METHOD FOR CONTROLLING A PRODUCTION PROCESS FOR THE PRODUCTION OF COMPONENTS

Abstract

A method for controlling a production process for components, wherein at least the components or a production device used for the production of the components have a plurality of characteristics that are recorded by measurement techniques, wherein at least one characteristic is defined as the main characteristic and at least one additional characteristic as a secondary characteristic; and comprising at least the following steps: a) defining a verification plan to record at least one main characteristic and at least one secondary characteristic by means of verifications, wherein the at least one main characteristic is measured with a first verification frequency and the at least one secondary characteristic is measured with a second verification frequency; wherein for the at least one main characteristic at least one stability criterion is defined;b) produce the components and carry out in parallel the verification plan for the generation of verification results, in which only at least one main characteristic is verified with the first verification frequency; c) evaluate the determined verification results; and, if at least one verification result for at least one main characteristic violates the stability criterion, d) continue with the execution of the verification plan, provided that at least one secondary characteristic is also verified.

Description

METHOD FOR CONTROLLING A PRODUCTION PROCESS FOR THE bn i znn / zznz / E / YiAi PRODUCTION OF COMPONENTS Field of Invention The present invention relates to a method for controlling a production process, where components are produced by means of the production process. Background of the Invention During the mass production of components, various quality and process measurements are carried out regularly. These measurements are usually performed randomly. For example, a specific characteristic might be recorded at regular intervals for a given number of components (e.g., the weight of a component produced by sintering). Other measurements relate to other component characteristics, such as length, density, surface parallelism, or diameter. Alternatively or in addition to this, process parameters, such as the pressure curve on a machine during the pressing process in the production of green parts, might also be recorded at regular intervals. Through a purely statistical definition of Ref. 331511: Verification intervals do not address the constantly changing risks that arise in a real-world process. In stable processes that produce consistently high-quality components, too many measurements are often taken for this reason. This leads to avoidable costs. In unstable processes, on the other hand, there is a risk that, due to too few measurements, defects will not be adequately recorded. Detailed Description of the Invention Based on this, the object of the present invention is at least to alleviate or even solve the problems described with reference to the prior art. In particular, a method for controlling a production process for the manufacture of components shall be disclosed, in which a verification cost is calculated based on the stability of the production process. To accomplish these tasks, a method is proposed in accordance with the features of claim 1. The favorable improvements are the subject of the dependent patent claims. The features individually specified in the patent claims can be combined with each other in a technologically significant manner and can be supplemented by facts explained in the description, which shows other embodiments of the invention. A method is proposed for controlling a component production process, i.e., for the production of components. At least the components, or a production facility used for the production of the components, have a plurality of characteristics that are recorded using measurement techniques (e.g., a volume, weight, density, surface parallelism, or diameter of the component; a temperature, pressure, or energy consumption of the production equipment). At least one characteristic is defined as a primary characteristic, and at least one other characteristic as a secondary characteristic. The method includes, at a minimum, the following steps: a) Establish a verification plan to record the minimum of a primary characteristic and the minimum of a secondary characteristic by verification (i.e., for example, by measurement), wherein at least the minimum of a primary characteristic is measured with (a first value) of a first verification frequency and the minimum of a secondary characteristic with a second verification frequency; wherein for the minimum of a primary characteristic at least one stability criterion is defined; b) Produce the components and parallel execution of the verification plan for the production of verification results, in which (temporarily) only at least one main characteristic is verified, in the first verification frequency; c) Evaluate the determined verification results; provided that, if at least one verification result for at least one main characteristic violates the stability criterion, d) Continue the execution of the verification plan, ensuring that at least the minimum of one secondary characteristic is also verified. In this case, we propose using correlations between the various characteristics of a component and / or a production facility. For this purpose, we establish, for example, a single quality and / or process characteristic (or a small group of them) as the primary characteristic. This primary characteristic can be measured automatically and preferably for each component produced. The other characteristics (secondary characteristics) are measured at a significantly lower frequency (i.e., not for all components); ideally, rarely or not at all. In particular, the verification plan bn i znn / zznz / E / YiAi establishes and includes only one main feature. By analyzing the time course of the main characteristic (or the small group of main characteristics) and the correlations between all the characteristics, it is possible to derive, in particular, forecasts about the quality status of a measured component. If changes in the distribution of the main characteristic(s) lead to a higher probability of a potential lack of tolerance for one or more of the other component characteristics, then