Machine and method for operating the same

Abstract

The invention relates to a method for operating a machine, in particular a cyclically operating machine, in particular a packaging machine for packaging products, such as products of the tobacco industry, wherein the machine has at least one automation device (21) for controlling, influencing and / or monitoring the machine, which automation device has at least one multicore processor (26) having a plurality of processor cores (22 a - 22 d), in particular as part of an industrial PC of the automation device (21), to which calculation tasks of the automation device (21) are distributed and are executed thereby. The method is characterized in that the same or the identical time slots (time periods) which are arranged in temporal succession, in particular each directly succeed one another, are preferably permanently assigned to each processor core (22 a - 22 d) of those processor cores (22 a - 22 d) to which these calculation tasks are distributed, in each case in particular with regard to the starting time, end time and duration, in which time slots the processor cores each execute the calculation tasks assigned to each of them in succession, preferably using a predetermined timing scheme (timing diagram).

Description

[0001] Machine and method for operating the same

[0002] Description

[0003] The present invention relates to a method for operating a machine, in particular one that operates intermittently, in particular a packaging machine for packaging products, such as products of the tobacco industry. The machine comprises at least one automation device for controlling, influencing, and / or monitoring the machine, which has at least one multi-core processor with multiple processor cores, in particular as a component of an industrial PC of the automation device, to which the computing tasks of the automation device are distributed and processed. The invention further relates to a machine operated using this method.

[0004] In packaging machines, the control of various components is usually carried out by automation devices that include a software-based industrial PC (IPC). As the demands on packaging machines increase (e.g., modularity, faster execution times, Industry 4.0), so do the demands on the computing power of the IPCs. It is known that higher computing power can be achieved by multiplying the number of CPUs or processor cores (multi-core processors), whereby the computational tasks assigned to the multi-core processor or that it must perform are distributed among the individual processor cores.

[0005] Coordinating this distribution presents a particular challenge.

[0006] The object of the present invention is to further develop the method of the aforementioned type and the aforementioned machine. This object is achieved by a method having the features of claim 1 and by a machine having the features of claim 16.

[0007] The method according to the invention is characterized in that each processor core of those processor cores to which these calculation tasks are distributed is assigned, in particular with regard to start time, end time and duration, the same or identical time slots (time periods) arranged one after the other in time (in particular in a row), preferably immediately following one another, in particular dynamically, in which they each successively process the calculation tasks assigned to them, preferably using a predetermined timing scheme (timing diagram).

[0008] This makes it possible, among other things, to develop the procedure for operating the machine in a secure, real-time manner.

[0009] Preferably, the respective duration (and / or start time and / or end time) of the machine's time slots is determined dynamically during operation, preferably depending on the respective actual or actual calculation times and / or predicted calculation times of the respective calculation tasks executed by the processor cores. This is particularly based on the calculation time required by the processor core that requires the longest time of all considered processor cores to process its calculation tasks in the corresponding time slot.

[0010] For example, it can be provided that the (further) time slot following (in the series) a previous time slot in which calculation tasks are currently being processed is only released for its calculation tasks when the last calculation task of all calculation tasks to be processed in the previous time slot by the processor cores in question has been processed; or in other words, when the last calculation task of the processor core of the processor cores that takes the longest to complete its calculation tasks in the previous time slot has been processed. It can further be provided that the time slots repeat cyclically one after the other within a predetermined total time period, in particular within a predetermined total time period that is shorter than the clock time of the multi-core processor / processor cores, in order to ensure that the corresponding calculations and, if applicable,Output data / output signals are generated or transmitted in a timely manner within one clock cycle. The clock cycle of the multi-core processor(s) can be, for example, 1 ms.

[0011] The individual time slots preferably follow one another in a series in such a way that a time slot following one another in the series starts immediately, in particular with no or only minimal time offset, as soon as the previous time slot ends.

[0012] The processor cores to which the calculation tasks are distributed (this may, but need not be, all of the processor cores of the multi-core processor) process their respective calculation tasks in the manner mentioned, each (in time) in parallel within the same time slot of the several assigned time slots, in particular in a correspondingly defined and / or coordinated manner.

[0013] Preferably, all processor cores start their calculation tasks together automatically immediately at the start time of a respective time slot.

