Intelligent control system and method for chip processing equipment

Abstract

The invention discloses an intelligent control system and method for chip processing equipment, and relates to the technical field of equipment control, and the control method comprises the following steps: constructing an equipment control data management system to generate a corresponding control record; performing control step division and feature extraction on any control record; randomly selecting one control record, and performing first feature evaluation on the control step; analyzing the presented state of the chip in the control step, and performing second feature evaluation on the control step; extracting abnormal features based on the difference between the two evaluation results, and analyzing the influence degree of each abnormal feature; an anomaly recognition model is constructed for anomaly recognition, and a recognition standard for judging whether anomaly exists in any control step or not is set; whenever the chip is processed, abnormal feature extraction is carried out, abnormal recognition is carried out through the abnormal recognition model, and whether abnormal reminding is carried out or not is judged.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment control, and in particular to an intelligent control system and method for chip processing equipment. Background Art

[0002] Chip processing equipment refers to the machinery and tools used to perform various process steps in the semiconductor manufacturing process. It is a vital component of the semiconductor manufacturing process. The chip manufacturing process usually includes multiple steps, each of which requires specific equipment to complete. The equipment used in each step needs to have extremely high precision and complexity to meet the strict requirements of modern chip manufacturing for size and performance.

[0003] Since chip manufacturing involves multiple complex process steps, each step may introduce different types of anomalies. Traditional chip processing equipment lacks real-time monitoring capabilities and cannot detect anomalies immediately during the processing process. This will cause the problem to be amplified in subsequent steps. When the anomaly is finally discovered, it is necessary to trace back the occurrence of the problem step by step, increasing the difficulty and cost of repair. Summary of the invention

[0004] The purpose of the present invention is to provide an intelligent control system and method for chip processing equipment to solve the problems raised in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solution: an intelligent control method for chip processing equipment, the control method comprising the following steps:

[0006] Step S100: constructing an equipment control data management system to record each processing process and generate corresponding control records; dividing any control record into control steps, extracting features of each control step, and analyzing the completion status of the control record;

[0007] Step S200: randomly selecting a manipulation record, extracting several features of any manipulation step in the manipulation record, and performing a first feature evaluation on the manipulation step based on the several features; analyzing the state of the chip in the manipulation step, and performing a second feature evaluation on the manipulation step;

[0008] Step S300: Based on the difference between the two evaluation results, the abnormal features of any manipulation step are extracted, and the influence of each abnormal feature is analyzed; an abnormal recognition model is constructed to perform abnormal recognition on any manipulation record, and based on the completion status of the manipulation record, a recognition standard for determining whether any manipulation step has an abnormality is set;

[0009] Step S400: Whenever the chip is processed, the current operation step of the chip is evaluated; based on the evaluation result, abnormal features are extracted, and the abnormality of the current operation step of the chip is identified through the abnormality identification model, and it is determined whether to issue an abnormal reminder.

[0010] Furthermore, step S100 includes the following steps:

[0011] Step S101: Whenever a chip is placed on any chip processing equipment for processing, monitoring data monitored in the chip processing equipment is obtained; all chip processing equipment involved in any control record are obtained, and a control step of the control record is set for each chip processing equipment;

[0012] Step S102: setting the monitoring data monitored in any chip processing equipment as processing data of the corresponding step; arbitrarily selecting a control step, presetting a number of operation types and data types of each operation type for the control step, dividing the processing data according to the data type, and matching it with each operation type to obtain a number of operation types of the control step, setting any matched operation type as a feature of the control step, and generating a feature set of the control step;

[0013] Step S103: preset a fixed number and fixed order of expected manipulation steps for the complete chip processing process, and preset a number of expected features for each expected manipulation step; arbitrarily select the i-th manipulation step from the manipulation record, and count the number n of the same features between the i-th manipulation step and the j-th expected manipulation step ij , set a feature number threshold n max， If n is satisfied ij >n max And n ij =Max(n ia ), then the i-th control step is matched with the j-th expected control step, where a∈(1,b), Max(n ia ) is the maximum number of identical features between the ith manipulation step and each expected manipulation step, and b is the expected number of manipulation steps; the number of successfully matched manipulation steps in the manipulation record is counted m, and the degree of completion of the manipulation record is obtained as C=m / m total , where m total is the total number of expected control steps; the control steps are matched with the expected control steps in order to determine whether the chip has any missed operations, wrong operations, etc. during the actual processing, or even whether the processing is interrupted due to serious chip abnormalities. The completion degree can directly reflect the abnormal situation of each control record.

