Search

CN-121760108-B - Yarn fineness online detection method and system based on self-learning

CN121760108BCN 121760108 BCN121760108 BCN 121760108BCN-121760108-B

Abstract

The invention provides a yarn fineness online detection method and system based on self-learning, and relates to the technical field of spinning. The method aims to solve the problems that the prior art depends on manual preset parameters, has poor adaptability and has low automation degree. The method comprises the steps of collecting original fineness information through a yarn fineness dynamic detection device, automatically judging whether the yarn enters a stable operation state, automatically entering a self-learning stage after stable operation, collecting a learning sample and analyzing the learning sample to calculate a standard characteristic value representing the current yarn fineness and a legal threshold range, switching to an online detection mode after self-learning is finished, comparing the real-time fineness information with the threshold range, and judging that the yarn is abnormal when the yarn exceeds the range and sending a control signal. According to the invention, through a self-learning mechanism, different yarns can be automatically adapted without preset parameters, so that the automation, the stable and reliable real-time monitoring of the detection process is realized, and the product quality is ensured.

Inventors

  • CHEN FENG
  • WANG JUNWEI
  • CUI JIEHAO

Assignees

  • 上海高适软件有限公司

Dates

Publication Date
20260505
Application Date
20260305

Claims (7)

  1. 1. The yarn fineness on-line detection method based on self-learning is characterized by comprising the following steps of: Step S1, non-contact optical signal acquisition is carried out on yarns moving in a production line through a photoelectric detection structure, optical signals are converted into electric signals, original fineness information is obtained, high-speed sampling and quantization are carried out on the original fineness information, and a real-time digital signal data stream is formed; s2, analyzing characteristic parameters of the digital signal data stream in real time, and automatically judging whether the yarn enters a continuous and stable running state; Step 3, when the yarn is judged to enter a stable running state, automatically triggering a self-learning process, collecting a digital signal data stream within a preset time length as a learning sample set, analyzing the learning sample set through a statistical algorithm to obtain a standard characteristic value and a fluctuation characteristic parameter of the normal fineness of the yarn in the current batch, and automatically generating a legal threshold range based on the standard characteristic value and the fluctuation characteristic parameter; step S4, continuously collecting real-time digital signal data streams of the yarns, comparing the numerical value of each sampling point with the legal threshold range, judging that the fineness of the yarns is abnormal if the numerical value of the sampling points exceeds the legal threshold range, and triggering corresponding abnormal processing operation; The characteristic parameter is a signal amplitude, and the specific judgment logic is as follows: presetting a first threshold value, wherein the first threshold value is higher than the background noise level and lower than the normal yarn signal level, and judging that the yarn enters a continuous and stable running state when the signal value is detected to be continuously larger than the first threshold value within a preset time; Or the characteristic parameter is signal energy, and the specific judgment logic is as follows: After the system is powered on, firstly, collecting background noise signals with preset duration, and calculating average short-time energy E noise of the background noise signals, and then continuously calculating short-time energy E realtime of the real-time digital signal data stream according to a preset sliding time window, and judging that the yarn enters a continuous and stable running state when E realtime continuously exceeds the preset multiple of E noise and continuously lasts for the preset duration; In step S3, the statistical algorithm includes: Calculating an arithmetic mean value mu of all data points in the learning sample set as a reference characteristic value, and calculating a standard deviation sigma as a fluctuation characteristic parameter; The legal threshold range is set to [ mu ] K.sigma, mu+k.sigma ], wherein k is a preset coefficient.
  2. 2. The method for online detection of yarn fineness based on self-learning according to claim 1, wherein in the self-learning process of step S3, after the learning sample set is collected, the learning sample set is first subjected to kalman filtering processing to filter out signal noise, and then the filtered sample data is analyzed by a statistical algorithm to obtain a reference characteristic value and a fluctuation characteristic parameter.
  3. 3. A yarn fineness on-line detection system based on self-learning, which is characterized by being used for realizing the detection method of any one of claims 1-2, wherein the system comprises a yarn fineness dynamic detection device and a sensor control chip; The yarn fineness dynamic detection device is used for realizing non-contact optical signal acquisition of yarns and comprises a guide assembly, a photoelectric sending device and a photoelectric receiving device, wherein the guide assembly is used for restraining movement tracks of the yarns to enable the yarns to stably pass through a preset detection area; The sensor control chip is used as a core processing unit, is integrated with a processor, a memory and an analog-to-digital converter, and is used for driving the photoelectric transmitting device to work, sampling and quantizing the electric signal output by the photoelectric receiving device, executing yarn running state judgment, self-learning, on-line detection and abnormality judgment processes and outputting an abnormality processing control signal.
  4. 4. The self-learning based yarn fineness online detection system according to claim 3, wherein the guide assembly is a U-shaped groove; the arc-shaped inner wall of the U-shaped groove is used for restraining the radial position of the yarn and preventing the yarn from swinging greatly; the yarn to be detected passes through a preset detection area from the bottom of the U-shaped groove under the traction of production equipment.
  5. 5. The self-learning based yarn fineness online detection system according to claim 3, wherein the guide assembly is a spiral guide tube; the inner wall of the spiral guide pipe is provided with a continuous spiral groove; When the yarn passes through the spiral guide pipe, stable circumferential rotation is generated under the action of the spiral groove, so that the sections of the yarn at different angles can be detected.
  6. 6. The on-line yarn fineness detection system based on self-learning according to claim 3, wherein the photoelectric transmitting device is an infrared light emitting diode, and the photoelectric receiving device is a photodiode or a phototransistor; The sensor control chip precisely controls the luminous intensity and the switching state of the photoelectric transmitting device through the luminous control module, and ensures the stability of the light source.
  7. 7. The on-line yarn fineness detection system based on self-learning as claimed in claim 3, further comprising an abnormality actuator electrically connected to the sensor control chip for receiving an abnormality processing control signal outputted from the sensor control chip and performing a corresponding operation; the abnormal executing mechanism comprises a motor controller of the yarn conveying equipment and an audible and visual alarm, wherein the motor controller receives the control signal and then controls the yarn conveying equipment to stop running, and the audible and visual alarm receives the control signal and then sends out an alarm prompt.

