Intelligent monitoring method and system for flange forging production

Abstract

The present application belongs to the technical field of production line monitoring, and particularly relates to an intelligent monitoring method and system for flange forging production, which comprises the following steps: generating a monitoring data set; obtaining a data segment to be measured, calculating a characteristic index to be measured according to the data distribution in the data segment to be measured; calculating a measured average deviation, a measured trend symbol and a measured change index according to the difference between the characteristic index to be measured and the data to be measured in the same data segment to be measured, the fluctuation condition of adjacent data to be measured and the fluctuation degree; integrating the obtained data to form a dynamic vector to be measured; calculating a local density to be measured according to the distribution distance of the standardized characteristic component; calculating an abnormality judgment threshold according to the dispersion degree of the local density to be measured in the historical dynamic vector to be measured; and comparing the abnormality judgment threshold in real time to control the system. The present application solves the problems of single monitoring dimension and insufficient feature mining in the flange forging production process.

Description

Technical Field

[0001] This invention relates to the field of production line monitoring technology. More specifically, this invention relates to an intelligent monitoring method and system for flange forging production. Background Technology

[0002] Flange forgings are key components in petrochemical, energy, and shipbuilding industries. During their production, parameters such as temperature, pressure, and deformation directly determine the strength, toughness, and density of the product. If abnormal parameters are not detected in time, it can lead to the scrapping of forgings or even safety accidents. Therefore, real-time monitoring of the production process is necessary.

[0003] Traditional flange forging production monitoring relies on manual inspections, with staff using handheld devices to periodically collect parameters. This not only fails to capture instantaneous parameter fluctuations but also introduces human error in recording data. Some companies use a single sensor to monitor temperature, which only reflects local conditions and cannot cover key influencing factors such as pressure and deformation, resulting in delayed anomaly identification and a high rate of missed detections.

[0004] To achieve comprehensive and precise monitoring and control of the forging production process, the industry is currently committed to building efficient management systems using various advanced technologies. Among related technologies, for example, Chinese patent document CN118760109B, entitled "A Management System for an Intelligent Forging Production Line Based on the Internet of Things," discloses a scheme that connects equipment on the production line using IoT technology, collects real-time temperature data from the heating process to output a temperature control performance index, and collects parameters such as rolling force from the forging process to output a forging control performance index, thereby monitoring and controlling the process accordingly.

[0005] However, in actual monitoring, existing technologies have problems such as focusing only on typical parameters of specific processes and insufficient data feature mining, making it difficult to capture subtle and complex abnormal changes in the production process, unable to conduct detailed analysis of the production status from an overall perspective, and unable to achieve comprehensive and in-depth control over the production process, thus having a limited effect on improving production efficiency and product quality. Summary of the Invention

[0006] To address the problems of limited monitoring dimensions and insufficient feature mining in existing technologies, this invention provides solutions in the following aspects.

[0007] In a first aspect, the present invention provides an intelligent monitoring method for flange forging production, comprising:

[0008] At the same preset frequency, relevant data from temperature sensors, pressure sensors, and displacement sensors are simultaneously acquired, and the obtained data are arranged in chronological order to generate a monitoring dataset containing temperature data, pressure data, and deformation data. A test data segment containing temperature data segments, pressure data segments, and deformation data segments is acquired; based on the data distribution in the test data segment, a test characteristic index is calculated; based on the difference between the test characteristic index and the test data in the same test data segment, the test average deviation is calculated; based on the fluctuation status and fluctuation degree of adjacent test data in the same test data segment, the test trend is calculated. The system calculates the potential sign and the change index to be measured; it integrates the target feature index, target average deviation, target trend sign, and target change index corresponding to the target data segment to form the target dynamic vector; it counts the number of feature component types in the target dynamic vector, and calculates the target local density based on the number of feature component types and the distribution distance of the standardized feature components; it acquires and calculates historical target dynamic vectors, and calculates the anomaly judgment threshold based on the dispersion of the target local density in the historical target dynamic vectors; it compares the target local density with the anomaly judgment threshold in real time, and controls the system based on the comparison results.

