Method and system for controlling an extrusion system for additive manufacturing
A method and system for controlling an extrusion system ( 1 ) having a plurality of nozzle heads ( 7 ) for depositing a predetermined interconnected arrangement for concurrently forming individual three-dimensional structures ( 3 ) in parallel. An operation of each of the plurality of nozzle heads is monitored in order to detect failing nozzle heads, wherein a first value is calculated which is indicative of an average print time for finishing a number of three-dimensional structures when the print job is continued using one or more non-failing nozzle heads of the plurality of nozzle heads which are still operational, and a second value is calculated indicative of an average print time for finishing the number of three-dimensional structures when the print job is promptly interrupted and a new print job is initiated with the one or more failing nozzle heads of the plurality of nozzles heads being repaired. The controller is configured to restart the print job with the one or more failing nozzle heads being repaired, when the second value is smaller than the first value.
1 . A method for controlling an extrusion system for additive manufacturing by means of an automated feedback loop that is configured to evaluate at least one detected operational failure event and based on that decide whether to continue the printing process, the extrusion system being configured to perform at least one print job in which a three-dimensional structure is manufactured in parallel, wherein the extrusion system includes a controller and an extrusion unit having a plurality of nozzle heads by means of which filaments of build material are deposited in a predetermined interconnected arrangement for concurrently forming individual three-dimensional structures in parallel, wherein each nozzle head includes one or more nozzle outlets through which the build material is extruded, wherein the method includes monitoring an operation of each of the plurality of nozzle heads during the print job by acquiring sensor data indicative of the operation of the nozzle heads in order to detect one or more operational failure events linked to one or more failing nozzle heads, wherein, in case one or more operational failure events are detected, the controller is configured to perform the steps of:
determining a first value indicative of an average print time for finishing a number of three-dimensional structures when the print job is continued using one or more non-failing nozzle heads of the plurality of nozzle heads which are still operational,
determining a second value indicative of an average print time for finishing the number of three-dimensional structures when the print job is promptly interrupted and a new print job is initiated with the one or more failing nozzle heads of the plurality of nozzle heads being repaired,
determining that the second value is smaller than the first value;
in response to determining that the second value is smaller than the first value: initiating a repair procedure on the one or more failing nozzle heads, and
automatically restarting the print job with the one or more failing nozzle heads being repaired.
2 . The method according to claim 1 , wherein the first value and second value are calculated using a statistical model.
3 . The method according to claim 2 , wherein the statistical model is configured to take into account a probability of operational failure events occurring.
4 . The method according to claim 1 , wherein the first average print time and the second average print time are calculated based on a predefined probability of an operational failure event occurring per time unit.
5 . The method according to claim 4 , wherein the predefined probability per time unit is constant.
6 . The method according to claim 4 , wherein the predefined probability per time unit is variable in function of elapsed time since a start of the print job.
7 . The method according to claim 4 , wherein the predefined probability per time unit is variable in function of types of error events.
8 . The method according to claim 4 , wherein the predefined probability per time unit is variable in function of a percentage of completion of the print job.
9 . The method according to claim 4 , wherein the predefined probability per time unit is variable in function of an amount of build material already extruded by the plurality of nozzle heads.
10 . The method according to claim 4 , wherein the predefined probability per time unit is variable in function of previously detected error events during the print job.
11 . The method according to claim 4 , wherein the extrusion unit includes more than two nozzle heads.
12 . The method according to claim 11 , wherein the extrusion unit includes more than five nozzle heads.
13 . The method according to claim 12 , wherein the extrusion unit includes more than ten nozzle heads.
14 . The method according to claim 1 , wherein the first value and the second value are at least partially calculated by means of a trained machine learning model.
15 . A non-transitory computer-readable medium for performing, when run on a controller of an extrusion system with an extrusion unit including a plurality of nozzle heads each having one or more nozzle outlets, the method according to claim 1 .
16 . An extrusion system arranged for manufacturing a three-dimensional structure by means of an automated feedback loop that is configured to evaluate at least one detected operational failure event and based on that decide whether to continue the printing process, wherein the extrusion system is configured to perform at least one print job in which a three-dimensional structure is manufactured in parallel, the system comprising:
a controller,
an extrusion unit comprising a plurality of nozzle heads by means of which filaments of build material are deposited in a predetermined interconnected arrangement such as to concurrently form individual three-dimensional structures in parallel, wherein each nozzle head includes one or more nozzle outlets through which the build material is extruded, and
a monitoring unit configured to monitor an operation of each of the plurality of nozzle heads during the print job by acquiring sensor data indicative of the operation of the nozzle heads in order to detect one or more operational failure events linked to one or more failing nozzle heads,
wherein, in case one or more operational failure events are detected by means of the monitoring unit, the controller is configured to:
determine a first value indicative of an average print time for finishing a number of three-dimensional structures in case the print job is continued using one or more non-failing nozzle heads of the plurality of nozzle heads which are still operational,
determine a second value indicative of an average print time for finishing the number of three-dimensional structures in case the print job is promptly interrupted and a new print job is initiated with the one or more failing nozzle heads of the plurality of nozzle heads being repaired,
determine that the second value is smaller than the first value,
in response to determining that the second value is smaller than the first value: initiate a repair procedure on the one or more failing nozzle heads, and
automatically restart the print job with the one or more failing nozzle heads being repaired.
17 . The extrusion system according to claim 16 , wherein the first value and second value are determined using a statistical model configured to take into account a probability of operational failure events occurring, the probability of operational failure events occurring being a predefined probability per time unit.
18 . The extrusion system according to claim 17 , wherein:
the predefined probability per time unit is constant;
the predefined probability per time unit is variable in function of elapsed time since a start of the print job;
the predefined probability per time unit is variable in function of types of error events;
the predefined probability per time unit is variable in function of a percentage of completion of the print job;
the predefined probability per time unit is variable in function of an amount of build material already extruded by the plurality of nozzle heads; or
the predefined probability per time unit is variable in function of previously detected error events during the print job.
19 . The extrusion system according to claim 17 , wherein the first value and the second value are at least partially calculated by means of a trained machine learning model.
20 . A method for additive manufacturing, the method comprising using the extrusion system according to claim 16 for manufacturing three-dimensional structures in parallel.