[0014] Further, step S200 includes the following steps:

[0015] Step S201: arbitrarily selecting an i-th manipulation step from any manipulation record, obtaining an expected manipulation step matching the i-th manipulation step, and presetting corresponding evaluation rules for each expected feature in the expected manipulation step;

[0016] Step S202: extract the feature set of the i-th manipulation step, obtain the processing data of any feature in the manipulation record, obtain the feature value presented by the feature in the manipulation record based on the preset evaluation rule, and calculate the average value of the feature values ​​presented by all features to obtain the first feature value Z1 of the i-th manipulation step i The first eigenvalue is an evaluation of the accuracy of the control step, which is used to compare with the actual accuracy evaluation results of the chip to analyze the abnormal impact between the control steps;

[0017] Step S203: Obtain the expected processing time of the expected manipulation step as T ex , the preset chip real-time completion evaluation rule evaluates the completion status of the chip at any time during the processing, and sets the completion feature value at time t1 in the expected control step as W(t1), according to the formula:

[0018]

[0019] Calculate the total completion characteristic value CW of the control step within the expected processing time ex ; The definite integral calculation results in the accumulation of completion, which can reflect the total amount of work within the processing time;

[0020] Step S204: Obtain the change of the actual completion status of the chip in the i-th control step over time, set the actual completion characteristic value at time t2 in the i-th control step as W(t2), according to the formula:

[0021]

[0022] Among them, T ac is the actual processing time of the i-th manipulation step; the second eigenvalue Z2 of the i-th manipulation step is calculated i The second eigenvalue is used to feedback the actual chip processing accuracy evaluation results, which is conducive to the subsequent identification of abnormal features.

[0023] Further, step S300 includes the following steps:

[0024] Step S301: randomly select a manipulation record and obtain the first characteristic value Z1 of the i-th manipulation step in the manipulation record i, the first characteristic average value from the start of processing to the completion of the i-th manipulation step in the manipulation record is obtained as Where e is a positive integer and e∈(1,i), Z1 e is the first eigenvalue of the e-th manipulation step; the eigenvalue presented by the j-th feature in the e-th manipulation step is tz ej , if tz ej <(Z1 i ) ave , then the j-th feature is set as an abnormal feature; the first feature average value is based on the fact that the completion degrees of each manipulation step are independent of each other and will not affect each other. By making this assumption and combining it with the subsequent correlation analysis, the impact of each abnormal feature can be accurately obtained;

[0025] Step S302: Multiply the first characteristic values ​​of each manipulation step in the manipulation record from the start of processing to the completion of the i-th manipulation step to obtain the first characteristic comprehensive value from the start of processing to the completion of the i-th manipulation step The first characteristic comprehensive value reflects the interdependence between various control steps;

[0026] Step S303: Obtain the first characteristic difference value ΔZ1 from the start of processing to the completion of the i-th manipulation step in the manipulation record i =(Z1 i ) ave -(Z1 i ) cv , all abnormal features in the manipulation record from the start of processing to the completion of the i-th manipulation step are summarized to obtain an abnormal feature set, and the k-th abnormal feature is selected from the abnormal feature set, according to the formula:

[0027]

[0028] Among them, tz k is the characteristic value presented by the kth abnormal feature, g i is the number of abnormal features in the abnormal feature set; calculate the influence degree Y of the kth abnormal feature in the manipulation record up to the i-th manipulation step k ;

[0029] Step S304: Obtain the second characteristic value Z2 of the i-th manipulation step i , it is calculated that the influence of the remaining factors except for each abnormal feature in the control record up to the i-th control step is ΔZ2 i =(Z1 i ) cv -Z2 i; Obtain the influence amplitude of the remaining factors up to the i-th control step in each control record, and calculate the average value to obtain the average influence amplitude ΔZ2(i);

[0030] Step S305: Set the completion degree of the control record to C. If C < 1 and i = C × m total , where m total is the total number of expected manipulation steps, then the first feature value presented in any manipulation step from the start of processing to the completion of the i-th manipulation step in the manipulation record and the influence degree of each abnormal feature included are obtained to construct an abnormality recognition model:

[0031]

[0032] Where e is a positive integer and e∈(1,i), Z1 e is the first eigenvalue of the e-th manipulation step, k is a positive integer and k∈(1,g e ), g e is the number of abnormal features contained from the beginning of processing to the e-th manipulation step, Y k is the influence degree of the kth abnormal feature included, ΔZ2(e) is the influence amplitude of the remaining factors except each abnormal feature from the start of processing to the eth manipulation step; the abnormal value F of the i-th manipulation step of the manipulation record is obtained by the abnormal recognition model i ;

[0033] Step S306: extract each control record with a completion degree C < 1, obtain the abnormal value of the last control step in each extracted control record, and select the abnormal value with the smallest value as the abnormal threshold for determining whether any control step has an abnormality.

