CN-121977655-A - Crimping quality real-time detection method and system

Abstract

The invention relates to the technical field of semiconductor packaging, in particular to a method and a system for detecting crimping quality in real time, wherein the method comprises the steps of collecting pressure sensor data and displacement encoder data in the crimping process in real time to generate a pressure-displacement curve; extracting characteristic parameters of a pressure-displacement curve, including a pressure fluctuation amplitude of a contact stage, a local slope change value of a main pressure interval and a high-frequency oscillation frequency of a pressure maintaining stage, synchronously normalizing the pressure fluctuation amplitude, the local slope change value and the high-frequency oscillation frequency into dimensionless risk values, respectively carrying out weighted fusion on the risk values through respective weight coefficients, calculating comprehensive quality scores, and judging the compression joint quality according to the comprehensive quality scores. According to the method, three-stage characteristic parameters are extracted through the pressure-displacement curve, a traditional single-point threshold method is replaced, transient anomalies are accurately captured, quick response can be achieved, and the hysteresis problem of offline analysis of crimping quality is broken through.

Inventors

• TANG GUANGMING
• JIANG SHUAI
• Xu chunxing

Assignees

• 常州科瑞尔科技有限公司

Dates

Publication Date

20260505

Application Date

20260403

Claims (10)

1. 1. The real-time detection method of the crimping quality is applied to a servo crimping system of a connector and a PCB, and is characterized by comprising the following steps: collecting pressure sensor data and displacement encoder data in the crimping process in real time, and generating a pressure-displacement curve; Extracting characteristic parameters of the pressure-displacement curve, wherein the characteristic parameters comprise the pressure fluctuation amplitude of a contact stage, the local slope change value of a main pressure interval and the high-frequency oscillation frequency of a pressure maintaining stage; Synchronously normalizing the pressure fluctuation amplitude, the local slope change value and the high-frequency oscillation frequency into dimensionless risk values A, S and F; The weight coefficients W 1 、W 2 and W 3 are used for carrying out weighted fusion on A, S and F respectively, and the comprehensive mass fraction Q is calculated, wherein Q=W 1 ×(1-A)+W 2 ×(1-S)+W 3 × (1-F); and judging the crimping quality according to the comprehensive quality score.
2. 2. The method of claim 1, further comprising a parameter sensitivity enhancement process comprising: Applying zero-mean Gaussian noise disturbance to the original pressure-displacement curve to generate N groups of disturbance curves, wherein N is more than or equal to 100; Calculating the variation coefficient of each characteristic parameter; And strengthening the corresponding weight coefficient according to the sensitivity of each characteristic parameter.
3. 3. The method of claim 1, wherein the dynamic adjustment of the weight coefficients W 1 、W 2 and W 3 comprises: monitoring defect type distribution of a plurality of continuous PCBs; And when the proportion of the similar defects exceeds the set proportion, the weight coefficient of the corresponding parameters of the defects is improved.
4. 4. The method according to claim 1, wherein the step of determining the magnitude of the pressure fluctuation in the contact stage includes: identifying a displacement point with the pressure continuously greater than the initial contact pressure threshold as a contact starting point; And calculating the standard deviation of the pressure in a preset displacement interval after the contact starting point, and taking the standard deviation as the pressure fluctuation amplitude.
5. 5. The method according to claim 1, wherein the determining of the local slope change value of the main pressure section includes: calculating the pressure change rate in real time by using a sliding time window or a displacement window when the displacement is in the main pressure stroke interval; And recording the absolute value peak value of the pressure change rate between adjacent windows as the local slope change value.
6. 6. The crimping quality real-time detection method according to claim 1, wherein the step of determining the high-frequency oscillation frequency of the dwell stage includes: Carrying out frequency domain transformation analysis on pressure data in the later stage of the pressure maintaining stage; Calculating the ratio of the energy of a preset risk frequency band to the energy of a full frequency band, and taking the ratio as a quantized value of the high-frequency oscillation frequency; The preset risk frequency band is set based on resonance characteristics of the PCB base material.
7. 7. The method for detecting crimp quality in real time according to claim 6, wherein the frequency band distribution of the high-frequency oscillation frequency in the pressure maintaining stage is monitored in real time; if the high-frequency energy is detected to migrate to the low frequency, the weight coefficient W 1 of the pressure fluctuation amplitude is improved, and the weight coefficient W 3 of the high-frequency oscillation frequency is reduced; If the transition from low-frequency energy to high-frequency energy is detected, the weight coefficient W 2 of the local slope change value is improved, and the weight coefficient W 1 of the pressure fluctuation range is reduced.
8. 8. The crimping quality real-time detection method according to claim 1, wherein an environmental disturbance decoupling process is performed before extracting the characteristic parameter, comprising: Acquiring environmental data of a crimping area in real time through a temperature and humidity sensor; Inputting the environmental data into a pre-trained LSTM network, and outputting the pressure compensation quantity of a pressure-displacement curve; A new pressure-displacement curve is generated for the extraction of the characteristic parameters based on the pressure compensation quantity.
9. 9. The method of claim 8, further comprising dynamically compensating for reliability monitoring, comprising: adopting a normalized ratio of the absolute value of the pressure compensation quantity to a preset compensation upper limit as a compensation confidence coefficient; when the compensation confidence is below a threshold, performing any of the following operations: If the displacement difference value of the curve before and after compensation at the contact starting point exceeds the limit, extracting the characteristic parameters by adopting an original curve; otherwise, weighting and fusing the comprehensive mass fractions of the compensation front and rear curves according to the confidence.
10. 10. The utility model provides a crimping quality real-time detection system, is applied to servo crimping system of connector and PCB board, its characterized in that includes: The data acquisition module acquires the data of the pressure sensor and the data of the displacement encoder in real time and generates a pressure-displacement curve; The characteristic extraction module is used for extracting characteristic parameters of the pressure-displacement curve, including the pressure fluctuation amplitude of the contact stage, the local slope change value of the main pressure section and the high-frequency oscillation frequency of the pressure maintaining stage; the quality judging module is used for: Synchronously normalizing the pressure fluctuation amplitude, the local slope change value and the high-frequency oscillation frequency into dimensionless risk values A, S and F; The weight coefficients W 1 、W 2 and W 3 are used for carrying out weighted fusion on A, S and F respectively, and the comprehensive mass fraction Q is calculated, wherein Q=W 1 ×(1-A)+W 2 ×(1-S)+W 3 × (1-F); and judging the crimping quality according to the comprehensive quality score.

