A performance monitoring method for reflow oven temperature control system based on data drive

By establishing a mathematical model of monitoring indicators through a data-driven approach, the problems of false alarms and missed alarms in the performance monitoring of the reflow oven temperature control system are solved, real-time detection of system degradation is achieved, and the quality of circuit board welding is improved.

CN114297906BActive Publication Date: 2025-10-03四川启睿克科技有限公司
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202111571751.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-21
Publication Date
2025-10-03
Estimated Expiration
2041-12-21

AI Technical Summary

Technical Problem

Existing performance monitoring methods for reflow oven temperature control systems suffer from false alarms and missed alarms, making it difficult to accurately monitor system degradation in real time, thus affecting the soldering quality of circuit boards.

Method used

A data-driven approach is adopted to establish a mathematical model of monitoring indicators, determine the parameter b, set the threshold using confidence intervals, and judge the system status in combination with decision logic to achieve performance monitoring of the reflow oven temperature control system.

Benefits of technology

It achieves fast and accurate system performance monitoring, improves the stability of circuit board welding quality, and reduces the risk of false alarms and missed alarms.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114297906B_ABST
    Figure CN114297906B_ABST
Patent Text Reader

Abstract

The present invention discloses a performance monitoring method for a reflow oven temperature control system based on data drive, comprising: establishing a mathematical model of a monitoring index f; determining a parameter b based on historical data; determining a threshold J based on a confidence interval; th,z Based on the obtained temperature values, a real-time indicator f(z) is calculated, and decision logic is used to determine whether the system is in a healthy or degraded state. This invention, based on historical temperature data from a reflow oven temperature control system and the characteristics of the data, establishes a mathematical model capable of rapidly monitoring system performance. This model effectively addresses the performance monitoring issues of reflow oven temperature control systems, as well as the shortcomings of traditional control system difference measurement methods and ν-gap measurement methods. It also offers high computational efficiency and real-time performance.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of reflow oven temperature control, and in particular to a performance monitoring method of a reflow oven temperature control system based on data driving. Background Art

[0002] The temperature control system of the reflow oven is the core part of the SMT production line. The performance of the temperature control system directly affects the product process temperature curve, and thus affects the soldering quality of the circuit board. Maintaining the temperature of each temperature zone accurately and constantly at the target value is the main indicator of the temperature control system, such as Figure 1 As shown, where A (A>0) is a constant.

[0003] However, as the system degrades, the temperature control system's control effect on the temperature of each temperature zone will deteriorate, resulting in large fluctuations in the temperature of each temperature zone around the target temperature value, such as Figure 2 As shown, B (B>A) is a constant. This has a significant impact on the soldering quality of the PCB. If the performance of the temperature control system can be accurately monitored, system degradation can be detected in a timely manner to prevent the degradation of PCB soldering quality.

[0004] Although several performance monitoring methods for temperature control systems exist in practical applications, such as empirical thresholds, traditional control system difference metrics (such as the infinite norm), and ν-gap metrics, these methods still suffer from false positives and negative alerts. In empirical threshold methods, the threshold setting significantly impacts the results, easily leading to false positives, negative alerts, and implementation issues. Although traditional control system difference metrics monitor control system performance by directly evaluating closed-loop system performance, their impact on closed-loop system performance is minimal. Furthermore, in actual engineering, due to internal factors such as component aging, it is difficult to accurately identify the system model through data. Furthermore, system identification methods may not always achieve real-time identification, or may even be impossible. Therefore, the ν-gap metric approach presents certain challenges. Summary of the Invention

[0005] The purpose of the present invention is to provide a performance monitoring method for a reflow oven temperature control system based on data drive, in order to solve the technical problems existing in the background technology.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A data-driven performance monitoring method for a reflow oven temperature control system, comprising:

[0008] Establish a mathematical model for monitoring indicator f;

[0009] Determine parameter b based on historical data;

[0010] Determine the threshold J based on the confidence interval th,z ;

[0011] The real-time indicator f(z) is calculated based on the obtained temperature value, and the decision logic is used to determine whether the system is in a healthy state or a degraded state.

[0012] The mathematical model for establishing the monitoring index f includes:

[0013]

[0014] Where b is a positive integer, then f(z) is The hyperbola is centered at , f(z) is the indicator for monitoring system performance, z(t) is the actual temperature, and z0 is the ideal temperature.