[0034] Furthermore, step S400 includes the following steps:

[0035] Step S401: Get the current step of the chip being processed, and set the current step to the i-th step. ’ The i-th control step is obtained by using the preset evaluation rules. ’ The first characteristic values ​​of each manipulation step before the first manipulation step are averaged to obtain a first characteristic average value; and abnormal features satisfying that the characteristic value is less than the first characteristic average value are extracted;

[0036] Step S402: Obtain each abnormal feature in the i-th ’ The impact degree obtained from the feedback of the i-th control step is input into the abnormality recognition model, and the output is the current state of the chip in the i-th control step. ’ The abnormal value F presented in the control step; set the abnormal threshold to F th, if F>F th , an abnormal reminder is sent, in the i ’ Stop processing when one control step is reached.

[0037] In order to better implement the above method, an intelligent control system for chip processing equipment is also proposed. The control system includes a control record analysis module, a control step evaluation module, an abnormal feature processing module and a control abnormality recognition module;

[0038] The control record analysis module is used to build an equipment control data management system to record each processing process and generate corresponding control records; divide any control record into steps, extract features for each step, and analyze the completion status of the control record;

[0039] A manipulation step evaluation module is used to select a manipulation record at random, extract several features of any step in the manipulation record, and perform a first feature evaluation on the step based on the several features; analyze the state presented by the chip in the step, and perform a second feature evaluation on the step;

[0040] The abnormal feature processing module is used to extract the abnormal features of any operation step based on the difference between the two evaluation results, and analyze the influence of each abnormal feature; build an abnormal recognition model to identify the abnormality of any operation record, and set the recognition standard for judging whether any operation step has an abnormality based on the completion status of the operation record;

[0041] The control anomaly recognition module is used to evaluate the current control step of the chip whenever the chip is processed; extract abnormal features based on the evaluation results, and use the abnormal recognition model to identify the abnormality of the current control step of the chip, and determine whether to issue an abnormal reminder.

[0042] Further, the control record analysis module includes a control record generation unit and a control step analysis unit;

[0043] The control record generation unit is used to build an equipment control data management system to record each processing process and generate corresponding control records; the control step analysis unit is used to divide any control record into steps, extract features of each step, and analyze the completion status of the control record.

[0044] Further, the manipulation step evaluation module includes a first feature evaluation unit and a second feature evaluation unit;

[0045] The first feature evaluation unit is used to arbitrarily select a manipulation record, extract several features of any step in the manipulation record, and perform a first feature evaluation on the step based on the several features; the second feature evaluation unit is used to analyze the state of the chip in the step, and perform a second feature evaluation on the step.

[0046] Further, the abnormal feature processing module includes an abnormal feature extraction unit and an abnormal identification processing unit;

[0047] The abnormal feature extraction unit is used to extract the abnormal features of any operation step based on the difference between the two evaluation results, and analyze the influence of each abnormal feature; the abnormal identification processing unit is used to construct an abnormal identification model to identify the abnormality of any operation record, and set the identification standard for judging whether any operation step has an abnormality based on the completion status of the operation record.

[0048] Furthermore, the control anomaly identification module includes a real-time step evaluation unit and an anomaly identification and judgment unit;

[0049] The real-time step evaluation unit is used to evaluate the current operation step of the chip whenever the chip is processed; the abnormality identification and judgment unit is used to extract abnormal features based on the evaluation results, and to identify the abnormality of the current operation step of the chip through the abnormality identification model, and to determine whether to issue an abnormal reminder.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1. The present invention can help workers to find abnormal chips in time during the processing process by analyzing abnormal conditions in real time during the processing of chips, so as to avoid finding problems in chips after production is completed, which will increase the difficulty of repair and maintenance, and can greatly improve the chip processing efficiency;

[0052] 2. The present invention can effectively analyze the abnormal characteristics that affect the evaluation results by comparing the step accuracy evaluation results in each control link and the chip processing result accuracy evaluation results, which is beneficial to the subsequent calculation of abnormal values ​​of chip processing and accurately identifying abnormalities in the chip processing process;