[0009] This invention utilizes multiple sensors to synchronously collect temperature, pressure, and deformation data at the same frequency, covering the entire production process from heating and forging to forming. The characteristics of these parameters are then integrated into a dynamic vector, transforming it into easily identifiable indicators. Simultaneously, adjustable anomaly thresholds are set based on historical normal data, rather than fixed standards. This allows for real-time and comprehensive monitoring of the production status, immediately triggering warnings in case of anomalies, and automatically adjusting equipment for persistent anomalies, eliminating the need for constant manual monitoring. This effectively prevents forgings from being scrapped due to parameter anomalies and also prevents safety accidents. It to some extent compensates for the shortcomings of traditional monitoring in terms of real-time performance, comprehensiveness, and accuracy, making flange forging production monitoring more reliable.

[0010] Preferably, the calculation of the feature index to be measured includes:

[0011] Select temperature, pressure, and deformation data from N consecutive time points to form a data segment, denoted as the temperature data segment, pressure segment, and deformation data segment. Then, denote three data segments containing the same number of data points to be measured as the measured data segment. All the first-order data segments to be tested The sum of the specific values ​​of the nth test data and the nth All the first-order data segments to be tested The ratio of the sum of the squares of the specific values ​​of the data to be measured is denoted as the characteristic index to be measured.

[0012] Preferably, the average deviation to be measured satisfies the following expression:

[0013] ;

[0014] In the formula, Indicates the first The average deviation of the data segment to be tested; Indicates the first The first of the data segments to be tested The specific value of each data point to be tested; Indicates the first The feature indicators to be tested in the data segment to be tested; Indicates the first The number of data points to be tested in the data segment to be tested; This represents the absolute value function.

[0015] This invention calculates the average deviation by first determining the absolute difference between each monitoring data point and the characteristic coefficient, and then averaging these differences. This method provides a direct understanding of the gap between each monitoring data point and the ideal production state. For example, during the heating stage, it reveals the average difference between the actual temperature and the ideal temperature; during the forging stage, it reveals the average difference between the pressure and the ideal pressure. If the difference becomes too large, parameter fluctuations can be detected promptly, preventing persistent temperature spikes and pressure instability, thus preventing quality defects in forgings due to parameter fluctuations and making the assessment of production parameter stability more accurate.

[0016] Preferably, the trend sign to be measured is calculated to satisfy the following expression:

[0017] ;

[0018] In the formula, Indicates the first The trend symbol to be measured for each data segment to be measured; Indicates the first The number of data points to be tested in the data segment to be tested; , Indicates the first The first of the data segments to be tested The first data to be tested, the first The specific value of each data point to be tested; Represents a symbolic function.

[0019] This invention calculates trend signs by comparing changes in adjacent monitoring data to determine whether the parameter is rising, falling, or remaining stable, and then summing these trends. This allows for a clear understanding of the overall direction of change in production parameters, such as whether temperature is continuously rising or pressure is continuously falling. If such trend changes are not detected in time, anomalies may gradually escalate. However, through this calculation, abnormal trends can be detected early, preventing the anomalies from continuing to expand and ensuring that production parameters always develop in a stable direction.

[0020] Preferably, the measured change index is calculated to satisfy the following expression:

[0021] ;

[0022] In the formula, Indicates the first The change index to be measured for the data segment to be measured; , Indicates the first The first of the data segments to be tested The first data to be tested, the first The specific value of each data point to be tested; Indicates the first The number of data points to be tested in the data segment to be tested; This represents the absolute value function.

[0023] This invention calculates the average chord slope by first determining the absolute difference between adjacent monitoring data, and then averaging these differences. This method can detect the rate of change in production parameters, such as whether there is a sudden rise in temperature or a sudden drop in pressure. Drastic parameter changes can easily lead to forging quality problems; for example, a sudden temperature rise may cause localized overheating of the forging. Through this calculation, drastic changes can be detected in a timely manner, allowing staff or the system to intervene early, avoiding the impact of drastic fluctuations on forging quality and maintaining the stability of production parameter changes.