[0014] However, it is conceivable that within the respective time slot in which the processor cores are currently processing their calculation tasks, one, several or all of the processor cores may already complete their respective calculation tasks before the end time of the time slot is reached.

[0015] Further features of the present invention emerge from the appended patent claims, the following description of embodiments and the appended drawings.

[0016] Therein shows: Fig. 1 a section of a packaging machine operated or controlled according to the invention for producing packs for tobacco products,

[0017] Fig. 2 is a side view of the packaging machine along the viewing direction II in Fig. 1 and a schematically illustrated IPC with a multi-turn processor comprising

[0018] Automation device that controls the packaging machine,

[0019] Fig. 3 shows a timing scheme which is used in particular in the normal operation of a machine, in particular the packaging machine in Figs. 1 and 2, for distributing the calculation tasks to the individual processor cores of a multi-core processor,

[0020] Fig. 4 shows a timing scheme used in an initialization operation of such a machine.

[0021] Using the method according to the invention, for example, a packaging machine 10 for packaging cigarettes 11 or for producing cigarette packs containing cigarettes 11 can be operated or controlled, as shown by way of example in Figs. 1 and 2 and known to those skilled in the art in various forms (only partially shown in the figures). However, it is understood that the method according to the invention can also be applied to various other types of machines.

[0022] Fig. 1 shows part of such a packaging machine 10, in which, in a first area, cigarettes 11 are pushed or slid in groups and in cycles from a cigarette magazine 13 into machine elements held in precise positions, moved (further) in cycles by a drive C in the machine cycle and in this case designed as pushers 12, of a circulating pocket conveyor 15, in this case designed as pockets 14. The pocket conveyor 15 then conveys the cigarette groups in cycles towards a circulating driver conveyor 16, which has individual machine elements moved by a drive B and in this case designed as drivers 17, which move on a conveyor track running transversely to the conveying plane of the pockets 14.

[0023] In a transfer area 18, in which the conveying or movement path of the pockets 14 and the conveying or movement path of the carriers 17 intersect, an upper carrier part of one of the carriers 17, adapted to the inner contour of the respective pocket 14, is moved lengthwise through a pocket 14 held ready in each case and in the process conveys the group of cigarettes located in the pocket 14 (taking them along) transversely to the conveying plane of the pockets 14 out of the pocket 14 in the direction of a subsequent winding station 19, in which an inner blank 20 is then placed on the group of cigarettes and folded around it.

[0024] Drives A, B, and C can be individually controlled and, in this case, are each designed as servo drives with a servo motor and corresponding position control. The control of the illustrated movements, or rather, of drives A, B, and C, as well as various other components of machine 10, is carried out in the example mentioned via an automation device 21, which in this case, for example, comprises or consists of an industrial PC (IPC).

[0025] For the purposes of this application, the term "automation device" encompasses all automation devices and devices comprising such automation devices, each of which is suitable for controlling, influencing, and / or monitoring machines, each of which can be used in automation solutions and is known to the person skilled in the art in the field of automation technology and mechanical engineering at the time of filing. In particular, these may be controllers such as programmable logic controllers, process computers, (industrial) computers, and the like, drive controllers, frequency converters, and the like, as they are used or can be used to control, regulate, and / or monitor technological processes, e.g., for transforming or transporting material, energy, or information, etc., whereby energy is applied or converted, in particular, via suitable technical devices, such as sensors or actuators.The control of a machine, such as the packaging machine 10, which is to be operated using the method according to the invention, is preferably carried out by means of machine software executed on a computing unit of such an automation device of the machine, for example, on the IPC in the present example. As the complexity of the respective machine and its components to be controlled increases, so does the computing power required by the computing unit.

[0026] For this reason, multi-core processors are now being used for the computing units of automation systems, on which the machine software runs and which generally process the calculation tasks of the automation systems. It is well known that the upcoming calculation tasks are distributed among the processor cores.

[0027] In the context of this application, all computational steps that such a multi-core processor or its individual processor cores process are collectively referred to as "computational tasks".

[0028] Calculation tasks can, for example, include the recording or collection of actual machine signals or, where appropriate, derived from them, particularly current machine data from the machine's hardware. This includes, in particular, the collection of actual machine data, which can originate from sensors on the machine, among other things, but also from other hardware units on the machine that generate machine data. The sensors can be any type of sensor. For example, encoder signals from encoders assigned to the actual positions of drives, such as drives A, B, or C.