[0053] 3. The present invention analyzes the abnormal situation of the current operation step of the chip by constructing an abnormality recognition model, helping the staff to capture the abnormal situation in real time and take corresponding measures, reducing losses and minimizing the difficulty of repair. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A schematic diagram of the steps of an intelligent control method for chip processing equipment;

[0055] Figure 2 The figure is a schematic diagram of the structure of an intelligent control system for chip processing equipment. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] Example: Figure 1 to Figure 2 As shown, the present invention provides an intelligent control method for chip processing equipment, and the control method comprises the following steps:

[0058] Step S100: constructing an equipment control data management system to record each processing process and generate corresponding control records; dividing any control record into control steps, extracting features of each control step, and analyzing the completion status of the control record;

[0059] Wherein, step S100 includes the following steps:

[0060] Step S101: Whenever a chip is placed on any chip processing equipment for processing, monitoring data monitored in the chip processing equipment is obtained; all chip processing equipment involved in any control record are obtained, and a control step of the control record is set for each chip processing equipment;

[0061] Step S102: setting the monitoring data monitored in any chip processing equipment as processing data of the corresponding step; arbitrarily selecting a control step, presetting a number of operation types and data types of each operation type for the control step, dividing the processing data according to the data type, and matching it with each operation type to obtain a number of operation types of the control step, setting any matched operation type as a feature of the control step, and generating a feature set of the control step;

[0062] Example 1: A control record is set to include several steps such as wafer processing, oxidation, photolithography, etching, etc., and is divided based on different processing equipment, and subsequent feature extraction is performed;

[0063] Step S103: preset a fixed number and fixed order of expected manipulation steps for the complete chip processing process, and preset a number of expected features for each expected manipulation step; arbitrarily select the i-th manipulation step from the manipulation record, and count the number n of the same features between the i-th manipulation step and the j-th expected manipulation stepij , set a feature number threshold n max， If n is satisfied ij >n max And n ij =Max(n ia ), then the i-th control step is matched with the j-th expected control step, where a∈(1,b), Max(n ia ) is the maximum number of identical features between the ith manipulation step and each expected manipulation step, and b is the expected number of manipulation steps; the number of successfully matched manipulation steps in the manipulation record is counted m, and the degree of completion of the manipulation record is obtained as C=m / m total , where m total is the total number of expected control steps.

[0064] Step S200: randomly selecting a manipulation record, extracting several features of any manipulation step in the manipulation record, and performing a first feature evaluation on the manipulation step based on the several features; analyzing the state of the chip in the manipulation step, and performing a second feature evaluation on the manipulation step;

[0065] Wherein, step S200 includes the following steps:

[0066] Step S201: arbitrarily selecting an i-th manipulation step from any manipulation record, obtaining an expected manipulation step matching the i-th manipulation step, and presetting corresponding evaluation rules for each expected feature in the expected manipulation step;

[0067] Step S202: extract the feature set of the i-th manipulation step, obtain the processing data of any feature in the manipulation record, obtain the feature value presented by the feature in the manipulation record based on the preset evaluation rule, and calculate the average value of the feature values ​​presented by all features to obtain the first feature value Z1 of the i-th manipulation step i ;

[0068] Step S203: Obtain the expected processing time of the expected manipulation step as T ex , the preset chip real-time completion evaluation rule evaluates the completion status of the chip at any time during the processing, and sets the completion feature value at time t1 in the expected control step as W(t1), according to the formula:

[0069]

[0070] Calculate the total completion characteristic value CW of the control step within the expected processing time ex ;

[0071] Step S204: Obtain the change of the actual completion status of the chip in the i-th control step over time, set the actual completion characteristic value at time t2 in the i-th control step as W(t2), according to the formula:

[0072]

[0073] Among them, T ac is the actual processing time of the i-th manipulation step; the second eigenvalue Z2 of the i-th manipulation step is calculated i .

