An automatic adjustment method for sampling rates on an intelligent manufacturing production line

Abstract

The present invention relates to a method for automatically adjusting the sampling rate on an intelligent manufacturing production line. First, an observation window is introduced and the probability of product failures occurring on the production line within this window is statistically analyzed. After time translation and smoothing processing, the probability distribution of product failures occurring within the next observation window is estimated based on the observed data within this observation window. Then, a fault tolerance probability is set according to the requirements of the production line and a random threshold under this fault tolerance probability is calculated. If the failure rate of the next observation window is less than this threshold, it indicates that the sampling rate is set too high, and the sampling rate is reduced by a specific adjustment step. Conversely, if the failure rate is greater than this threshold, it indicates that the sampling rate is set too low, and the sampling rate is increased by the adjustment step. Using the present invention can better reflect the randomness of fault occurrences on the intelligent manufacturing production line, scientifically and reasonably adjust the sampling rate according to the fault tolerance probability, which can not only minimize the impact of small-probability product failure events on the adjustment of the sampling rate, but also achieve a balance between detection efficiency and quality assurance in intelligent manufacturing.

Description

Technical Field

[0001] The present invention relates to the field of intelligent manufacturing, and particularly to an automatic adjustment method for the sampling inspection rate on an intelligent manufacturing production line. Through this sampling inspection technology, a scientific and reasonable balance can be achieved between quality assurance and inspection efficiency. Background Art

[0002] In thousands of current intelligent factories, for a finished product to be put on the market, from individual parts to assembly and then to packaging and delivery, there is a particularly important process in between, namely the detection of product quality, which is also known as the lifeline in the field of intelligent manufacturing. To ensure product quality, workers with rich experience and high professional skills need to be arranged for product inspection. Therefore, the inspection link occupies a large share of the labor cost in the manufacturing industry. To save labor costs, product sampling inspection can be carried out to improve inspection efficiency. Therefore, how to set a scientific and reasonable sampling inspection rate and balance between product quality assurance and inspection efficiency has become one of the key factors in reducing labor costs in intelligent manufacturing.

[0003] The current product sampling inspection rate is generally set based on empirical values, lacking scientific and reasonable numerical calculations, and once the sampling inspection rate is given, it remains unchanged. When the failure rate of the product increases significantly, if the product sampling inspection rate cannot be increased in a timely manner, some defective products will flow into the next link, resulting in product rework and re-inspection; the flow of unqualified products into the market will affect the enterprise's reputation and image. On the contrary, when the failure rate of the product remains at a relatively low level for a long time, if the sampling inspection rate is not reduced in a timely manner, the inspection efficiency will also be greatly reduced, causing waste of labor and time. Therefore, how to dynamically adjust the sampling inspection rate based on historical data on the intelligent manufacturing production line is crucial. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to overcome the deficiencies of low product inspection efficiency and labor waste on the existing intelligent manufacturing production line, and provide an automatic adjustment method for the sampling inspection rate on the intelligent manufacturing production line, which can improve the product inspection efficiency while ensuring the production quality of the product.

[0005] The technical solution of the present invention is as follows: An observation window is introduced and the probability of product failures on the intelligent manufacturing production line under this window is statistically analyzed. At the same time, time translation and smoothing processing are carried out, and the observation data of this observation window is used to estimate the probability distribution of product failures on the production line in the next observation window. Then, a fault tolerance probability is set according to the requirements of the production line, and a random threshold under this fault tolerance probability is calculated. If the failure rate in the next observation window is less than this threshold, it indicates that the sampling rate is set too high, and the sampling rate is reduced by a specific adjustment step. Otherwise, the sampling rate is set too low, and the adjustment step is increased on the sampling rate. The present invention scientifically and reasonably adjusts the sampling rate according to the fault tolerance probability, which can not only minimize the impact of small probability events on the adjustment of the sampling rate, but also achieve the balance between detection efficiency and quality assurance in intelligent manufacturing.

[0006] Adopting the present invention can better reflect the randomness of fault occurrence on the intelligent manufacturing production line. By scientifically and reasonably adjusting the sampling rate according to the fault tolerance probability, it can not only minimize the impact of small probability fault events on the adjustment of the sampling rate, but also achieve the balance between detection efficiency and quality assurance in intelligent manufacturing.