[0024] Preferably, forming the dynamic vector to be measured includes:

[0025] The test data segments containing temperature data segments, pressure data segments, and deformation data segments are integrated with the corresponding test feature indicators, test average deviation, test trend sign, and test change indicators to form a test dynamic vector; the test dynamic vector includes temperature test dynamic vector, pressure test dynamic vector, and deformation test dynamic vector.

[0026] Preferably, calculating the local density to be measured includes:

[0027] Obtain the dynamic vector to be tested, count the number of feature component types of all types in the dynamic vector to be tested, and denote it as the number of feature component types. Standardize each type of feature component, and calculate the local density to be tested based on the ratio of the standardized feature component to the number of feature component types.

[0028] Preferably, the anomaly detection threshold is calculated to satisfy the following expression:

[0029] ;

[0030] In the formula, Represents the first dynamic vector to be measured. Anomaly detection threshold for a given data segment to be tested; Represents the first in the historical dynamic vector to be measured. The local density of the data segment to be tested; Represents the first in the historical dynamic vector to be measured. The standard deviation of the local density of the data segment to be tested; This represents the adjustment factor.

[0031] When calculating the anomaly detection threshold, this invention references the local density and normal fluctuation range of historical normal production, and then combines them with an adjustment coefficient. This ensures the threshold is not fixed and can be flexibly adjusted according to actual normal production conditions. For example, when raw materials change slightly or equipment status differs slightly, the threshold can adapt to these minor changes, preventing normal situations from being mistakenly treated as anomalies, and also preventing the omission of genuine anomalies due to a fixed threshold, thus reducing false positives and false negatives and making anomaly detection more accurate.

[0032] Preferably, controlling the system includes:

[0033] Will and Perform real-time comparisons and control the system based on the comparison results; if The system continues to maintain real-time monitoring; if The system immediately triggers an anomaly warning mechanism; if three consecutive test data segments are detected... The system automatically sends control commands to the control systems of the forging press and the heating furnace: after receiving the command, the forging press will reduce its running speed and reduce the forging frequency; after receiving the command, the heating furnace will suspend the temperature increase operation and maintain the current temperature.

[0034] Secondly, the present invention provides an intelligent monitoring system for flange forging production, comprising a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned intelligent monitoring method for flange forging production is implemented.

[0035] By adopting the above technical solution, a computer program is generated from the intelligent monitoring method for flange forging production and stored in a memory so that it can be loaded and executed by a processor. This allows for the creation of a terminal device based on the memory and processor, making it convenient to use.

[0036] The beneficial effects of this invention are as follows: This invention has certain macro-level value for the flange forging manufacturing industry and downstream sectors. As core components in key fields such as petrochemicals, energy, and shipbuilding, the quality of flange forgings directly determines the safe operation and service life of downstream equipment. This invention provides strong assurance for the production quality of these key components. From the production perspective, this invention eliminates the need for frequent manual inspections, reducing labor costs and error rates associated with manual operations. Simultaneously, through automatic early warning and equipment control, it significantly reduces the probability of producing defective forgings, minimizing economic losses due to product scrap and improving production efficiency. From an industry development perspective, this invention promotes the transformation of flange forging production from traditional manual monitoring to intelligent and automated monitoring, breaking the limitations of single-parameter monitoring and providing the industry with a more scientific and comprehensive monitoring solution. Attached Figure Description

[0037] Figure 1 This is a flowchart illustrating an intelligent monitoring method for flange forging production according to the present invention. Detailed Implementation

[0038] This invention discloses an intelligent monitoring method for flange forging production, referring to... Figure 1 This includes steps S1-S4:

[0039] S1: Simultaneously acquire relevant data from temperature sensors, pressure sensors, and displacement sensors at the same preset frequency, and arrange the acquired data in chronological order to generate a monitoring dataset containing temperature data, pressure data, and deformation data.