[0074] Step S300: Based on the difference between the two evaluation results, the abnormal features of any manipulation step are extracted, and the influence of each abnormal feature is analyzed; an abnormal recognition model is constructed to perform abnormal recognition on any manipulation record, and based on the completion status of the manipulation record, a recognition standard for determining whether any manipulation step has an abnormality is set;

[0075] Wherein, step S300 includes the following steps:

[0076] Step S301: randomly select a manipulation record and obtain the first characteristic value Z1 of the i-th manipulation step in the manipulation record i , the first characteristic average value from the start of processing to the completion of the i-th manipulation step in the manipulation record is obtained as Where e is a positive integer and e∈(1,i), Z1 e is the first eigenvalue of the e-th manipulation step; the eigenvalue presented by the j-th feature in the e-th manipulation step is tz ej , if tz ej <(Z1 i ) ave , then the j-th feature is set as an abnormal feature;

[0077] Step S302: Multiply the first characteristic values ​​of each manipulation step in the manipulation record from the start of processing to the completion of the i-th manipulation step to obtain the first characteristic comprehensive value from the start of processing to the completion of the i-th manipulation step

[0078] Step S303: Obtain the first characteristic difference value ΔZ1 from the start of processing to the completion of the i-th manipulation step in the manipulation record i =(Z1 i ) ave -(Z1 i ) cv, all abnormal features in the manipulation record from the start of processing to the completion of the i-th manipulation step are summarized to obtain an abnormal feature set, and the k-th abnormal feature is selected from the abnormal feature set, according to the formula:

[0079]

[0080] Among them, tz k is the characteristic value presented by the kth abnormal feature, g i is the number of abnormal features in the abnormal feature set; calculate the influence degree Y of the kth abnormal feature in the manipulation record up to the i-th manipulation step k ;

[0081] Example 2: Assuming that there are 2 manipulation steps, and the characteristic values ​​are 90% and 80% respectively, the average value of the first characteristic is 85%, and the comprehensive value of the first characteristic is 72%; assuming that 2 abnormal features are extracted from the 2 manipulation steps, the characteristic value of one abnormal feature is 80%, and the characteristic value of the other is 82%, and the influence degree of one abnormal feature is calculated to be Y k =(85%-80%) / (5%+3%)×13%=8.125%;

[0082] Step S304: Obtain the second characteristic value Z2 of the i-th manipulation step i , it is calculated that the influence of the remaining factors except for each abnormal feature in the control record up to the i-th control step is ΔZ2 i =(Z1 i ) cv -Z2 i ; Obtain the influence amplitude of the remaining factors up to the i-th control step in each control record, and calculate the average value to obtain the average influence amplitude ΔZ2(i);

[0083] Step S305: Set the completion degree of the control record to C. If C < 1 and i = C × m total , where m total is the total number of expected manipulation steps, then the first feature value presented in any manipulation step from the start of processing to the completion of the i-th manipulation step in the manipulation record and the influence degree of each abnormal feature included are obtained to construct an abnormality recognition model:

[0084]

[0085] Where e is a positive integer and e∈(1,i), Z1 e is the first eigenvalue of the e-th manipulation step, k is a positive integer and k∈(1,g e ), g eis the number of abnormal features contained from the beginning of processing to the e-th manipulation step, Y k is the influence degree of the kth abnormal feature included, ΔZ2(e) is the influence amplitude of the remaining factors except each abnormal feature from the start of processing to the eth manipulation step; the abnormal value F of the i-th manipulation step of the manipulation record is obtained by the abnormal recognition model i ;

[0086] Example 3: Assuming that the selected manipulation record contains only one manipulation step, and the first characteristic value is 80%, and contains two abnormal characteristics, the influence degrees are 8% and 5% respectively, and the influence range of other factors is set to 10%, the risk value of the selected manipulation step is calculated to be F i =(1 / 0.8)×(1+0.08)×(1+0.05)×(1+0.1)=1.559;

[0087] Step S306: extract each control record with a completion degree C < 1, obtain the abnormal value of the last control step in each extracted control record, and select the abnormal value with the smallest value as the abnormal threshold for determining whether any control step has an abnormality.

[0088] Step S400: Whenever the chip is processed, the current operation step of the chip is evaluated; based on the evaluation result, abnormal features are extracted, and the abnormal operation step of the chip is identified through an abnormal identification model, and it is determined whether to issue an abnormal reminder;

[0089] Wherein, step S400 includes the following steps:

[0090] Step S401: Get the current step of the chip being processed, and set the current step to the i-th step. ’ The i-th control step is obtained by using the preset evaluation rules. ’ The first characteristic values ​​of each manipulation step before the first manipulation step are averaged to obtain a first characteristic average value; and abnormal features satisfying that the characteristic value is less than the first characteristic average value are extracted;

[0091] Step S402: Obtain each abnormal feature in the i-th ’ The impact degree obtained from the feedback of the i-th control step is input into the abnormality recognition model, and the output is the current state of the chip in the i-th control step. ’ The abnormal value F presented in the control step; set the abnormal threshold to F th , if F>F th , an abnormal reminder is sent, in the i ’ Stop processing when one control step is reached.