additional measurements are taken (automatically) for these characteristics (and, as required, process corrections are implemented). Process stability is monitored exclusively by the main characteristic. In particular, by using IoT (Internet of Things) technologies, modern production facilities can automatically record and store certain characteristics and their verification results for each individual component produced. These characteristics can correspond to quality or process characteristics. Examples include the component's weight or the pressing force of the production system. This results in a significantly higher measurement speed (e.g., 15 measurements per minute corresponding to a production speed of 15 strokes) compared to a traditional verification plan (e.g., 5 measurements every two hours).In other words, this makes it particularly possible to record trends in characteristics in a population of 100% of all components produced, without the need for (costly) manual measurements of characteristics (e.g., by extracting the components and performing a measurement in a laboratory that does not belong to the production facility, or similar). A production facility can, for example, be a press, a sintering furnace, or a complete production plant with both a press and a sintering furnace. Additionally, a production facility can also include, for example, a building housing a press and / or a sintering furnace. Other characteristics that can be recorded include ambient air temperature, air humidity, and air composition (proportions of shielding gas / oxygen, etc.). In particular, it can be assumed that there is a minimum number of key characteristics that are directly correlated with the quality of the components produced. Furthermore, it can be assumed that if this minimum number of key characteristics is in a stable equilibrium (i.e., for example, if no negative trend is observed), excessive verification of these characteristics is unnecessary for quality monitoring. If changes in the trend of the main characteristics (e.g., lack of boundary values, downward or upward deviations, oscillation within boundary values, atypical constellation of characteristics, dispersions, ...) lead to a higher probability of tolerance failures (or a lack of stability criteria) with respect to one or more characteristics (it could also be said, in the case of establishing a negative forecast), additional measurements for the affected characteristics should be implemented (automatically). The relevant characteristics are, in particular, the characteristics covered by the forecast, the forecast being established by observing at least one main characteristic. Any necessary process corrections can be made, in particular, based on all the verification results (i.e., bn i znn / zznz / E / YiAi of the main and secondary characteristics). For this purpose, one can use not only the results of verifications from the ongoing production of the components, but also historical data from previous production runs. A production process involves the manufacture of a component, for example, the production of a component using sintering technology. For this, a material in powder form is supplied, compressed to obtain a green part, and then treated with sintering (i.e., thermal) technology. If necessary, machining with chip removal can be performed. As part of the production process, a large number (for example, at least 100, 1000, or several 10,000) of ideally identical components are produced. As a primary characteristic to be recorded, one can record, for example, the weight of a component produced by sintering, as an additional primary characteristic, or as a secondary characteristic, such as length, density, surface parallelism, or diameter of the component. Furthermore, a burr or the surface appearance of the component, for example, matte or glossy, can also be defined as a characteristic. Similarly, it is possible to record a characteristic of a production facility, for example, an incident pressure, such as press pressure (as a primary characteristic), or energy consumption (as a secondary characteristic). A first verification frequency could represent, for example, an initial value for a newly started production process of the component or components. This value could, for example, comprise one verification of one component for every 1,000 components, or one verification of a random sample of several components (for example, five components manufactured sequentially) for every 1,000 components, or something similar. Alternatively or simultaneously, the first verification frequency could be set based on a time period of the production process. Then, for example, a characteristic of a production facility could be recorded, or a verification (on one or more components) could be carried out for each hour of the production process. It is also possible to consider the number of components produced during that period or that passed through a specific processing stage. The results of the verification can be determined by evaluating a single value (the result of the verification of the investigated characteristic) or by evaluating several values ​​simultaneously. The evaluation or judgment based on multiple values ​​can be based on mathematical methods, such as calculating the average and / or standard deviation within a random sample. A stability criterion can be established as a defined deviation from a nominal value for the respective characteristic. The stability criterion should not, in this particular case, correspond to a permissible tolerance (a value outside the tolerance would mean, for example, that the component is not in order, i.e., defective) for the nominal value. Specifically, the tolerance is established as wider than the stability criterion. Consequently, for example, a component would not be defective simply because it does not meet the stability criterion (and is therefore still within the tolerance). In particular, it can also be anticipated that a violation of the stability criterion depends on a method with which, for example, a random sample of components analyzed during a verification is evaluated. For example, it is possible to define as a subset of the components of a random sample (the random sample bn i znn / zznz / E / YiAi includes, for example, five components) a number beyond which a violation of the