[0040] It should be noted that flange forgings are key components in petrochemical, energy, and shipbuilding industries. Temperature, pressure, and deformation during production directly determine their strength, toughness, and density. Abnormal parameters can easily lead to forging scrap or safety accidents, necessitating real-time monitoring. Traditional manual inspection struggles to capture instantaneous fluctuations and suffers from recording errors. Single sensors only measure temperature, failing to cover all key parameters at every stage. Multi-parameter feature integration and analysis techniques are insufficient, resulting in delayed anomaly identification and a high rate of missed detections. This invention addresses these issues by using temperature, pressure, and displacement sensors to collect data simultaneously, generating a monitoring dataset. Data segment features are extracted, and parameters such as feature coefficients and average deviations are calculated and integrated into a dynamic vector. Based on the dynamic vector, local density is calculated, and anomaly thresholds are calculated using historical data. Real-time comparison with the thresholds ensures continuous monitoring if normal data is detected. If anomalies are detected, the system triggers an early warning. If continuous anomalies are detected, the system automatically adjusts equipment, achieving intelligent monitoring of the entire flange forging production process and ensuring production quality and safety.

[0041] It should be noted that flange forging production is divided into three key stages: heating, forging, and forming. The parameters of each stage have different impacts on product quality. Insufficient temperature during the heating stage can lead to insufficient metal plasticity; insufficient pressure during the forging stage can cause loose grains; and excessive deformation during the forming stage can affect dimensional accuracy. Using only one sensor can only monitor the status of some stages and cannot provide a comprehensive understanding of the production process. Therefore, multiple sensors are needed to collect data simultaneously to ensure that the key parameters of each stage are captured in real time, providing a complete data foundation for subsequent monitoring.

[0042] Specifically, a temperature sensor is installed at the outlet of the heating furnace in the flange forging production workshop; a pressure sensor is installed under the worktable of the forging press; and a displacement sensor is installed on the side of the forming inspection station. The sampling frequency of the three sensors is set to a times / second, and the temperature data, pressure data, and deformation data collected by the three sensors are arranged in order of collection time to form a monitoring dataset.

[0043] It should be noted that the thermocouple temperature sensor has a measurement range of 0-1200℃, covering the temperature range for forging heating. The piezoelectric pressure sensor has a measurement range of 0-200MPa, matching the pressure requirements of the forging process. The laser displacement sensor has a measurement range of 0-500mm, covering the deformation range of common flange forgings, and can acquire deformation data of the forging in real time. To ensure timely capture of instantaneous changes in temperature, pressure, and deformation, a=1 is chosen as the sampling frequency.

[0044] At this point, a monitoring dataset containing temperature data, pressure data, and deformation data has been obtained.

[0045] S2: Obtain the test data segment containing temperature data segment, pressure data segment, and deformation data segment; calculate the test characteristic index based on the data distribution in the test data segment; calculate the test average deviation based on the difference between the test characteristic index and the test data in the same test data segment; calculate the test trend sign and the test change index based on the fluctuation status and fluctuation degree of adjacent test data in the same test data segment; integrate the test characteristic index, test average deviation, test trend sign, and test change index corresponding to the test data segment to form the test dynamic vector.

[0046] It should be noted that the monitoring dataset consists of temperature, pressure, and deformation data at different times. This data only reflects the parameter magnitude at a specific moment and cannot reflect the trend and overall fluctuation of the parameters. For example, if the temperature is normal at a certain moment, but rises rapidly for 10 consecutive seconds, this trend change may indicate an anomaly, but it cannot be detected from a single data point. This invention aims to extract feature parameters that reflect trends and fluctuations, and then integrate them into a feature vector. This transforms the raw data into comprehensive information that reflects the production status, facilitating subsequent judgment of whether anomalies exist.

[0047] Specifically, temperature data, pressure data, and deformation data from N consecutive time points are selected to form a data segment, denoted as the temperature data segment, pressure data segment, and deformation data segment. Three data segments containing the same number of data points to be measured are denoted as the measured data segment. The calculation of the... All the first ones in the data segment to be tested The sum of the specific values ​​of the nth test data is denoted as the total sum of the test data; calculate the nth... All the first ones in the data segment to be tested The sum of the squares of the specific values ​​of the data to be tested is denoted as the sum of squares of the data to be tested; based on the ratio of the sum of squares of the data to be tested to the sum of squares of the data to be tested, the characteristic coefficients of the data segment to be tested are calculated and denoted as the characteristic indices to be tested, including:

[0048] The feature index to be measured satisfies the following expression:

[0049] ;

[0050] In the formula, Indicates the first The feature indicators to be tested in the data segment to be tested; Indicates the first The number of data points to be tested in the data segment to be tested; Indicates the first The first of the data segments to be tested The specific value of each data point to be tested.