[0092] An intelligent control system for chip processing equipment, the control system includes a control record analysis module, a control step evaluation module, an abnormal feature processing module and a control abnormality identification module;

[0093] The control record analysis module is used to build an equipment control data management system to record each processing process and generate corresponding control records; divide any control record into steps, extract features for each step, and analyze the completion status of the control record;

[0094] A manipulation step evaluation module is used to select a manipulation record at random, extract several features of any step in the manipulation record, and perform a first feature evaluation on the step based on the several features; analyze the state presented by the chip in the step, and perform a second feature evaluation on the step;

[0095] The abnormal feature processing module is used to extract the abnormal features of any operation step based on the difference between the two evaluation results, and analyze the influence of each abnormal feature; build an abnormal recognition model to identify the abnormality of any operation record, and set the recognition standard for judging whether any operation step has an abnormality based on the completion status of the operation record;

[0096] The control anomaly recognition module is used to evaluate the current control step of the chip whenever the chip is processed; extract abnormal features based on the evaluation results, and use the abnormal recognition model to identify the abnormality of the current control step of the chip, and determine whether to issue an abnormal reminder.

[0097] Among them, the control record analysis module includes a control record generation unit and a control step analysis unit;

[0098] The control record generation unit is used to build an equipment control data management system to record each processing process and generate corresponding control records; the control step analysis unit is used to divide any control record into steps, extract features of each step, and analyze the completion status of the control record.

[0099] Wherein, the manipulation step evaluation module includes a first feature evaluation unit and a second feature evaluation unit;

[0100] The first feature evaluation unit is used to arbitrarily select a manipulation record, extract several features of any step in the manipulation record, and perform a first feature evaluation on the step based on the several features; the second feature evaluation unit is used to analyze the state of the chip in the step, and perform a second feature evaluation on the step.

[0101] Among them, the abnormal feature processing module includes an abnormal feature extraction unit and an abnormal identification processing unit;

[0102] The abnormal feature extraction unit is used to extract the abnormal features of any operation step based on the difference between the two evaluation results, and analyze the influence of each abnormal feature; the abnormal identification processing unit is used to construct an abnormal identification model to identify the abnormality of any operation record, and set the identification standard for judging whether any operation step has an abnormality based on the completion status of the operation record.

[0103] Among them, the control abnormality identification module includes a real-time step evaluation unit and an abnormality identification and judgment unit;

[0104] The real-time step evaluation unit is used to evaluate the current operation step of the chip whenever the chip is processed; the abnormality identification and judgment unit is used to extract abnormal features based on the evaluation results, and to identify the abnormality of the current operation step of the chip through the abnormality identification model, and to determine whether to issue an abnormal reminder.

[0105] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A chip processing equipment intelligent control method, characterized in that: The control method comprises the following steps: Step S100: constructing an equipment control data management system to record each processing process and generate corresponding control records; dividing any control record into control steps, extracting features of each control step, and analyzing the completion status of the control record; Step S200: randomly selecting a manipulation record, extracting several features of any manipulation step in the manipulation record, and performing a first feature evaluation on the manipulation step based on the several features; analyzing the state of the chip in the manipulation step, and performing a second feature evaluation on the manipulation step; Step S300: Based on the difference between the two evaluation results, the abnormal features of any manipulation step are extracted, and the influence of each abnormal feature is analyzed; an abnormal recognition model is constructed to perform abnormal recognition on any manipulation record, and based on the completion status of the manipulation record, a recognition standard for determining whether any manipulation step has an abnormality is set; Step S400: Whenever the chip is processed, the current operation step of the chip is evaluated; based on the evaluation result, abnormal features are extracted, and the abnormality of the current operation step of the chip is identified through the abnormality identification model, and it is determined whether to issue an abnormal reminder.