[0029] Calculation tasks can also include the validation of such machine data, such as comparing this machine data with stored target values ​​or similar. Consistency checks can also be performed to verify whether the machine data is consistent and / or logical and can serve as a basis for subsequent calculations.

[0030] Calculation tasks can also include the further processing of this machine data. For example, using formulaic relationships to calculate results based on the validated data. For example, from validated actual positions of drives A, B, and C, target values ​​for drives A, B, and C can be calculated, or general control signals / control data can be generated, based on which drives A, B, and C or other processes are then moved or controlled. However, further processing of the machine data can also include, for example, detecting inconsistencies in the machine data or calculating inconsistent data based on this machine data. In such a case, corresponding error data / error signals are generated.

[0031] Calculation tasks can further include the output of the machine data further processed into control data, etc., to controllable elements of the machine hardware, such as drives A, B, C, particularly as control signals. In the above-described case of further processing the machine data into error data or error signals, the calculation tasks could include the output of this data as alarm or warning signals, for example, to a display or output device, such as a screen or the like.

[0032] The described distribution of the calculation tasks on the multi-core processor among its individual processor cores results in a corresponding proportional reduction in the total computing time required.

[0033] According to the invention, it is now particularly important that each processor core of the processor cores to which the calculation tasks are distributed is assigned the same or identical time slots (time periods) arranged one after the other, in particular directly following one another, in particular with regard to start time, end time and duration, in which they each successively process the calculation tasks assigned to them.

[0034] The time slots are preferably part of a timing scheme for the processor cores. This enables parallel processing of the computational tasks, particularly in real time. To ensure real-time capability, it can be provided, in particular, that the time slots must repeat cyclically one after the other within a predetermined total time period, preferably during normal operation of the machine, in particular within a predetermined total time period that is shorter than the clock time of the multi-core processor(s).

[0035] It would be conceivable to provide for or to define the parameters of the time slots, in particular the number and their total duration, such that the time slots in total may not take up or comprise more than 80%, preferably not more than 70%, of the clock time or cycle time of the multi-core processor or processor cores.

[0036] For the normal operation of a machine, in particular in cycles, a timing scheme 23 for four processor cores 22 ad of a multi-core processor 26 is shown in Fig. 3.

[0037] At the beginning of a cycle or clock, a first time slot 1 starts automatically, in which all processor cores 22 ad then process their calculation tasks 24.

[0038] Then, time slot 2 starts immediately after time slot 1, then time slot 3 starts immediately after time slot 2, and finally time slot 4 starts immediately after time slot 3.

[0039] The respective start time and / or the respective end time and / or the respective length of each of the time slots 1-4 are determined dynamically for each processor core 22 ad, as already described above, preferably during operation of the machine 10 or the automation device 21. Although, at least theoretically, a one-time static predetermination is also conceivable.

[0040] The dynamic determination can be made depending on the respective actual or actual calculation times and / or predicted calculation times of the respective calculation tasks executed by the processor cores. This is particularly based on the calculation time required by the processor core that requires the longest time of all considered processor cores to process its calculation tasks in the corresponding time slot. In the present example, the processor cores 22 ad process the calculation tasks described above in time slot 1 (collection time slot), which include collecting, in particular, current machine data from the machine hardware.

[0041] In time slot 2 (validation time slot), the processor cores 22 ad process the calculation tasks described above, which include validating the machine data collected in time slot 1.

[0042] In time slot 3 (further processing time slot), the calculation tasks described above are processed, which include the further processing of the collected machine data.

[0043] In time slot 4 (output time slot), the calculation tasks described above are processed, which include the output of machine data further processed in the further processing time slot.

[0044] As can be seen in each case, see, for example, the processor cores 22 c and 22 d and their calculation tasks 24 in time slot 1, in the event that calculation tasks within a time slot are completed before the specified end time of the time slot, an idle process can be provided and / or inserted to fill the time until the end time is reached (naturally independent of the example in Fig. 3).

[0045] Furthermore, it can be provided that one of the processor cores of the multi-core processor of a machine - Main Dispatcher -, preferably a predetermined processor core of the multi-core processor that does not change during operation (in Fig. 3 the processor core 22a), distributes the calculation tasks that the processor cores of the automation device have to process among the several processor cores of the multi-core processor and / or initiates a simultaneous release of the individual time slots, so that the individual processor cores can start their calculation tasks in the respectively released time slot.