Description

Crimping quality real-time detection method and system Technical Field The invention relates to the technical field of semiconductor packaging, in particular to a method and a system for detecting crimping quality in real time. Background In the crimping process of the connector and the PCB, the traditional quality detection mainly depends on manual sampling inspection or off-line analysis, and has the problems of low efficiency and poor real-time performance. The existing online monitoring system is mostly based on a single static threshold value for judgment, such as a maximum pressure value or a displacement end point value, and cannot effectively capture complex dynamic behavior characteristics in the crimping process. In an actual production environment, sensor data is easy to drift under the interference of environmental temperature and humidity fluctuation, mechanical vibration of equipment and the like, so that the misjudgment rate is increased. At present, a quality judgment scheme capable of fusing multidimensional dynamic parameters and self-adapting environmental changes in real time is lacking, and the improvement of the intelligent level of the production line is restricted. Disclosure of Invention The invention provides a method and a system for detecting crimping quality in real time, which can effectively solve the problems pointed out in the background technology. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: a real-time detection method of crimping quality is applied to a servo crimping system of a connector and a PCB, and comprises the following steps: collecting pressure sensor data and displacement encoder data in the crimping process in real time, and generating a pressure-displacement curve; Extracting characteristic parameters of the pressure-displacement curve, wherein the characteristic parameters comprise the pressure fluctuation amplitude of a contact stage, the local slope change value of a main pressure interval and the high-frequency oscillation frequency of a pressure maintaining stage; Synchronously normalizing the pressure fluctuation amplitude, the local slope change value and the high-frequency oscillation frequency into dimensionless risk values A, S and F; The weight coefficients W 1、W2 and W 3 are used for carrying out weighted fusion on A, S and F respectively, and the comprehensive mass fraction Q is calculated, wherein Q=W 1×(1-A)+W2×(1-S)+W3 × (1-F); and judging the crimping quality according to the comprehensive quality score. Further, the method also comprises a parameter sensitivity strengthening process, which comprises the following steps: Applying zero-mean Gaussian noise disturbance to the original pressure-displacement curve to generate N groups of disturbance curves, wherein N is more than or equal to 100; Calculating the variation coefficient of each characteristic parameter; And strengthening the corresponding weight coefficient according to the sensitivity of each characteristic parameter. Further, the weight coefficients W 1、W2 and W 3 implement dynamic adjustment, including: monitoring defect type distribution of a plurality of continuous PCBs; And when the proportion of the similar defects exceeds the set proportion, the weight coefficient of the corresponding parameters of the defects is improved. Further, the step of determining the pressure fluctuation amplitude of the contact stage includes: identifying a displacement point with the pressure continuously greater than the initial contact pressure threshold as a contact starting point; And calculating the standard deviation of the pressure in a preset displacement interval after the contact starting point, and taking the standard deviation as the pressure fluctuation amplitude. Further, the determining step of the local slope change value of the main pressure interval includes: calculating the pressure change rate in real time by using a sliding time window or a displacement window when the displacement is in the main pressure stroke interval; And recording the absolute value peak value of the pressure change rate between adjacent windows as the local slope change value. Further, the step of determining the high-frequency oscillation frequency of the dwell stage includes: Carrying out frequency domain transformation analysis on pressure data in the later stage of the pressure maintaining stage; Calculating the ratio of the energy of a preset risk frequency band to the energy of a full frequency band, and taking the ratio as a quantized value of the high-frequency oscillation frequency; The preset risk frequency band is set based on resonance characteristics of the PCB base material. Further, the frequency band distribution of the high-frequency oscillation frequency in the pressure maintaining stage is monitored in real time; if the high-frequency energy is detected to migrate to the low frequency, the weight coefficient W 1 of the pressure flu