stability criterion is concluded. That is, a minimum number of components in a random sample (i.e., for example, at least two components of the five-component random sample) must have, for example, a certain minimum deviation from a nominal value of the first characteristic, and only then is a violation of the stability criterion considered to have occurred. The violation of the stability criterion makes it possible to identify instability in the process, so that through a corrective intervention in the production process it is possible to avoid an imminent tolerance failure. The stability criterion can be applied, for example, to a measure of the variation in a verification result. In particular, a violation of the stability criterion can be determined if the verification results show a trend or a sudden change that occurs sporadically. In particular, the production of the components and the implementation of the verification plan, i.e., the execution of measurements of the first characteristic, are carried out temporarily in parallel relative to each other. In particular, the first verification frequency comprises a verification of at least 50% of the components (i.e., the relevant characteristic of the component or production facility is verified for at least 50% of all components), preferably at least 75%, and especially preferably 100%. In particular, the second verification frequency differs from the first verification frequency (i.e., it is higher or lower). Specifically, the second verification frequency is lower or even significantly lower than the first verification frequency. Preferably, the second verification frequency comprises, in total, at most 50% (preferably at most 20%, more preferably at most 5%) of the number of verifications carried out as a consequence of the first verification frequency. Specifically, only at least one primary feature is checked temporarily, whereas at a second, significantly lower verification frequency (continuously and overlapping with the first verification frequency), at least one secondary feature is checked. Preferably, the second verification frequency comprises a verification of 0% of the components and a verification of at least one secondary feature is adopted only if a stability criterion of at least one primary feature was violated. Specifically, at least one main characteristic (preferably only one) is recorded at one or more process steps during the first verification frequency. If it falls within predefined limits for a predefined period (i.e., the stability criterion is not violated), no further verifications or measurements are performed. If the stability criterion limits are violated, all characteristics are recorded and verified according to the verification plan. Specifically, the main individual characteristic is, for example, the weight of a component. This main characteristic is verified, at least in the context of the component's automated production. In the case of a component's weight, it is possible to weigh the number of components in question. In the case of pressure / press pressure, it is possible to record, for example, the pressure applied by the production system for the respective number of components. In particular, it is possible that, as a trigger to carry out the additional measurements or all of them in accordance with the verification plan, changes in the variability and distribution properties of the verification results of at least one main characteristic may be used (in addition). Specifically, there are at least a plurality of secondary characteristics, and a violation of a stability criterion for a primary characteristic only requires verification of a selection of secondary characteristics. That is, not all secondary characteristics identified in the verification plan are verified, but only a selection. Furthermore, it is possible that additional secondary characteristics will be verified or the selection will be changed. In particular, there is a plurality of main characteristics (e.g., weight and surface color of the component), so the violation of a stability criterion of a main characteristic (e.g., weight) makes it necessary to verify a first selection of secondary characteristics (e.g., length, width of the component; pressure of the bn ι znn / zznz / E / YiAi production facility) and the violation of a stability criterion of another main characteristic (e.g., surface color) the verification of a second selection of secondary characteristics (e.g., composition of the starting materials for the component, temperature in the production facility, pressure of the production equipment). In particular, a violation of a stability criterion includes at least one of the following conditions: • the verification results follow a trend (they continuously rise or fall); • the verification results fluctuate within a range that exceeds the limit values ​​defined by the stability criterion; • The results of the verification of at least several different main characteristics differ from a predefined correlation of the main characteristics (e.g., weight and volume, or height, length, width, or weight and pressure depend on each other) In particular, it is possible that in the case of changing verification results (for example, in the case of a trend, a sudden change, or the appearance of large intervals) of the main feature or of the main features, a verification of other (or all) of the features specified in the verification plan may be initiated. Specifically, the relevant verification frequencies are only increased for those features whose verification results have the listed properties. In particular, in the case of changing verification results (for example, in the presence of a trend, a sudden change, or excessively large intervals) of the main characteristic(s), it is possible to verify only a selection of characteristics (i.e., not all the characteristics in the verification plan). The selection of characteristics can be determined by experts. Alternatively, a new selection can be made for each case or stored as a fixed selection in the verification plan. At the beginning of a component production process, the weight of the produced components could be continuously recorded, for example, and considered the primary characteristic. In a subsequent step, additional primary characteristics could be added or replaced by those