[0046] It is also conceivable that the multi-core processor has several such main dispatchers, each of which fulfills the aforementioned distribution or release tasks, with each of these main dispatchers being assigned its own group of processor cores for which it fulfills these tasks. In this context, it is also conceivable that in this case, one of the processor cores of the multi-core processor, which may also be one of the aforementioned main dispatchers, assumes the role of a higher-level main dispatcher, to which the other main dispatchers are assigned, and assumes correspondingly higher-level distribution, release, and / or coordination tasks (for / between the assigned main dispatchers).

[0047] The main dispatcher, or any main dispatcher, can generally ensure that the correct programs / program parts are executed on the correct processor core and at the correct time. Thus, the main dispatcher serves as a higher-level coordination point. The main dispatcher can distribute or assign individual calculations to the processor cores using various approaches:

[0048] For example, if the computing effort and the computing capacity required for this are already known before the machine is started, a defined distribution or allocation to individual processor cores can be determined in advance.

[0049] Accordingly, time-critical, complex calculations could not be assigned to all processor cores of the multi-core processor, but rather to a subgroup (fixed number) of processor cores in order to ensure their timely processing.

[0050] An automatic or dynamic allocation of calculations to individual cores based on their utilization is also conceivable. This would involve real-time recording of the utilization of the individual processor cores, and the allocation of calculations would be automatic and variable based on their available capacity.

[0051] Another possible distribution approach would be to use artificial intelligence to sub-distribute the computational tasks among the individual cores. This would enable automatic assignment of computational tasks based on trained models. Regarding the timing of a dynamic distribution of computational tasks among the individual processor cores during operation, this could, for example, be reset after the expiration or completion of a complete clock cycle of the multi-core processor(s) or after the completion of all time slots. However, it is also conceivable to provide for a redistribution for the subsequent time slot after the completion or completion of a previous time slot.

[0052] It can further be provided that the other processor cores ultimately dependent on the main dispatcher—sub-dispatchers—in Fig. 3, processor cores 22b-d, report back to the main dispatcher when their computational tasks have been completed, and that the main dispatcher only releases the next time slot when the last sub-dispatcher has reported the completion of its computational tasks. If this last processor core or sub-dispatcher fails to do so by the end of the corresponding time slot, an error message could be generated and / or the machine could be stopped.

[0053] Fig. 3 also shows a special "message exchange." For example, if data required by multiple (software) components of the machine is generated within one of the described time slots 1-4, a hierarchically independent distribution of this data can be provided in the form of a comprehensive message exchange—global messaging. This message exchange is preferably executed on the main dispatcher and is also released by the main dispatcher once per clock cycle after all time slots have been processed.

[0054] For example, to enable software components of individual sub-dispatchers to access this global messaging system, a message adapter can be implemented as a program component in each software component of the sub-dispatcher. This message adapter can log on to and / or off from global messaging at any time. Each message adapter can have its own inbox and outbox, which can be configured as a ring buffer. A message is to be transmitted by a message adapter, it is stored in the outbox. If a message is to be received by the message adapter, the message adapter must first "subscribe" to the message, i.e., it must inform global messaging which data is necessary for it and its further calculations. The inbox can then be checked cyclically for messages.

[0055] Global Messaging cyclically scans all outboxes of the registered message adapters. If it finds a message, it checks all other message adapters to see if they are subscribed to the message. If so, it adds the message to the message adapter's inbox.

[0056] Through this higher-level exchange of data, direct addressing (what to whom) of each software component is no longer necessary, as is required in the current state of the art. Furthermore, unnecessary data transfer can be prevented because if no software component subscribes to the data message, the main dispatcher will not deliver such data messages independently (no passive communication). This, in turn, saves computing power and time, thus ensuring faster processes.

[0057] While processing computational tasks, the individual processor cores must access memory areas in one or more of their assigned data memories, for example, to copy, synchronize, and / or exchange data. Especially when the same data is used across multiple processor cores or when dependent data is used mutually by different processor cores, this data should preferably be protected from simultaneous access by the processor cores.