[0015] The method for determining the parameter b includes:

[0016] Select a section of healthy data and degradation data that meets the preset conditions from the system's historical temperature data, and establish a function e(b) that is only related to b using formula (4):

[0017] e(b)=min|Q i -μ Q |-max|Q i -μ T |+max|T i -μ Q |-min|T i -μ T | (4)

[0018] Where Q i and T i are the health indicators of the i-th healthy data and degradation data, μ Q and μ T are the average values ​​of Q and T respectively. min|Q i -μ Q | represents the minimum distance from Q to the average value of Q, max|Q i -μ T | represents the maximum distance from the average value in Q to T, max|T i -μ Q | represents the maximum distance from the average value of Q in T, max|Q i -μ T | represents the minimum distance from T to the average value of T; calculate the positive integer that makes e(b)>0 and e(b) minimum.

[0019] The threshold J is determined based on the confidence interval th,z ,include:

[0020] Get the index f h The mean and variance of

[0021] Get the index f h The formulas for the mean and variance of are:

[0022]

[0023] Where M1 is the number of z; fh is the health index obtained through health data;

[0024] According to f h The characteristics of , need to set an upper limit, so the threshold is

[0025] J th,z =μ z +β z ×σ z (7)

[0026] Among them, β z It is a constant determined based on actual engineering data.

[0027] The calculation of the real-time indicator f(z) based on the obtained temperature value and the use of decision logic to determine whether the system is in a healthy state or a degraded state include:

[0028] Whenever a new temperature value is obtained, the real-time index f(z) is obtained by formula (2);

[0029] The decision logic is:

[0030]

[0031] Among them, f(z) is a real-time indicator.

[0032] The beneficial effects that may be brought about by the data-driven performance monitoring method of a reflow oven temperature control system disclosed in this application include but are not limited to:

[0033] Based on the historical temperature data of the reflow oven temperature control system and the characteristics of the data, a mathematical model that can quickly monitor the system performance is established. This model effectively solves the performance monitoring problem of the reflow oven temperature control system and the shortcomings of the traditional control system difference measurement method and ν-gap measurement method. It also has high computational efficiency and real-time performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 The temperature curve of the temperature control system in the healthy stage is shown in Figure 1. The ordinate is the temperature and the abscissa is the sampling time. At this time, the temperature oscillates between [target temperature - A - α, target temperature + A + α], where α (α > 0) is the temperature change caused by the disturbance in the healthy stage.

[0035] Figure 2 The temperature curve of the temperature control system during the degradation phase is plotted on the y-axis, with the temperature plotted on the x-axis and the sampling time plotted on the x-axis. The temperature oscillates between [target temperature - B - ε, target temperature + B + ε], where ε (ε > 0) is the temperature change caused by the disturbance during the degradation phase.

[0036] Figure 3 The hyperbola graph of actual temperature z(t) and index g. g(z) is a hyperbola centered at (z0,1). According to the fluctuation range of z(t), g(z) fluctuates between positive and negative, and when z(t) = z 0- When g(z)→+∞; when z(t)=z 0+ When , g(z)→-∞. These are not conducive to monitoring degradation.

[0037] Figure 4 The adjusted hyperbola graph is shown in the figure. In the figure, [H1, H2] is the healthy interval, [0, F1] ∪ [F2, +∞] is the fault interval, and [F1, H1] ∪ [H2, F2] is the degradation interval. This way, the f corresponding to all fault data can be f are greater than the f corresponding to the health data h , and only need to set the f corresponding to the health data h The upper limit can be set.

[0038] Figure 5 Flowchart of the present invention. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0040] On the contrary, this application covers any alternatives, modifications, equivalents, and solutions made within the spirit and scope of this application as defined by the claims. Furthermore, to facilitate a better understanding of this application, certain specific details are described in detail below in the detailed description of this application. Those skilled in the art will be able to fully understand this application without these details.

[0041] The following is a detailed description of a performance monitoring method for a reflow oven temperature control system based on data drive according to an embodiment of the present application. It should be noted that the following embodiments are only used to explain the present application and do not constitute a limitation of the present application.

