Blockage detection

Abstract

The invention relates to a method for detecting and evaluating an event during an operation of a sorting installation, wherein the method steps are carried out on a monitoring system for monitoring machine parameters and / or material data streams of the sorting installation such that sensor data relating to assemblies and / or sensor data relating to machines and / or sensor data relating to machine components and / or sensor data relating to material streams are continuously detected (2) by the monitoring system; and such that when the sensor data of at least one sensor has fulfilled a detection criterion (5) defined for the respective sensor, in particular a defined threshold value, a time stamp is set in the memory, an event criterion is created, the existing sensor data remains stored so as to be linked to the event criterion, and a pattern detection routine is started, wherein by means of the pattern recognition routine, the sensor data detected starting from the time stamp in a time interval with a specified duration is evaluated such that a confidence value (5) is repeatedly determined and / or such that, during the time interval, the sensor data and / or the currently determined confidence value (5) is used to decide that a fault indication is to be assigned to the event criterion and is transmitted (13, 15, 16) to at least one fault location or that an event notification is to be assigned to the event criterion and is transmitted to a notification location and the pattern recognition routine is executed again or that the pattern recognition routine is terminated (14).

Description

[0001] BLOCKAGE DETECTION

[0002] The present invention relates to the technical field of sorting systems for solid waste materials.

[0003] Sorting plants for solid waste materials usually consist of several process steps, with the process steps being interconnected via conveyor technology.

[0004] Solid waste materials can include, for example, household waste, dry recyclable mixtures, commercial waste, lightweight packaging, plastic bottles, films, paper, cardboard, electronic waste, mixed construction material, bulky waste, wood, scrap metal, or metal-containing waste.

[0005] The process steps can include, for example, comminution, dosing, sieving, screening, magnetic separation, eddy current separation, sensor-based sorting or manual sorting.

[0006] Conveyor technology can include, for example, belt conveyors, chain conveyors, trough conveyors, acceleration belts, vibratory conveyor technology or chutes.

[0007] Particularly due to their lack of bulk material properties, solid waste materials are prone to malfunctions such as material blockages. Blockages or entanglements can occur within the process steps and during material transport. In the case of material blockages, particles in the material flow become wedged together and / or against the machinery, thus impeding or completely blocking the material flow. Often, this initial material blockage intensifies over time until the material flow ceases entirely, the machinery becomes mechanically blocked, or manual intervention by operating personnel, such as a shift supervisor stopping the sorting system, results in an unplanned plant shutdown.

[0008] Due to limited input bunker sizes and / or uneven material deliveries, plant operators may attempt to compensate for downtime caused by material blockages by increasing throughput to prevent the input bunker from overfilling. This can lead to increased stress on process and conveying equipment, higher operating costs, reduced purity and / or lower material yield, and / or a higher probability of new material blockages forming.

[0009] Sorting systems consist of numerous process steps and individual components, such as conveyor units. Operating personnel can therefore often only detect malfunctions after a considerable delay, for example, only when there has been a significant impairment of the sorting system's performance or a complete shutdown, such as due to an overload of a process or conveyor unit. It is important to note that material blockages accumulate over time and are therefore increasingly difficult to remove manually the later they are detected. Since removing the blockage can often only be done while the system and / or subprocesses are shut down, this leads to further downtime and losses for the system operator. Furthermore, the time required to remove the blockage can sometimes increase faster than the detection delay can be minimized.This means that the cleaning time can be reduced by five to ten minutes, depending on the throughput, for example, if the blockage is detected one minute earlier. Since the blocking material is usually loosened but not removed from the processing line, clearing a blockage that is detected too late poses the risk of further material blockages in subsequent units.

[0010] Disadvantages of the state of the art include, among other things, unplanned shutdowns of the entire plant and / or sub-processes and thus reduced availability of the process steps or the entire plant, increased wear and / or damage to process technology or conveyor technology, impairment of the sorting performance of the sub-processes and / or the entire plant due to the disruption, as well as increased operating costs and / or lower revenues due to these aspects.

[0011] The technical objective of the invention is to improve the state of the art and to offer alternatives to it.

[0012] The technical problem is solved by an object with the technical features according to the independent claim. Advantageous embodiments are the subject of the dependent claims, the description, and the figures.

[0013] All features described in connection with individual embodiments of the invention can be provided in different combinations in the invention in order to simultaneously achieve their advantageous effects. The scope of protection of the present invention is defined by the claims and is not limited by the features described in the description or illustrated in the figures.

[0014] For the purposes of the invention, a fault location can be understood, for example, as a control room of the sorting system. A fault location can also be understood as a location within the sorting system itself and / or a location on an affected and / or another individual component of the sorting system. For example, a corresponding programming interface can also be understood as a fault location.

[0015] In accordance with the invention, a recognition criterion can be fulfilled, for example, due to exceeding a threshold value based on the sensor data and / or, for example, due to exceeding a statistical parameter calculated from the sensor data and / or, for example, due to exceeding a mathematical expression calculated from the sensor data and / or, for example, due to a pattern recognition routine.