[0058] This principle of securing memory space ensures the integrity of the data. This integrity is especially necessary for the aforementioned cross-processor core data sharing. For example, when a first processor core processes computational tasks that are required by another processor core for its own computational tasks. This protection prevents overlaps during the processing of the (divided) computational tasks, preventing data that has not yet been fully computed from being reused prematurely and leading to incorrect computational results.

[0059] To ensure data integrity, it is generally intended that memory areas of a data memory that a first and a second processor core can or must access when executing their computing tasks are protected from simultaneous access by both processor cores. This is achieved, in particular, by locking the memory area for the second processor core for the duration of the first processor core's access if the second processor core accesses later than the first processor core or at the same time. The lock is released after the first processor core's access has been completed.

[0060] Furthermore, it can be provided that, in order to ensure data integrity, memory areas of a data memory in which data are located that depend on calculations of a first processor core are locked during these calculations of the first processor core before access by a second processor core, and that this lock is released when the calculations are completed.

[0061] In this case, it can be provided that, in the event that the data area is blocked due to access by the first processor core, the second processor core performs a predetermined number of further access attempts. In the event that the data area is blocked due to access by the first processor core, it can then be provided that an error signal is generated and / or the machine is stopped if the second processor core has performed a predetermined number of unsuccessful access attempts during the blocking.

[0062] Another special feature of the invention concerns the initialization or setup of a machine outside of normal operation, which usually takes place before the latter.

[0063] For such an initialization (for example during initial commissioning or recommissioning of the machine), in particular before normal operation of the machine, an initialization operation which is in particular not repeated cyclically can be provided, which has a time slot - initialization time slot -, in Fig. 4 time slot I, in which calculation tasks are processed which serve to set up or configure the machine.

[0064] The calculation tasks can in particular comprise the collection of actual machine data of the machine hardware, preferably actual machine data originating from sensors of the machine or from other hardware units of the machine that generate machine data, and / or the validation of collected machine data, and / or the further processing of validated machine data and / or the output of further processed data to controllable elements of the machine hardware.

[0065] List of reference symbols

[0066] 10 packaging machines

[0067] 11 cigarettes

[0068] 12 kickers

[0069] 13 cigarette magazine

[0070] 14 pockets

[0071] 15 pocket conveyors

[0072] 16 flight conveyors

[0073] 17 drivers

[0074] 18 Takeover area

[0075] 19 changing station

[0076] 20 Internal cutting

[0077] 21 Automation device

[0078] 22 a processor core

[0079] 22 b processor core

[0080] 22 c processor core

[0081] 22 d processor core

[0082] 23 Timing scheme

[0083] 24 calculation tasks

[0084] 25 Idle process

[0085] 26 multi-core processor

Claims

Patent claims 1. Method for operating a machine, in particular a packaging machine for packaging products such as tobacco products, which operates in a cycle-by-cycle manner, wherein the machine has at least one automation device (21) for controlling, influencing and / or monitoring the machine, which has at least one multi-core processor (26) with several processor cores (22a - 22d), in particular as part of an industrial PC of the automation device (21), to which the computational tasks of the automation device (21) are distributed, characterized in that each processor core (22a - 22d) of those processor cores (22a - 22d) to which these computational tasks are distributed is assigned the same or the same parameters, in particular with regard to start time, end time and duration.the identical, temporally sequential, in particular immediately consecutive time slots (time periods) are assigned, in which they each successively process the calculation tasks assigned to them, preferably using a predetermined timing scheme (timing diagram).

2. Method according to claim 1, characterized in that the computational tasks which the processor cores (22 a - 22 d) process in a first time slot - collection time slot - comprise the collection of, in particular, current machine data from the machine hardware, especially actual machine data originating from sensors of the machine or from other machine data generating hardware units of the machine.

3. Method according to claim 2, characterized in that the computational tasks performed by the processor cores (22 a - 22 d) in a second time slot - validation time slot - include validating the collected machine data.

4. Method according to claim 2 or 3, characterized in that the computation tasks which the processor cores (22 a - 22 d) perform in the time slot of a further time slot, in particular immediately after the validation time slot Processing time slots - further processing time slots - includes the further processing of the collected machine data, in particular the further processing into control signals for the machine.