[0051] In the formula, Indicates all the first elements in the data segment to be tested. The first of the data segments to be tested The sum of the specific values ​​of each data point to be tested; Indicates all the first elements in the data segment to be tested. The first of the data segments to be tested The sum of the squares of the specific values ​​of the data to be tested; This ratio represents the central tendency of the data, calculated by dividing the overall level by the dispersion. A larger ratio indicates that the data is more concentrated around a certain level, while a smaller ratio indicates that the data is more dispersed. Larger means The larger, that is, the first The first of the data segments to be tested The more concentrated the data to be tested, the better.

[0052] It's important to note that in flange forging production, the stability of parameters such as heating temperature and forging pressure directly determines the quality of the forgings. Calculating the average deviation of the measured values ​​clearly shows the degree of deviation between each monitored data point and the ideal state within a production period. For example, a small average deviation in furnace temperature indicates that the actual temperature during this period meets the ideal temperature trend, heating is stable, and the forging material is uniform. A large average deviation suggests fluctuating temperatures, potentially leading to overheating or incomplete burning in certain areas of the forging, affecting quality. Similarly, a large average deviation in pressure data indicates a problem with the hydraulic system, requiring immediate repair to prevent quality defects in the forgings.

[0053] Preferably, a data segment of the same length as the one used to calculate the target feature index is obtained. Based on the difference between the target data and the target feature index in the data segment, the average deviation of the data segment is calculated and denoted as the average deviation of the target data segment, including:

[0054] The measured average deviation satisfies the following expression:

[0055] ;

[0056] In the formula, Indicates the first The average deviation of the data segment to be tested; Indicates the first The first of the data segments to be tested The specific value of each data point to be tested; Indicates the first The feature indicators to be tested in the data segment to be tested; Indicates the first The number of data points to be tested in the data segment to be tested; This represents the absolute value function.

[0057] In the formula, Indicates the first The first of the data segments to be tested The absolute difference between the specific value of each data point to be tested and the corresponding characteristic indicator to be tested; Indicates the first Within a given data segment, the sum of the absolute differences between all the data to be tested and their corresponding characteristic indicators; This means averaging the total deviation across each data point to obtain the average deviation.

[0058] It should be noted that in the monitoring scenario of flange forging production, traditional manual inspection and single-sensor monitoring have shortcomings, making it difficult to accurately capture parameter fluctuations and the overall status of flange forging production. To solve this problem, this invention introduces the calculation of the trend sign of the data to be measured. By comparing the changes of adjacent data in the data segment to be measured, the sign function is used to determine whether each adjacent data is rising, falling, or stable. Then, these trend signs are accumulated, and the sum can clearly show whether the overall trend of the data segment is rising, falling, or stable.

[0059] Preferably, a data segment of the same length as the one used to calculate the target feature index is obtained, and the trend sign of the data segment is calculated and denoted as the target trend sign, including:

[0060] The trend symbol to be tested satisfies the following expression:

[0061] ;

[0062] In the formula, Indicates the first The trend symbol to be measured for each data segment to be measured; Indicates the first The number of data points to be tested in the data segment to be tested; , Indicates the first The first of the data segments to be tested The first data to be tested, the first The specific value of each data point to be tested; Represents a symbolic function.

[0063] In the formula, Indicates the first The first of the data segments to be tested The first test data and the first The changing trend of the data to be tested; Indicates the first In a given data segment to be tested, the sum of the trend signs of all adjacent data points to be tested is used to describe the trend sign of the data segment to be tested; if A positive value with a large absolute value indicates that the data being tested is trending upwards. A negative value with a large absolute value indicates that the measured data is showing a downward trend. A value of 0 indicates that the data being tested fluctuates smoothly.