[0007] In the first aspect, the specific implementation of the sampling rate automatic adjustment scheme of the present invention is as follows:

[0008] Step 1: Set the fault tolerance probability P according to the actual requirements of the intelligent manufacturing production line Q , that is, the probability of allowing misjudgment, generally a relatively small parameter;

[0009] Step 2: Set the length n of the observation window. Within the observation window [t, t + n), whether the products at the t, t + 1,..., t + n positions on the production line fail is represented by the random variable sequence ξ(t), ξ(t + 1),..., ξ(t + n), where ξ(t) = 1 indicates that the t-th product fails, and ξ(t) = 0 indicates that the t-th product does not fail. Similarly, ξ(t + 1) = 1 indicates that the (t + 1)-th product fails, and ξ(t + 1) = 0 indicates that the (t + 1)-th product does not fail, and so on. The random variable sequence is independent and identically distributed and follows a Bernoulli distribution, with the probability distribution being π(t); noting that the probability of fault occurrence varies within different observation windows, so the probability distribution π(t) will change with different observation windows;

[0010] Step 3: Initialize the sampling rate a of the intelligent manufacturing production line, set the sampling rate adjustment step Δ (generally one percent of the sampling rate a, that is ), the smoothing factor ρ (generally taking 0.95 - 0.98), initialize the number l of the observation window (l = 1) and the probability distribution π(t) (generally taking a small positive number between (0, 1));

[0011] Step 4: In the l-th observation window [t + (l - 1)n, t + ln), whether each product fails within this window is represented by random variables ξ(t + (l - 1)n), ξ(t + (l - 1)n + 1), ξ(t + ln). The total number of product failures m(t + (l - 1)n) within this window is counted, where m(t + (l - 1)n) = ξ(t + (l - 1)n) + ξ(t + (l - 1)n + 1) + … + ξ(t + ln);

[0012] Step 5: Obtain the rate function based on the probability π(t + (l - 1)n) of the current observation window and the to-be-determined threshold c

[0013] Step 6: Use e -nv(c) = P Q to inversely deduce the value of c, that is, the theoretical threshold of the failure rate within the l-th observation window [t + (l - 1)n, t + ln);

[0014] Step 7: Compare to check if it holds. If it indicates that the sampling rate setting is too high, and the sampling rate is adjusted downwards to a → a - Δ. If it indicates that the sampling rate setting is too low, and the sampling rate is adjusted upwards to a → a + Δ;

[0015] Step 8: Perform smoothing processing by integrating historical data and the current failure rate situation, and update to obtain the probability of failures within the next observation window [t + ln, t + (l + 1)n);

[0016] Step 9: Enter the next observation window, that is, l = l + 1, and return to Step 4 to continue the dynamic adjustment of the sampling rate until all products are inspected.

[0017] In a second aspect, the present invention provides an electronic device, including a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete communication with each other through the communication bus;

[0018] The memory is used to store a computer program;

[0019] The processor is used to execute the computer program stored on the memory, and when executed, it implements the above method for automatically adjusting the sampling rate.

[0020] In a third aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the method for automatically adjusting the sampling rate of the present invention.

[0021] The advantages of the present invention compared with the prior art are as follows:

[0022] (1) The present invention discloses a method for automatically adjusting the sampling rate to achieve a balance between production quality and detection efficiency on an intelligent manufacturing production line. Based on an observation window and feedback parameters, the present invention gives a discriminant condition for whether the sampling rate setting on the intelligent manufacturing production line is appropriate in a random sense, and gives an automatic adjustment method for the sampling rate, so as to reduce manual participation, improve detection efficiency and save production costs.

[0023] (2) By adopting the sampling rate automatic adjustment method of the present invention, the sampling rate on the intelligent manufacturing production line can be scientifically and reasonably set according to the error tolerance probability, so as to achieve a balance between detection efficiency and quality assurance in intelligent manufacturing. Description of the Drawings

[0024] Figure 1 is a flowchart for implementing the method of the present invention. Detailed Embodiments

[0025] The present invention will be described in detail below with reference to the drawings and embodiments.

[0026] Before the products on the intelligent manufacturing production line leave the factory, they need to go through a series of inspections. If the inspection ratio is too low, undetected products will occur, making it impossible to guarantee product quality; if the inspection ratio is too high, the inspection efficiency of the production line will be greatly reduced, resulting in waste of time and labor costs. Therefore, it is crucial to set the sampling rate of products in a scientific and reasonable manner so that it can achieve a balance between inspection efficiency and quality assurance.

[0027] As Figure 1 shown, the method of the present invention is specifically implemented as follows:

[0028] Step 1: Set the error tolerance probability P Q , that is, the probability of allowing misjudgment (generally a small parameter, which can be set according to actual needs, such as about 0.0002, indicating that the probability of the failure occurrence rate being greater than the threshold does not exceed 0.0002, that is, two ten-thousandths).

