050] Application layer software: Adopting a web front-end and back-end separation architecture, it provides real-time monitoring, alarm management, historical data analysis and system configuration functions.

[0051] II. An Intelligent Industrial Temperature Detection Method

[0052] This method is executed by the system, such as Figure 3 As shown, it includes the following steps:

[0053] 1. Data Acquisition and Feature Calculation: Temperature data of the monitored area is acquired using a dedicated infrared thermal imager; the temperature distribution characteristics of the area are calculated in real time, including the average temperature of the area, the highest / lowest temperature points, and the temperature gradient.

[0054] 2. Adaptive threshold calculation and alarm judgment: A dynamic threshold algorithm is adopted, with the alarm threshold ΔT Dy namicThe dynamic threshold is dynamically adjusted based on the current temperature distribution characteristics and historical data of the area (e.g., dynamic threshold = base threshold + k1 * current temperature gradient + k2 * historical fluctuation standard deviation (where k1 and k2 are adjustable weighting coefficients)). The alarm condition is: when the temperature T at any monitoring point (POI) reaches a certain value... POI With the regional average temperature T Region_Avg The difference exceeds the dynamic threshold ΔT Dynamic When, that is, when T is satisfied POI -T Region_Avg >ΔT Dynamic An alarm is triggered at that time.

[0055] 3. Data classification and storage: The system classifies data into regular sampling data and alarm data; a differentiated storage strategy is adopted, with alarm data (including thermal images) being synchronized to the remote server first.

[0056] 4. Multi-dimensional data analysis and display: The centralized control platform provides multi-dimensional data analysis functions based on timelines, such as alarm image sequence playback, temperature change trend charts, and regional temperature distribution heat maps.

[0057] Working principle

[0058] The system uses dedicated infrared thermal imagers deployed at industrial sites to collect temperature data of the monitored area in real time. The thermal imagers continuously acquire temperature field information at a frame rate of ≥30Hz, and multiple Points of Interest (POIs) can be set at each monitoring point. Simultaneously, it calculates characteristic parameters such as the highest temperature, average temperature, and temperature gradient within the area.

[0059] The collected temperature data is transmitted to the local control device (embedded industrial computer) via an industrial router. At the edge, the system performs the following core processes: real-time calculation of regional temperature distribution characteristics; execution of an adaptive dynamic threshold algorithm; and preliminary alarm judgment and data classification.

[0060] The core of the system is an adaptive dynamic threshold algorithm. The working principle of the algorithm is described in [link to algorithm description]. Figure 3 The algorithm formula is: T POI -T Region_Avg >ΔT Dynamic , (T Region_Avg The average temperature of the region is calculated as the sum of the temperatures of all valid pixels within the region divided by the number of pixels; ΔT Dynamic (This refers to the dynamic alarm threshold), where the dynamic threshold ΔT Dynamic Adjust dynamically based on the current temperature gradient and historical fluctuations.

[0061] The system employs an intelligent data classification strategy. Regular data is collected at a preset frequency and transmitted using compressed data; alarm data is transmitted in real-time with priority, containing complete data packets. The reliability of data transmission is ensured through the breakpoint resume function of the industrial router.

[0062] The remote server cluster receives and stores various types of data, providing the following core functions: real-time data visualization; historical data backtracking and analysis; multi-dimensional data comparison; and intelligent report generation, such as... Figure 1 As shown.

[0063] Technological breakthrough

[0064] In traditional temperature monitoring, we typically focus on whether the temperature at a certain point exceeds a threshold (e.g., the temperature of a calender roll exceeds 500°C). However, the temperature gradient in this invention describes the rate and direction of temperature change in space within a specific region. It is a vector that contains both magnitude and direction information.