[0064] It should be noted that in flange forging production monitoring, existing technologies struggle to capture parameter fluctuations in a timely and accurate manner, while single-sensor monitoring suffers from dimensional limitations, easily leading to problems such as delayed anomaly identification. This invention obtains a data segment of equal length to the measured characteristic index, compares the absolute differences between adjacent data points, and then averages the results to obtain the measured change index. This index quantifies the steepness of the average change in the data segment; the more drastic the change, the greater the average sloping angle, thus accurately reflecting the severity of production parameter fluctuations.

[0065] Preferably, a data segment of the same length as the one used to calculate the characteristic index to be measured is obtained, and the average sine slope of the data segment to be measured is calculated and denoted as the change index to be measured, including:

[0066] The change index to be measured satisfies the following expression:

[0067] ;

[0068] In the formula, Indicates the first The change index to be measured for the data segment to be measured; , Indicates the first The first of the data segments to be tested The first data to be tested, the first The specific value of each data point to be tested; Indicates the first The number of data points to be tested in the data segment to be tested; This represents the absolute value function.

[0069] In the formula, Indicates the first The first of the data segments to be tested The first test data and the first The absolute difference between the test data; Indicates the first The average steepness of change in the data segment to be tested; The larger the value, the more likely it is to be the first. The steeper the average change of the data segment to be tested, the greater the steepness.

[0070] It should be noted that single-sensor monitoring can only reflect a local state and is insufficient to cover key parameters such as temperature, pressure, and deformation, leading to delayed anomaly identification and a high rate of missed detection. This invention integrates the characteristic coefficients, average deviations, trend signs, and average sine slopes corresponding to the three types of data segments to be measured—temperature, pressure, and deformation—to form a dynamic vector containing the key features of each of the three parameters. This integration method transforms scattered single-parameter features into comprehensive feature information that fully reflects the production status, avoiding the omission of key anomaly signals due to the limitations of single-parameter monitoring.

[0071] Preferably, the test feature index, test average deviation, test trend sign, and test change index corresponding to the test data segments containing temperature data segments, pressure data segments, and deformation data segments are integrated to form a test dynamic vector; the test dynamic vector includes a temperature test dynamic vector, a pressure test dynamic vector, and a deformation test dynamic vector; the temperature, pressure, and deformation test dynamic vectors include the temperature, pressure, and deformation test feature indexes, the temperature, pressure, and deformation test average deviations, the temperature, pressure, and deformation test trend signs, and the temperature, pressure, and deformation test change indexes.

[0072] At this point, the dynamic vector to be tested has been obtained.

[0073] S3: Count the number of feature component types in the dynamic vector to be tested. Based on the number of feature component types and the distribution distance of the standardized feature components, calculate the local density to be tested. Obtain and calculate the historical dynamic vector to be tested. Calculate the anomaly judgment threshold based on the dispersion of the local density to be tested in the historical dynamic vector to be tested.

[0074] It should be noted that the dynamic vector to be tested contains comprehensive features of multi-dimensional parameters, but these features cannot be directly used to judge anomalies. They need to be transformed into a quantifiable indicator, and the thresholds for normal and abnormal should be determined. The thresholds cannot be fixed, because the state of raw materials and equipment may change slightly during the production process. Fixed thresholds are prone to misjudgment or missed judgment. This invention sets the thresholds based on the statistical characteristics of historical normal data, which makes the thresholds more in line with the actual production situation.

[0075] Specifically, the dynamic vector to be measured is obtained, the number of feature component types of all types in the dynamic vector to be measured is counted and denoted as the number of feature component types, and each type of feature component is standardized. Based on the ratio of the standardized feature component to the number of feature component types, the local density to be measured is calculated, including:

[0076] The local density to be measured satisfies the following expression:

[0077] ;

[0078] In the formula, Represents the first dynamic vector to be measured. The local density of the data segment to be tested; Represents the first dynamic vector to be measured. A type of dynamic vector; This represents the number of all types of dynamic vectors in the dynamic vector to be tested; This represents the normalization function.