previously used. Conceivable additional main characteristics bn ι znn / zznz / E / YiAi would be, for example, a press force or a pressure applied to the component by the production device, or even individual measurements of the produced components. In particular, at least the first verification frequency is modified, such that: • (or) the first verification frequency is increased if at least one verification result for at least one main feature violates the stability criterion; or • the first verification frequency is reduced when a certain number of successive verification results of at least one main feature are consistent with the stability criterion. Specifically, the first verification frequency (assigned to at least one main feature or to all main features) is modified; that is, the first value of the first verification frequency is adjusted to a second value. Therefore, the first verification frequency is increased if, for example, at least one verification result for at least one main feature violates the stability criterion. Conversely, the first verification frequency is reduced if all verification results comply with the stability criterion. Therefore, it is proposed here that when the production process develops stably (and the verification results do not show a major deviation, at least with respect to at least one main characteristic, i.e., the verification results agree with the stability criterion), the first verification frequency can be reduced. If a possible instability of the production process is identified (i.e., if, for example, at least one verification result with respect to at least one main characteristic shows a major deviation, i.e., that the verification result violates the stability criterion), the verification frequency can be increased. The expert will use, for example, what is called an intervention limit, a tolerance limit, or a stability limit to define a stability criterion. Therefore, if the verification results for the characteristic being checked fall within the defined limits, the verification frequency is progressively reduced. Conversely, the verification frequency is increased as soon as the results bn i znn / zznz / E / YiAi of the verification of the characteristic to be verified are outside the defined limits, since this transgression of the limit indicates a greater quality risk for the component. Under stable production process conditions, the measurements of the characteristics to be verified can be minimized, and it is still possible to guarantee a high quality of the components produced. In particular, based on a record of measurement values ​​and a record of measurement values ​​linked to a computer, it is possible to continuously analyze the verification results and evaluate them as to whether the production process is stable or whether interventions in the production process are required. Based on the measurements taken and the resulting verification results, it is possible to identify a trend in the change of a characteristic (e.g., weight or height). If the verification results fall within defined limits over a defined period or across a certain number of random samples (satisfying the stability criterion), it can be concluded that the process for the characteristic being verified is stable, and therefore the verification frequency can be reduced. This reduction can also be implemented in stages, and ideally, the verification frequency can be set to a defined minimum. However, if one or more verification results violate at least one stability criterion in a single verification, this indicates an unstable production process. In such cases, the verification frequency is increased, or other characteristics (primary and / or secondary characteristics) are reviewed, and the production process is monitored more closely. Simultaneously, intervention may be implemented in the production process to restore stability, including the rejection of potentially defective components. The increased verification frequency or the verification of additional characteristics ensures that the production process is monitored more closely over a defined period.Only if the verification results satisfy at least one stability criterion over a defined period of time or over a certain number of random or component verifications, can the verification frequency and / or the number of monitored features be reduced again (gradually). In particular, at least the main features (and possibly additional secondary features) are linked to each other by a model. Therefore, based on the verification results of at least one main feature, conclusions are drawn about a change in other main or secondary features under consideration by the model. A verification of at least one of these main or secondary features is initiated as soon as a violation of a stability criterion associated with that feature is predicted. For example, a continuous increase in the weight of a component can be used as the basis for a forecast. Based on this forecast, a future violation of a stability criterion could be predicted. Furthermore, checks of other characteristics could be initiated based on this forecast, allowing for proactive corrective intervention in a production process if necessary (even before a stability criterion is violated). In particular, sensitivity and dependency analyses can be performed. Based on these, both the main characteristics and their influence on the remaining (secondary) characteristics can be identified. These analyses will also determine, in particular, the verification frequency with which the main characteristics should be recorded, whether it should be with high frequency (e.g., 100% of the components) or perhaps only through periodic measurements at fixed time and / or production intervals. In particular, derived main characteristics are foreseen that are not directly measurable, but rather can be calculated from several measured characteristics. An example is the density of a component, which is calculated from the measured weight and volume (which in turn are derived, for example, from the measured height of the component). Specifically, a model is created that, based on the values ​​of the main features, can estimate the distribution of the values ​​of the remaining features. The model is based, in particular, on the sensitivity and dependency analyses mentioned earlier. Specifically, feature