Description

Yarn fineness online detection method and system based on self-learning Technical Field The invention relates to the technical field of textile, in particular to a yarn fineness online detection method and system based on self-learning. Background Uniformity of yarn fineness is one of the key factors in determining the quality of the final fabric during yarn production and processing in the textile industry. The traditional yarn fineness detection method is mainly off-line sampling detection, namely, randomly extracting part of yarn samples from production batches, and measuring by using a special instrument. The method has obvious hysteresis, cannot monitor the production process in real time and comprehensively, and once the problems occur, the method often causes batchwise quality defects, and has low detection efficiency and high labor cost. To overcome the shortcomings of off-line detection, a variety of on-line detection techniques have been developed. Patent document CN111691028a discloses a roving frame with a monitoring system for producing rovings from sliver, the roving frame having a plurality of roving production positions, each roving production position comprising a drafting device, a shaft for receiving a bobbin and a flyer for winding the roving onto the bobbin. The roving frame comprises a measuring device comprising at least one measuring unit arranged between the transport roller of the drafting device and the flyer at each roving production location, wherein the plurality of measuring units each comprise at least one sensor and the measuring device is configured for continuously measuring at least one quality and/or production parameter of the roving at each roving production location by means of the measuring unit. The patent document sets a measuring device consisting of an optical sensor and a guiding device on a production line, continuously measures a yarn moving at a high speed, and compares the real-time measured value with a preset target value and an allowable deviation range. When the measured value exceeds the preset range, the system triggers an alarm or controls the production equipment to stop running. However, such solutions have the common disadvantage that they rely heavily on the operator to manually preset a standard fineness value and corresponding tolerance threshold according to the specifications of the yarn to be produced. When a different type or batch of yarn needs to be replaced, the manual setting must be re-performed. The mode depending on preset parameters is complicated in operation, poor in adaptability and easy to cause inaccurate detection due to human errors under the current flexible production trend of small batches and multiple varieties. In addition, how to ensure that the yarn moving at high speed can stably pass through a narrow detection area and how to enable a system to automatically judge the start-stop state of production so as to realize unmanned monitoring is also a problem which cannot be effectively solved by the prior art. Disclosure of Invention Aiming at the defects in the prior art, the invention aims to provide a yarn fineness online detection method and system based on self-learning, so as to solve the problems that an online yarn fineness detection scheme in the prior art depends on manual preset parameters, has poor adaptability and low automation degree, and aims to realize an online fineness detection scheme which can automatically adapt to yarns of different types and batches, does not need manual intervention and is stable and reliable in detection process. The invention provides a yarn fineness online detection method based on self-learning, which comprises the following steps: Step S1, non-contact optical signal acquisition is carried out on yarns moving in a production line through a photoelectric detection structure, optical signals are converted into electric signals, original fineness information is obtained, high-speed sampling and quantization are carried out on the original fineness information, and a real-time digital signal data stream is formed; s2, analyzing characteristic parameters of the digital signal data stream in real time, and automatically judging whether the yarn enters a continuous and stable running state; Step 3, when the yarn is judged to enter a stable running state, automatically triggering a self-learning process, collecting a digital signal data stream within a preset time length as a learning sample set, analyzing the learning sample set through a statistical algorithm to obtain a standard characteristic value and a fluctuation characteristic parameter of the normal fineness of the yarn in the current batch, and automatically generating a legal threshold range based on the standard characteristic value and the fluctuation characteristic parameter; And S4, continuously collecting real-time digital signal data streams of the yarns, comparing the numerical value of each sampling point with the leg