[0079] In the formula, Represents the first dynamic vector to be measured. The first type of dynamic vector The standardized values ​​of each test data point are denoted as feature components; This means that the first element in the dynamic vector to be measured is... All feature components of a given feature component are summed; there are a total of 12 feature components in this invention, namely... =12, for example, there are two types of temperature measurement characteristic indicators and temperature measurement average deviation; This represents the local density to be measured in the dynamic vector to be measured. The larger the value, the greater the local density of the dynamic vector to be measured;

[0080] It should be noted that this invention accurately distinguishes production states by calculating an anomaly judgment threshold. This threshold is calculated based on local density data under historical normal production conditions. It uses typical local density levels during normal production as a foundation, combined with the fluctuation range of local density values ​​under normal conditions, and then adjusts the threshold's leniency through a coefficient adjustment, ultimately obtaining an anomaly judgment standard that adapts to actual production changes. This calculation logic avoids misjudgments caused by minor changes in raw materials and equipment conditions during production due to fixed thresholds, while also allowing local densities exceeding the threshold under abnormal conditions to promptly capture genuine production anomalies, providing a more reliable basis for monitoring flange forging production.

[0081] Preferably, the process involves acquiring and calculating dynamic vectors of the same length and type under historical normal production conditions, denoted as historical dynamic vectors; and calculating anomaly judgment thresholds based on these historical dynamic vectors, including:

[0082] The anomaly detection threshold satisfies the following expression:

[0083] ;

[0084] In the formula, Represents the first dynamic vector to be measured. Anomaly detection threshold for a given data segment to be tested; Represents the first in the historical dynamic vector to be measured. The local density of the data segment to be tested represents the typical density level of normal production. Represents the first in the historical dynamic vector to be measured. The standard deviation of the local density of the data segment to be tested reflects the fluctuation range of the density value under normal conditions; This represents the adjustment coefficient, used to control the leniency of the threshold; here, m=2. In this process, thresholds are calculated based on the mean and standard deviation of relevant data under historical production conditions, which can ensure that most normal production conditions are met. All The following also allows abnormal states to be... Can exceed This not only avoids normal states being misjudged as abnormal, but also allows for timely detection of genuine anomalies.

[0085] Thus, the local density to be measured and the anomaly detection threshold were obtained.

[0086] S4: Compare the local density to be measured with the anomaly judgment threshold in real time, and control the system based on the comparison results.

[0087] It should be noted that this invention forms a complete anomaly response mechanism by comparing the real-time calculated local density to be measured with an anomaly judgment threshold determined based on historical normal data. When the local density to be measured is within the normal range, it indicates that the production parameters conform to normal fluctuation patterns, and the system continuously monitors to ensure stable production. Once the local density exceeds the normal range, it indicates that the parameters deviate from the normal state, and multi-form early warning is immediately activated to quickly notify the staff. If the anomaly persists, the production equipment is further automatically adjusted to avoid the production of more defective forgings due to the expansion of the anomaly.

[0088] Specifically, and Perform real-time comparisons and control the system based on the comparison results; if The system continues to maintain real-time monitoring, updating the data segment to be tested every 'a' seconds and recalculating. And compare; if The system immediately triggers the anomaly warning mechanism, which activates a red alarm light and a buzzer on the central control panel in the production workshop. The system also automatically sends a text message to the production manager's mobile phone, containing the specific time the anomaly occurred, the type of parameters involved, and the current status. Values ​​and thresholds Value; if all three consecutive test data segments are detected The system automatically sends control commands to the control systems of the forging press and the heating furnace: after receiving the command, the forging press will reduce its running speed and reduce the forging frequency; after receiving the command, the heating furnace will suspend the temperature increase operation and maintain the current temperature.

[0089] This completes the intelligent monitoring of the entire flange forging production process.

[0090] This invention also discloses an intelligent monitoring system for flange forging production, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent monitoring method for flange forging production according to the present invention is implemented.