measurements are automatically initiated if a higher probability of a predicted tolerance violation or a breach of a stability criterion is calculated for one or more features, or if the forecast interval (which measures the uncertainty) of an estimated feature becomes too large. The verification results provided by these automatically initiated measurements are used, in particular, to validate the model and / or improve it online (in real time). In particular, the model is continuously validated and modified based on the verifications performed and the results of the verifications. A model validation state, expressed particularly in the binary distinction (gradation) good or bad, can be associated with the generally known concept of model quality such that a model quality that results in not being found within a defined bandwidth is equivalent to the state of injury to a stability criterion. Automatically initiated measurements or checks are used not only to validate the model but also to extend its scope. For example, measured features may be assigned a higher level of confidence than features that are only estimated. Consequently, features measured as a result of automatic initiation can temporarily become core features, which can then be used to initiate broader measurements or replace existing ones. bn i znn / zznz / E / YiAi In particular, the model for selecting the main or secondary feature to be verified additionally takes into account at least one of the following factors: • costs of additional verification; • time required for additional verification; • Availability of the required measuring equipment. In particular, when measurements or verifications are automatically initiated, the marginal conditions of these measurements are replicated in the model to ensure the most favorable measures are always used. For example, factors such as effort, costs, duration, load on the measurement system, and measurement reliability are modeled and considered when selecting the measures. In particular, additional measurements or verifications are evaluated specifically. Multiple initiations may point, for example, to systematic or underlying faults in processes and / or machines / tools. These faults must be identified and subsequently eliminated. Furthermore, dynamic adjustments to the maintenance intervals of the installations are conceivable, based on the frequency of characteristic measurements. In particular, the main characteristics and dependencies (on other characteristics) can be re-identified for each new product or process. However, to reduce this expense, it is possible to determine product / process classes that share similar behaviors / requirements. To this end, existing measured values ​​(e.g., quality, process, and machine data) can be evaluated, for example, according to similar patterns in deviations. Furthermore, it is possible to establish dependencies between these patterns and the properties of the components to be produced (e.g., size, weight, geometry, density, etc.). By comparing a new component to be produced with components already produced, a suitable base model class (defining the recommended checks) can be proposed for the new component. In particular, the basic models to be developed are becoming increasingly generic. Once defined, they are particularly easy to adapt to different new products and / or processes. For this purpose, multiple measurements can be taken at the start of a new production run. The model parameters and corresponding confidence intervals can be adapted, in particular automatically. As soon as a predefined stability criterion has been reached, the model can be switched to automatic control (and used for the process), after which measurements or verification are only carried out as needed. The stability criterion could be calculated in an illustrative way from the main characteristics and the confidence intervals of the remaining characteristics. Concrete example for the correct case (i.e., for the case of a working model): The main characteristics behave discretely and the confidence intervals are found for a predefined time within predefined limit values. In particular, the verification plan for a component to be produced is created at least partially in an automated manner during the production of the component, by identifying at least one primary feature and one secondary feature through the results of feature verification. In particular, in order to validate the verification plan, in addition to verifying at least one identified main feature, a plurality of other features can be verified periodically. Specifically, at the end of a production run, in addition to the verification results, the following information is saved: model parameters, parameter behavior and characteristics during the learning process, and the number and type of measurements or verifications initiated (automatically). This information can be automatically incorporated into the stored model during the next production run, thereby progressively improving its performance. In particular, based on the evaluation of the measurements or verifications initiated (automatically), the need for automation of measurements can be derived for all products and processes. In particular, the production of the component comprises a plurality of process steps, wherein a feature is present after a first process step, and is verified, and wherein there is an additional feature (recorded by metrology) after a second process step that follows the first process step. bn ι znn / zznz / E / YiAi In addition, a data processing device is proposed, comprising a processor that is configured to carry out the described method. As a precaution, it should be noted that the numerical terms used here (first, second, third, ...) serve primarily (and only) to distinguish between several identical objects, sizes, or processes; that is, they do not necessarily establish any dependency or sequence between these objects, sizes, or processes. Should a dependency or sequence be required, it is either explicitly stated here or becomes obvious to a qualified person studying the specifically described design. It is hereby stated that, as of this date, the best method known to the applicant for putting the aforementioned invention into practice is the one that is clear from the present description of the invention.