[0042] Due to the effects of control performance degradation and external disturbances, the actual temperature z(t) fluctuates within a certain range [z0 - δ1, z0 + δ2] (0 < δ1 < z0, δ2 > 0), where δ1 and δ2 are constants and z0 is the ideal temperature. When the system is not degraded, the actual temperature z(t) is only affected by external disturbances, and the fluctuation of the actual temperature z(t) is small. However, when the system is degraded, the actual temperature z(t) is also affected by the degradation of control performance, and the fluctuation of the actual temperature z(t) is significantly increased.

[0043] Based on this, a mathematical model of the monitoring system performance index g(z) and the actual temperature z(t) is established through formula (1). Figure 3 It is a hyperbola graph of actual temperature z(t) and index g.

[0044]

[0045] Obviously, g(z) is a hyperbola centered at (z0,1). According to the fluctuation range of z(t), g(z) fluctuates between positive and negative, and when z(t)=z 0- When g(z)→+∞; when z(t)=z 0+ When g(z)→-∞. These are not conducive to monitoring degradation. To this end, this patent improves formula (1), and the expression of the monitoring index is:

[0046]

[0047] Where b is a positive integer. In this case, f is A hyperbola centered on Figure 4 In order to facilitate the monitoring of system performance, according to the characteristics of the data, let [H1,H2] be the healthy interval, [0,F1]∪[F2,+∞] be the fault interval, and [F1,H1]∪[H2,F2] be the degradation interval.

[0048]

[0049] This will make the f corresponding to all fault data f are greater than the f corresponding to the health data h , and you only need to set the upper limit of the monitoring indicators corresponding to the health data.

[0050] In actual projects, due to severe data imbalance, that is, there are fewer fault data and more healthy data, it is difficult to obtain a more accurate F1. Therefore, a longer period of healthy data and a smaller amount of degradation data are selected from the system's historical temperature data. Then, Q corresponding to the healthy data and T corresponding to the degradation data are calculated according to formula (2).

[0051] Then, according to the maximum and minimum distance concept, a function e(b) related only to b is established through formula (4).

[0052] e(b)=min|Q i -μ Q |-max|Q i -μ T |+max|T i -μ Q |-min|T i -μ T | (4)

[0053] Where Q i and T i are the health indicators of the i-th healthy data and degradation data, μ Q and μ T are the average values ​​of Q and T respectively. min|Q i -μ Q | represents the minimum distance from Q to the average value of Q, max|Q i -μ T | represents the maximum distance from the average value in Q to T, max|T i -μ Q | represents the maximum distance from the average value of Q in T, max|Q i -μ T | represents the minimum distance from T to the mean of T.

[0054] Finally, the parameter b that makes e(b)>0 and minimizes e(b) is calculated through an intelligent optimization algorithm, such as a particle swarm optimization algorithm.

[0055] (3) Threshold setting

[0056] Since z is easily affected by external interference, if a smaller threshold is selected, it may lead to false positives. If a larger threshold is selected, it may lead to missed negatives. Therefore, this patent adopts the idea of ​​confidence interval in statistics to design the threshold. h The formulas for the mean and variance of are:

[0057]

[0058] Where M1 is the number of z; fh is the health indicator obtained through health data; formula (5) is used to obtain the average value of fh; and formula (6) is the standard deviation of fh.

[0059] Depend on Figure 4 It can be seen that the healthy interval is monotonic and is smaller than the fault interval. Therefore, according to f h In this section, we only need to set an upper limit. Therefore, the threshold is

[0060] J th,z =μ z +β z ×σ z (7)

[0061] Among them, β z It is a constant that can be determined based on actual engineering data.

[0062] Therefore, the decision logic is:

[0063]

[0064] This embodiment also provides a specific implementation method:

[0065] like Figure 5 As shown, a performance monitoring method of a reflow oven temperature control system based on data drive includes the following steps:

[0066] Step 1: Establish a mathematical model for monitoring index f.

[0067] First, a mathematical model of the monitoring system performance index g(z) and the actual temperature z(t) is established through formula (1).

[0068]

[0069] Then, considering that the mathematical model of formula (1) is not conducive to monitoring degradation, this patent improves it and obtains the expression of the monitoring index as follows:

[0070]

[0071] Where b is a positive integer. In this case, f is A hyperbola centered on Figure 2 shown.