[0091] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0092] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A smart monitoring method for flange forging production, characterized in that, include: At the same preset frequency, relevant data from temperature sensors, pressure sensors, and displacement sensors are acquired simultaneously, and the obtained data are arranged in chronological order to generate a monitoring dataset containing temperature data, pressure data, and deformation data. Obtain the test data segment containing temperature data segment, pressure data segment, and deformation data segment. Based on the data distribution in the test data segment, calculate the test characteristic index, including: selecting temperature data, pressure data, and deformation data from N consecutive time points to form a data segment, denoted as temperature data segment, pressure data segment, and deformation data segment; and denoting three data segments containing the same number of test data points as the test data segment; [The text abruptly ends here, likely due to an incomplete sentence or a missing section.] All the first-order data segments to be tested The sum of the specific values ​​of the nth test data and the nth All the first-order data segments to be tested The ratio of the sum of the squares of the specific values ​​of the data to be measured is denoted as the characteristic index to be measured. Calculate the average deviation of the target feature index based on its difference from the target data in the same data segment; calculate the trend sign and the change index based on the fluctuation status and degree of fluctuation of adjacent target data in the same data segment, satisfying the expression: In the formula, Indicates the first The change index to be measured for the data segment to be measured. , Indicates the first The first of the data segments to be tested The first data to be tested, the first The specific value of each data point to be tested. Indicates the first The number of data points to be tested in the data segment to be tested. Represents the absolute value function; The test feature index, test average deviation, test trend sign, and test change index corresponding to the test data segment are integrated to form the test dynamic vector; The system counts the number of feature component types in the dynamic vector to be tested, and calculates the local density based on the number of feature component types and the distribution distance of the standardized feature components. It also acquires and calculates historical dynamic vectors to be tested, and calculates the anomaly judgment threshold based on the dispersion of the local density in the historical dynamic vectors to be tested. The system compares the local density to be tested with the anomaly judgment threshold in real time, and controls the system based on the comparison results.

2. The intelligent monitoring method for flange forging production according to claim 1, characterized in that, The calculated average deviation of the test object satisfies the following expression: ； In the formula, Indicates the first The average deviation of the data segment to be tested; Indicates the first The first of the data segments to be tested The specific value of each data point to be tested; Indicates the first The feature indicators to be tested in the data segment to be tested; Indicates the first The number of data points to be tested in the data segment to be tested; This represents the absolute value function.

3. The intelligent monitoring method for flange forging production according to claim 1, characterized in that, The calculated trend symbol satisfies the following expression: ； In the formula, Indicates the first The trend symbol to be measured for each data segment to be measured; Indicates the first The number of data points to be tested in the data segment to be tested; , Indicates the first The first of the data segments to be tested The first data to be tested, the first The specific value of each data point to be tested; Represents a symbolic function.

4. The intelligent monitoring method for flange forging production according to claim 1, characterized in that, The process of forming the dynamic vector to be measured includes: The test data segments containing temperature data segments, pressure data segments, and deformation data segments are integrated with the corresponding test feature indicators, test average deviation, test trend sign, and test change indicators to form a test dynamic vector; the test dynamic vector includes temperature test dynamic vector, pressure test dynamic vector, and deformation test dynamic vector.

5. The intelligent monitoring method for flange forging production according to claim 1, characterized in that, The calculation of the local density to be measured includes: Obtain the dynamic vector to be tested, count the number of feature component types of all types in the dynamic vector to be tested, and denote it as the number of feature component types. Standardize each type of feature component, and calculate the local density to be tested based on the ratio of the standardized feature component to the number of feature component types.

6. The intelligent monitoring method for flange forging production according to claim 1, characterized in that, The calculated anomaly detection threshold satisfies the following expression: ； In the formula, Represents the first dynamic vector to be measured. Anomaly detection threshold for a given data segment to be tested; Represents the first in the historical dynamic vector to be measured. The local density of the data segment to be tested; Represents the first in the historical dynamic vector to be measured. The standard deviation of the local density of the data segment to be tested; This represents the adjustment factor.

7. The intelligent monitoring method for flange forging production according to claim 1, characterized in that, The control of the system includes: Will and Perform real-time comparisons and control the system based on the comparison results; if The system continues to maintain real-time monitoring; if The system immediately triggers an anomaly warning mechanism; if three consecutive test data segments are detected... The system automatically sends control commands to the control systems of the forging press and the heating furnace: after receiving the command, the forging press will reduce its running speed and reduce the forging frequency; after receiving the command, the heating furnace will suspend the temperature increase operation and maintain the current temperature.

8. An intelligent monitoring system for flange forging production, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement an intelligent monitoring method for flange forging production according to any one of claims 1-7.

Images