[0065] Temperature gradient itself is a diagnostic indicator. A small temperature gradient indicates a uniform temperature across the entire area, stable equipment operation, and a "quiet" state. In this case, the system considers the operating condition good, and any minor temperature anomaly warrants attention. Therefore, the alarm device will lower the dynamic threshold (ΔT). Dynamic This makes the system more sensitive, enabling it to detect subtle anomalies that stand out against a stable background, thus effectively reducing the risk of missed alarms. When the temperature gradient is large, it indicates significant uneven heating or cooling within the area, suggesting the equipment may be heating up, cooling down, or exhibiting inherent thermal inhomogeneity. In this context, normal fluctuations at individual measuring points may also be substantial. In this case, the alarm device will increase the dynamic threshold, making the system more tolerant and preventing normal operating condition fluctuations from being misinterpreted as fault alarms, thereby effectively reducing the false alarm rate.

[0066] Technical effect

[0067] 1. Significantly improved monitoring accuracy and reliability: Through deep integration of software and hardware and adaptive dynamic threshold algorithm, false alarms and missed alarms are effectively avoided, the system has a fast response speed and strong environmental adaptability.

[0068] 2. Highly efficient and optimized resource utilization: Intelligent data classification and differentiated storage strategies, combined with dedicated industrial hardware, significantly reduce the consumption of storage space and network bandwidth.

[0069] 3. Powerful data analysis capabilities: It provides visualization analysis tools that combine time and space dimensions, which can not only provide real-time alarms, but also provide in-depth data insights for predictive maintenance of equipment and process optimization.

[0070] 4. The system is a significant breakthrough: it abandons the fixed threshold alarm mechanism used in the industry for decades. It innovatively sets the alarm threshold as a dynamic variable (ΔT). DynamicThis variable is dynamically adjusted based on real-time calculations of regional temperature distribution characteristics and historical operating data. This means the system can "understand" whether the current operating conditions are normal, thus making more accurate judgments. This is a groundbreaking practice of applying artificial intelligence concepts to the underlying logic of industrial monitoring.

[0071] 1. Hardware Deployment: Install an infrared thermal imager above the inlet of the annealing furnace, such as... Figure 5 As shown, the industrial router connects to the industrial control computer, which in turn connects to a self-built server cluster.

[0072] 2. Software deployment: Deploy edge computing software on industrial control computers, deploy cloud microservices on servers, and deploy a centralized control platform in the control room.

[0073] 3. Workflow: The platform sets up monitoring areas and Points of Interest (POIs). The system continuously collects temperature data, calculates in real time, and uses dynamic thresholds for judgment. Once an anomaly is detected, an alarm is immediately triggered and complete data is recorded. Staff can use the platform for real-time monitoring and in-depth analysis, enabling precise management and predictive maintenance of the calender roll temperature.

[0074] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

Claims

1. A calender gradient control system, characterized by, 1) Data acquisition and feature calculation: Collect temperature data of the monitoring area through a dedicated infrared thermal imager; Real-time calculation of the temperature distribution characteristics of the area, including the average temperature, the highest / lowest temperature point and the temperature gradient; 2) Adaptive threshold calculation and alarm judgment: Use dynamic threshold algorithm, the alarm threshold is dynamically adjusted according to the current temperature distribution characteristics and historical data of the area; 3) Data classification and storage: The system classifies data into regular sampling data and alarm data; Use differentiated storage strategy, alarm data is synchronized to the remote server first; 4) Multi-dimensional data analysis and display: On the centralized control platform, provide multi-dimensional data analysis functions based on timeline, such as alarm picture sequence playback, temperature change trend chart, regional temperature distribution heat map, etc.

2. A calender gradient control system according to claim 1, wherein, The temperature data collected by the dedicated infrared thermal imager is transmitted to the local control device through the industrial router.

3. A calender gradient control system according to claim 1, wherein, The system uses intelligent data classification strategy, regular data is collected at a preset frequency and transmitted in compressed form; Alarm data is transmitted in real time and takes priority, including complete data packets.

4. A calender gradient control system according to claim 1, wherein, The remote server cluster receives and stores various data, and provides the following core functions: real-time data visualization, historical data backtracking analysis, multi-dimensional data comparison, and intelligent report generation.

Images