[0029] Step 2: Set the length n of the observation window. Within the observation window [t, t + n), whether the t-th, (t + 1)-th,..., (t + n)-th products on the production line have failures is represented by the random variable sequence ξ(t), ξ(t + 1),..., ξ(t + n):

[0030]

[0031] Where ξ(t) = 1 indicates that the t-th product fails, and ξ(t) = 0 indicates that the t-th product does not fail. Similarly, ξ(t + 1) = 1 indicates that the (t + 1)-th product fails, and ξ(t + 1) = 0 indicates that the (t + 1)-th product does not fail, and so on. The sequence of random variables is independent and identically distributed and follows a Bernoulli distribution with a probability distribution of π(t); noting that the probability of failure occurrence varies within different observation windows, so the probability distribution π(t) changes with different observation windows;

[0032] Step 3: Initialize the sampling rate a of the intelligent manufacturing production line, set the sampling rate adjustment step Δ (generally 1% of the sampling rate a, i.e., ), the smoothing factor ρ (generally taken as 0.95 - 0.98), initialize the number l (l = 1) of the observation window and the probability distribution π(t) (generally taken as a small positive number between (0, 1));

[0033] Step 4: In the l-th observation window [t + (l - 1)n, t + ln), whether each product in this window fails is represented by the random variables ξ(t + (l - 1)n), ξ(t + (l - 1)n + 1), ξ(t + ln), and count the total number m(t + (l - 1)n) of product failures in this window, where m(t + (l - 1)n) = ξ(t + (l - 1)n) + ξ(t + (l - 1)n + 1) + … + ξ(t + ln);

[0034] Step 5: Obtain the rate function according to the probability π(t + (l - 1)n) of the current observation window and the to-be-determined threshold c

[0035] Step 6: Use e -nv(c) = P Q to inversely deduce the value of c, that is, the theoretical threshold of failure occurrence in the l-th observation window [t + (l - 1)n, t + ln) under the fault tolerance probability P Q ;

[0036] Step 7: Compare to see if it holds. If it means that the sampling rate is set too high, and lower the sampling rate to a ← a - Δ. If it means that the sampling rate is set too low, and increase the sampling rate to a ← a + Δ;

[0037] Step 8: Perform smoothing processing by integrating historical data and the current failure rate situation, and update to obtain the probability of failure occurrence in the next observation window [t + ln, t + (l + 1)n)

[0038] Step 9: Enter the next observation window, i.e., l = l + 1, and return to Step 4 to continue the dynamic adjustment of the sampling rate until all products are inspected.

[0039] Based on the above steps, by introducing the fault tolerance probability P Q and generating a random threshold c, it can better reflect the randomness of the occurrence of faults, avoid blindly adjusting the sampling rate due to the random occurrence of a small probability event, and thus make the adjustment of the sampling rate more accurate. The mechanical judgment of whether to adjust the sampling rate is extended to a judgment based on the fault tolerance probability and the current failure rate, minimizing the impact of the occurrence of small probability fault events on the adjustment of the sampling rate.

[0040] In summary, the present invention can provide theoretical support for the setting and dynamic adjustment of the sampling rate of intelligent manufacturing production lines, and its function of adaptively adjusting the sampling rate can also help intelligent manufacturing companies save production costs.

[0041] Through the description of the above embodiments, those skilled in the art can clearly understand that the above embodiments can be implemented by software or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solutions of the above embodiments can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments of the present invention.

[0042] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention.

Claims

1. An automatic adjustment method for sampling rate on an intelligent manufacturing production line, characterized in that: Step 1: Set the fault tolerance probability according to the actual requirements of the intelligent manufacturing production line , that is, the probability of allowing misjudgment to occur Step 2: Set the length of the observation window , within the observation window , whether the th product on the production line has a fault is represented by a sequence of random variables . Among them = 1 indicates that the th product has a fault, = 0 indicates that the th product has no fault; = 1 indicates that the th product has a fault, = 0 indicates that the th product has no fault, and so on; The sequence of random variables is independent and identically distributed and follows a Bernoulli distribution, with the probability distribution being ; the probability of a fault occurring within different observation windows varies, so the probability distribution will vary with different observation windows; Step 3: Initialize the sampling rate of the intelligent manufacturing production line , set the step size for adjusting the sampling rate , smoothing factor , initialize the number of the observation window and probability distribution ; Step 4: At the th observation window , whether each product in this window fails is represented by a random variable . The total number of product failures m(t+(l-1)n) in this window is counted, where m(t+(l-1)n) = ξ(t+(l-1)n) + ξ(t+(l-1)n+1) + … + ξ(t+ln); Step 5: According to the probability distribution of the l-th observation window and the to-be-determined threshold to obtain the rate function ; The probability distributions of different observation windows are different, and the here corresponds to the probability distribution of the l-th observation window ; The probability distribution ; Step 6: Use to inversely deduce the value, that is, the theoretical threshold of the fault occurrence rate within the observation window ; Step 7: Compare to check if it holds; if , it indicates that the sampling rate is set too high. Lower the sampling rate to . If , it indicates that the sampling rate is set too low. Raise the sampling rate to ; Step 8: Perform smoothing processing by integrating historical data and the current failure rate situation, and update to obtain the next observation window Probability of internal failure ; Step 9: Enter the next observation window, i.e., , return to Step 4 to continue the dynamic adjustment of the sampling rate until all products are inspected.

2. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete mutual communication through the communication bus; The memory is used to store computer programs; The processor is used to execute the computer program stored on the memory, and when executed, it implements the method described in claim 1.

3. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method described in claim 1.

Images