5. Method according to claim 4, characterized in that the computational tasks which the processor cores (22 a - 22 d) perform in a further time slot - output time slot - which starts in particular immediately after the further processing time slot, include the output of machine data processed further in the further processing time slot, in particular to controllable organs of the machine and / or to display devices of the machine.

6. Method according to one or more of the preceding claims, characterized in that, in particular during normal operation of the machine, the time slots are cyclically repeated one after the other within a predetermined total time period, in particular within a predetermined total time period which is less than the clock time of the multi-core processor.

7. Method according to claim 6, characterized in that for the initialization of the machine, in particular before normal operation thereof, an initialization operation is provided which is not cyclically repeating and which has a time slot - initialization time slot - in which computational tasks are processed which serve the setup or configuration of the machine, in particular comprising the collection of actual machine data of the machine hardware, preferably of actual machine data which originate from sensors of the machine or from other hardware units of the machine generating machine data and / or the validation of collected machine data and / or the further processing of validated machine data and / or the output of further processed data to controllable bodies of the machine hardware.

8. Method according to one or more of the preceding claims, characterized in that the duration of each time slot and / or its respective end time is dynamically determined during operation of the automation device, in particular according to the calculation time that the The processor core required is the one that needs the longest time of all processor cores to complete its calculation tasks in the corresponding time slot, and / or in particular such that a subsequent time slot is only released when the last of the calculation tasks of the processor core that needs the longest time for its calculation tasks has been completed in the preceding time slot.

9. Method according to one or more of the preceding claims, characterized in that, in the event that computation tasks within a time slot are completed before the specified end time of the time slot, an idle process is provided and / or inserted to fill the time until the end time is reached.

10. Method according to one or more of the preceding claims, characterized in that, to ensure data integrity, a memory area of ​​a data storage device, which a first and a second processor core (22 a - 22 d) can or must access in the course of processing their computational tasks, is protected from simultaneous access by both processor cores (22 a - 22 d), in particular by locking the memory area for the second processor core (22 a - 22 d) for the duration of the access by the first processor core (22 a - 22 d), if the access by the second processor core (22 a - 22 d) occurs later than the access by the first processor core (22 a - 22 d) or at the same time, wherein the locking is lifted after completion of the access by the first processor core (22 a - 22 d).

11. Method according to one or more of the preceding claims, characterized in that, to ensure data integrity, memory areas of a data storage device containing data that depend on calculations of a first processor core (22 a - 22 d) are locked during these calculations of the first processor core (22 a - 22 d) from access by a second processor core (22 a - 22 d), and that this lock is lifted when the calculations are completed.

12. Method according to claim 11, characterized in that in the event of the data area being blocked due to access by the first Processor core (22 a - 22 d) the second processor core (22 a - 22 d) performs a predetermined number of further access attempts.

13. Method according to claim 11 or 12, characterized in that, in the event of the data area being locked due to access by the first processor core (22 a - 22 d), an error signal is generated and / or the machine is stopped if the second processor core (22 a - 22 d) has performed a predetermined number of unsuccessful access attempts during the locking process.

14. Method according to one or more of the preceding claims, characterized in that at least one of the processor cores (22 a) - Main Dispatcher -, preferably a predetermined processor core (22 a) of the multi-core processor (26) that does not change during operation, distributes the computational tasks that the processor cores (22 a - 22 d) of the automation device (21) have to process among the multiple processor cores (22 a - 22 d) of the multi-core processor (26) and / or initiates a simultaneous release of the individual time slots, so that the individual processor cores (22 a - 22 d) can start with their computational tasks in the respective released time slot.

15. Method according to one or more of the preceding claims, characterized in that the distribution of the computational tasks to the individual processor cores (22 a - 22 d) is carried out according to a defined distribution scheme, in particular one stored in a data memory, or that the distribution of the computational tasks to the individual processor cores (22 a - 22 d) is carried out depending on the current utilization of the processor cores (22 a - 22 d), or that the distribution of the computational tasks to the individual processor cores (22 a - 22 d) is carried out using an artificial intelligence method.

16. Machine with an automation device (21) for controlling, influencing and / or monitoring the machine, which has at least one multi-core processor (26) with several processor cores (22 a - 22 d), in particular as part of an industrial PC of the automation device (21), to which the computational tasks of the automation device (21) are distributed, wherein the machine is designed and configured such that it The method can be carried out according to one or more of the preceding claims. *****

Images