Claims

1. A method for controlling a production process for components, wherein at least the components or a production device used for the production of the components have a plurality of characteristics that are recorded by measurement techniques, wherein at least one characteristic is defined as the main characteristic and at least one additional characteristic as a secondary characteristic; characterized in that it comprises at least the following steps: a) establishing a verification plan to record at least one main characteristic and at least one secondary characteristic by means of verifications, wherein the at least one main characteristic is measured with a first verification frequency and the at least one secondary characteristic is measured with a second verification frequency; wherein for the at least one main characteristic at least one stability criterion is defined;b) produce the components and carry out in parallel the verification plan for the generation of verification results, whereby only at least one main characteristic is verified with the first verification frequency; c) evaluate the determined verification results; and, if at least one verification result for at least one main characteristic violates the stability criterion, d) continue with the execution of the verification plan, whereby at least one secondary characteristic is also verified.

2. Method according to claim 1, characterized in that the first verification frequency comprises a verification of each component.

3. Method in accordance with any of the preceding claims, characterized in that the second verification frequency is less than the first verification frequency.

4. A method in accordance with any of the preceding claims, characterized in that at least one main feature is verified automatically during the production of the component. bn ι znn / zznz / E / YiAi 5. A method in accordance with any of the preceding 31 claims, characterized in that there are at least a plurality of secondary features, wherein the violation of the stability criterion of at least one main feature requires verification of only a selection of the secondary features.

6. Method according to claim 5, characterized in that there are a plurality of principal features, wherein the violation of a stability criterion of a principal feature requires verification of a first selection of secondary features and the violation of a stability criterion of another principal feature requires verification of a second selection of secondary features.

7. A method in accordance with any of the preceding claims, characterized in that the violation of a stability criterion includes at least one of the following conditions: • the verification results follow a trend; • the verification results fluctuate within an interval that exceeds the limit values ​​defined by the stability criterion; • the verification results of at least different principal characteristics differ from a previously defined correlation of the principal characteristics.

8. A method in accordance with any of the preceding claims, characterized in that at least the first verification frequency is modified, wherein: • the first verification frequency is increased if at least one verification result for the at least one main characteristic violates the stability criterion, or • the first verification frequency is reduced if at least a certain number of successive verification results of the at least one main characteristic agree with the stability criterion.

9. A method in accordance with any of the preceding claims, characterized in that at least the main features are linked to each other by a model, wherein on the basis of the verification results of at least one main feature a variation of other main features or secondary features is concluded by taking into account the model; wherein a verification of at least one of these main features or secondary features is initiated as soon as a violation of a stability criterion assigned to this feature is predicted.

10. Method according to claim 9, characterized in that the model is continuously validated and modified on the basis of the verifications carried out and the results of the verifications.

11. Method in accordance with any of the preceding claims 9 and 10, characterized in that the model for selecting the main or secondary feature to be additionally verified takes into account at least one of the following factors: • costs of the additional verification; • time expenditure of the additional verification; • availability of the required measuring facility.

12. A method in accordance with any of the preceding claims, characterized in that the verification plan for a component to be produced is created at least partially in an automated manner during the production of the component, by identifying, through the verification results of the characteristics, at least one main characteristic and one secondary characteristic.

13. A method according to claim 12, characterized in that, for the validation of the bn i znn / zznz / E / YiAi verification plan, in addition to verifying at least one identified main feature, a plurality of other features are periodically verified.

14. A data processing apparatus characterized in that it comprises a processor configured to execute the method according to any of the preceding claims.