2. The intelligent control method for chip processing equipment according to claim 1, characterized in that: The step S100 includes the following steps: Step S101: Whenever a chip is placed on any chip processing equipment for processing, monitoring data monitored in the chip processing equipment is obtained; all chip processing equipment involved in any control record are obtained, and a control step of the control record is set for each chip processing equipment; Step S102: setting the monitoring data monitored in any chip processing equipment as processing data of the corresponding step; arbitrarily selecting a control step, presetting a number of operation types and data types of each operation type for the control step, dividing the processing data according to the data type, and matching it with each operation type to obtain a number of operation types of the control step, setting any matched operation type as a feature of the control step, and generating a feature set of the control step; Step S103: preset a fixed number and fixed order of expected manipulation steps for the complete chip processing process, and preset a number of expected features for each expected manipulation step; arbitrarily select the i-th manipulation step from the manipulation record, and count the number n of the same features between the i-th manipulation step and the j-th expected manipulation step ij , set a feature number threshold n max， If n is satisfied ij >n max And n ij =Max(n ia ), then the i-th control step is matched with the j-th expected control step, where a∈(1,b), Max(n ia ) is the maximum number of identical features between the ith manipulation step and each expected manipulation step, and b is the expected number of manipulation steps; the number of successfully matched manipulation steps in the manipulation record is counted m, and the degree of completion of the manipulation record is obtained as C=m / m total , where m total is the total number of expected control steps.

3. The intelligent control method for chip processing equipment according to claim 2, characterized in that: The step S200 includes the following steps: Step S201: arbitrarily selecting an i-th manipulation step from any manipulation record, obtaining an expected manipulation step matching the i-th manipulation step, and presetting corresponding evaluation rules for each expected feature in the expected manipulation step; Step S202: extract the feature set of the i-th manipulation step, obtain the processing data of any feature in the manipulation record, obtain the feature value presented by the feature in the manipulation record based on the preset evaluation rule, and calculate the average value of the feature values ​​presented by all features to obtain the first feature value Z1 of the i-th manipulation step i ; Step S203: Obtain the expected processing time of the expected manipulation step as T ex , the preset chip real-time completion evaluation rule evaluates the completion status of the chip at any time during the processing, and sets the completion feature value at time t1 in the expected control step as W(t1), according to the formula: Calculate the total completion characteristic value CW of the control step within the expected processing time ex ; Step S204: Obtain the change of the actual completion status of the chip in the i-th control step over time, set the actual completion characteristic value at time t2 in the i-th control step as W(t2), according to the formula: Among them, T ac is the actual processing time of the i-th manipulation step; the second eigenvalue Z2 of the i-th manipulation step is calculated i .

4. The intelligent control method for chip processing equipment according to claim 3, characterized in that: The step S300 includes the following steps: Step S301: randomly select a manipulation record and obtain the first characteristic value Z1 of the i-th manipulation step in the manipulation record i , the first characteristic average value from the start of processing to the completion of the i-th manipulation step in the manipulation record is obtained as Where e is a positive integer and e∈(1,i), Z1 e is the first eigenvalue of the e-th manipulation step; the eigenvalue presented by the j-th feature in the e-th manipulation step is tz ej , if tz ej <(Z1 i ) ave , then the j-th feature is set as an abnormal feature; Step S302: Multiply the first characteristic values ​​of each manipulation step in the manipulation record from the start of processing to the completion of the i-th manipulation step to obtain the first characteristic comprehensive value from the start of processing to the completion of the i-th manipulation step Step S303: Obtain the first characteristic difference value ΔZ1 from the start of processing to the completion of the i-th manipulation step in the manipulation record i =(Z1 i ) ave -(Z1 i ) cv , all abnormal features in the manipulation record from the start of processing to the completion of the i-th manipulation step are summarized to obtain an abnormal feature set, and the k-th abnormal feature is selected from the abnormal feature set, according to the formula: Among them, tz k is the characteristic value presented by the kth abnormal feature, g i is the number of abnormal features in the abnormal feature set; calculate the influence degree Y of the kth abnormal feature in the manipulation record up to the i-th manipulation step k ; Step S304: Obtain the second characteristic value Z2 of the i-th manipulation step i , it is calculated that the influence of the remaining factors except for each abnormal feature in the control record up to the i-th control step is ΔZ2 i =(Z1 i ) cv -Z2 i ; Obtain the influence amplitude of the remaining factors up to the i-th control step in each control record, and calculate the average value to obtain the average influence amplitude ΔZ2(i); Step S305: Set the completion degree of the control record to C. If C < 1 and i = C × m total , where m total is the total number of expected manipulation steps, then the first feature value presented in any manipulation step from the start of processing to the completion of the i-th manipulation step in the manipulation record and the influence degree of each abnormal feature included are obtained to construct an abnormality recognition model: Where e is a positive integer and e∈(1,i), Z1 e is the first eigenvalue of the e-th manipulation step, k is a positive integer and k∈(1,g e ), g e is the number of abnormal features contained from the beginning of processing to the e-th manipulation step, Y k is the influence degree of the kth abnormal feature included, ΔZ2(e) is the influence amplitude of the remaining factors except each abnormal feature from the start of processing to the eth manipulation step; the abnormal value F of the i-th manipulation step of the manipulation record is obtained by the abnormal recognition model i ; Step S306: extract each control record with a completion degree C < 1, obtain the abnormal value of the last control step in each extracted control record, and select the abnormal value with the smallest value as the abnormal threshold for determining whether any control step has an abnormality.