[0072] Finally, in order to facilitate the monitoring of system performance, according to the characteristics of the data, let [H1,H2] be the healthy interval, [0,F1]∪[F2,+∞] be the fault interval, and [F1,H1]∪[H2,F2] be the degradation interval.

[0073]

[0074] Step 2: Based on historical data and the maximum and minimum distance concept, the intelligent optimization algorithm is used to determine the parameter b;

[0075] A long period of healthy data and a small amount of degradation data are selected from the historical temperature data of the system, and a function e(b) that is only related to b is established through formula (4).

[0076] e(b)=min|Q i-μ Q |-max|Q i -μ T |+max|T i -μ Q |-min|T i -μ T | (4)

[0077] The parameter b that makes e(b)>0 and minimizes e(b) is calculated through an intelligent optimization algorithm, such as a particle swarm optimization algorithm.

[0078] Step 3 Determine the threshold J based on the confidence interval th,z .

[0079] Since z is easily affected by external interference, if a smaller threshold is selected, it may lead to false positives. If a larger threshold is selected, it may lead to missed positives. Therefore, this patent adopts the idea of ​​confidence interval in statistics to design the threshold. h The mean and variance of are:

[0080]

[0081] Where M1 is the number of z.

[0082] Depend on Figure 4 It can be seen that the healthy interval is monotonic and is smaller than the fault interval. Therefore, according to f h In this section, we only need to set an upper limit. Therefore, the threshold is

[0083] J th,z =μ z +β z ×σ z (7)

[0084] Among them, β z It is a constant that can be determined based on actual engineering data.

[0085] Step 4: Whenever a temperature value z(t) is obtained, the real-time index f(z) is calculated according to formula (2).

[0086]

[0087] Step 5: Determine whether the system is in a healthy state or a degraded state according to formula (8).

[0088] The decision logic is:

[0089]

[0090] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A performance monitoring method for a reflow oven temperature control system based on data drive, characterized in that: include: Establish a mathematical model for monitoring indicator f; Determine parameter b based on historical data; Determine the threshold J based on the confidence interval th,z ; Calculate the real-time indicator f(z) based on the obtained temperature value, and use decision logic to determine whether the system is in a healthy state or a degraded state; The mathematical model for establishing the monitoring index f includes: Where b is a positive integer, then f(z) is The hyperbola is centered at , f(z) is the indicator for monitoring system performance, z(t) is the actual temperature, and z0 is the ideal temperature.

2. The performance monitoring method of the reflow oven temperature control system based on data drive according to claim 1, characterized in that: The method for determining the parameter b includes: Select a section of healthy data and degradation data that meets the preset conditions from the system's historical temperature data, and establish a function e(b) that is only related to b using formula (4): e(b)=min|Q i -m Q |-max|Q i -m T |+max|T i -m Q |-min|T i -m T | (4) Where Q i and T i are the health indicators of the i-th healthy data and degradation data, μ Q and μ T are the average values ​​of Q and T respectively; min|Q i -μ Q | represents the minimum distance from Q to the average value of Q, max|Q i -μ T | represents the maximum distance from the average value in Q to T, max|T i -μ Q | represents the maximum distance from the average value of Q in T, max|Q i -μ T | represents the minimum distance from T to the average value of T; calculate the positive integer that makes e(b)>0 and e(b) minimum.

3. The performance monitoring method of the reflow oven temperature control system according to claim 1, characterized in that: The threshold J is determined based on the confidence interval th,z ,include: Get the index f h The mean and variance of Get the index f h The formulas for the mean and variance of are: Where M1 is the number of z; f h It is a health indicator obtained through health data; According to f h The characteristics of , need to set an upper limit, so the threshold is J th,z =μ z +b z ×s z (7) Among them, β z It is a constant determined based on actual engineering data.

4. The performance monitoring method of the reflow oven temperature control system according to claim 3, characterized in that: The calculation of the real-time indicator f(z) based on the obtained temperature value and the use of decision logic to determine whether the system is in a healthy state or a degraded state include: Whenever a new temperature value is obtained, the real-time index f(z) is obtained by formula (2); The decision logic is: Among them, f(z) is a real-time indicator.

Citation Information

Patent Citations

  • Temperature-curve-characteristic-based fault diagnosis method of satellite temperature control system

    CN107942994A

  • Heating control method for glass melting furnace

    CN108147647A