5. The intelligent control method for chip processing equipment according to claim 4, characterized in that: The step S400 includes the following steps: Step S401: Get the current step of the chip being processed, and set the current step to the i-th step. ’ The i-th control step is obtained by using the preset evaluation rules. ’ The first characteristic values ​​of each manipulation step before the first manipulation step are averaged to obtain a first characteristic average value; and abnormal features satisfying that the characteristic value is less than the first characteristic average value are extracted; Step S402: Obtain each abnormal feature in the i-th ’ The impact degree obtained from the feedback of the i-th control step is input into the abnormality recognition model, and the output is the current state of the chip in the i-th control step. ’ The abnormal value F presented in the control step; set the abnormal threshold to F th , if F>F th , an abnormal reminder is sent, in the i ’ Stop processing when one control step is reached.

6. An intelligent control system for chip processing equipment, used to execute an intelligent control method for chip processing equipment according to any one of claims 1 to 5, characterized in that: The control system includes a manipulation record analysis module, a manipulation step evaluation module, an abnormal feature processing module and a manipulation abnormality identification module; The control record analysis module is used to construct an equipment control data management system to record each processing process and generate corresponding control records; divide any control record into steps, extract features of each step, and analyze the completion status of the control record; The manipulation step evaluation module is used to select any manipulation record, extract several features of any step in the manipulation record, and perform a first feature evaluation on the step based on the several features; Analyze the state of the chip in the steps, and perform a second characteristic evaluation on the steps; The abnormal feature processing module is used to extract the abnormal features of any manipulation step based on the difference between the two evaluation results, and analyze the influence of each abnormal feature; construct an abnormal recognition model to recognize abnormalities in any manipulation record, and set the recognition standard for judging whether any manipulation step has abnormalities based on the completion status of the manipulation record; The control abnormality identification module is used to evaluate the current control step of the chip whenever the chip is processed; extract abnormal features based on the evaluation results, and identify the abnormality of the current control step of the chip through the abnormality identification model, and determine whether to issue an abnormal reminder.

7. The intelligent control system for chip processing equipment according to claim 6, characterized in that: The control record analysis module includes a control record generation unit and a control step analysis unit; The control record generation unit is used to construct an equipment control data management system to record each processing process and generate corresponding control records; the control step analysis unit is used to divide any control record into steps, extract features of each step, and analyze the completion status of the control record.

8. The intelligent control system for chip processing equipment according to claim 6, characterized in that: The manipulation step evaluation module includes a first feature evaluation unit and a second feature evaluation unit; The first feature evaluation unit is used to arbitrarily select a manipulation record, extract several features of any step in the manipulation record, and perform a first feature evaluation on the step based on the several features; the second feature evaluation unit is used to analyze the state of the chip in the step, and perform a second feature evaluation on the step.

9. The intelligent control system for chip processing equipment according to claim 6, characterized in that: The abnormal feature processing module includes an abnormal feature extraction unit and an abnormal identification processing unit; The abnormal feature extraction unit is used to extract the abnormal features of any manipulation step based on the difference between the two evaluation results, and analyze the influence of each abnormal feature; the abnormal identification processing unit is used to construct an abnormal identification model to identify the abnormality of any manipulation record, and set the identification standard for judging whether there is an abnormality in any manipulation step based on the completion status of the manipulation record.

10. The intelligent control system for chip processing equipment according to claim 6, characterized in that: The control abnormality identification module includes a real-time step evaluation unit and an abnormality identification and judgment unit; The real-time step evaluation unit is used to evaluate the current operation step of the chip whenever the chip is processed; the abnormal identification and judgment unit is used to extract abnormal features based on the evaluation results, and to identify abnormalities in the current operation step of the chip through the abnormal identification model, and to determine whether to issue an abnormal reminder.

Images