[0016] According to one aspect, the technical problem of the invention is solved by a method for detecting and evaluating an event during the operation of a sorting system, wherein the method steps are carried out on a monitoring system for machine parameters and / or material data streams of the sorting system such that the monitoring system continuously acquires sensor data from aggregates and / or sensor data from machines and / or sensor data from machine components and / or sensor data from material streams, such that, after the sensor data of at least one sensor fulfills a detection criterion defined for the respective sensor, in particular a defined threshold value, a timestamp is set in the memory, an event criterion is created, the existing sensor data remain stored assigned to the event criterion, and a pattern recognition routine is started.wherein the pattern recognition routine evaluates the sensor data acquired from the timestamp onwards within a time interval of a predetermined duration in such a way that a confidence value is repeatedly determined and / or that, within the time interval, decisions are repeatedly made based on the sensor data and / or on the currently determined confidence value, such that a fault message is assigned to the event criterion and this is sent to at least one fault reporting point, or that an event message is assigned to the event criterion and this is sent to a reporting point and the pattern recognition routine is executed again, or that the pattern recognition routine is aborted.

[0017] This allows for the early and, in particular, indirect detection of malfunctions such as material blockages and / or belt misalignment and / or an impending defect in a machine component and / or an upcoming need for maintenance.

[0018] For example, software monitors system parameters, such as motor load, which would increase in the event of a material blockage. Such characteristic behavior can be detected, for example, using pattern recognition, and countermeasures can be initiated early.

[0019] In a technically advantageous embodiment of the method, the pattern recognition routine includes the creation of a first statistical indicator of the sensor data over a first period, as well as a comparison with a second statistical indicator over a second period, wherein in particular the second period and the first period refer to the same or different time intervals.

[0020] In particular, it can be stipulated that the sensor data does not always have to be assigned to a specific machine. The initial detection of an event can be based on a variety of methods. In the simplest case, this can be a threshold value, but more complex algorithms can also be easily integrated.

[0021] Furthermore, in a technically advantageous embodiment of the method, it is provided that the sensor data, machine data, in particular of individual processes, and / or material data, in particular of individual processes, and / or conveying technology data, and / or level sensor data, and / or vibration sensors, and / or temperature sensors, and / or inline sensor data, and / or online sensor data, and / or imaging inline sensor data, and / or online sensor data, and / or event data, and / or error data, and / or warning message data from the plant control system are acquired.

[0022] Conveyor technology data can include, for example, data from a motor driving the conveyor system. Such motor data can include, for example, data on motor load, motor torque, motor current, motor temperature, motor speed, conveying velocity, conveying direction, sensor data from the motor fan, error messages, and / or warning messages.

[0023] Advantageously, this allows material blockages or maintenance needs to be detected indirectly, for example based on increased motor load data of the conveyor technology, and various detection methods can be created depending on the application requirements.

[0024] Furthermore, in a technically advantageous embodiment of the method, it is provided that, for the pattern recognition routine, the sensor data, in particular as sensor data streams, are evaluated individually and / or in combination using algorithms and / or models, wherein in particular an algorithm and / or a model used supplements predictions with a statement of the confidence value and / or with the statement of meta-information, wherein the meta-information is derived in particular by means of the acquisition of the sensor data of the machine components or material flows.

[0025] Advantageously, various methods, for example from the field of signal processing and data processing, as well as from the field of data science, can be used to improve the process.

[0026] The term meta-information can include, for example, machine parameters and / or material parameters and / or temporal information.

[0027] Furthermore, in a technically advantageous embodiment of the method, it is provided that measures to remedy the fault are initiated by means of the fault notification, whereby in particular operating personnel, preferably operating personnel who are in close proximity to the fault, are informed.

[0028] For example, it may be planned that operating personnel are informed via various communication channels, either individually or in combination. These communication channels may include messages via a process control system and / or a separate display, such as a computer program or a website. Visual warning signals, such as a warning light and / or a signal light, and / or audible warning signals, such as an announcement via a loudspeaker system and / or a radio message, may also be provided. Messages via a pager and / or a smartwatch may also be provided. Notification on a mobile device, such as a smartphone or tablet, may also be provided. Wayfinding systems or lighting systems may also be used as communication channels.

[0029] This can advantageously enable an efficient and quick resolution of the fault.

[0030] Furthermore, in a technically advantageous embodiment of the method, it can be provided that the probability of occurrence and / or the meta-information are taken into account when sending targeted notifications and / or initiating countermeasures, in particular by means of operating personnel and / or automatically by means of the plant control system, and in particular that fault messages are prioritized.

[0031] This allows for an appropriate response to a disruption, such as a material blockage, and enables the prioritization of multiple simultaneous disruptions. Particularly critical disruptions, for example those with a high confidence level, can be addressed first. For instance, the system can be configured to issue warnings tailored to the confidence level.

[0032] Furthermore, in a technically advantageous embodiment of the method, it may be provided that the fault message includes information regarding the time of the fault and / or the location of the fault.

[0033] For example, information about the time and / or location of the malfunction can be advantageously provided by specifying a position identifier and / or a machine identifier and / or a general identifier. This information can also include local coordinates and / or descriptions and / or circumlocutions and / or directional guidance and / or a symbol and / or be visual, using two-dimensional or three-dimensional representations or by marking a representation of the sorting system, such as a conveyor belt or a two-dimensional or three-dimensional visualization of the system.

[0034] Furthermore, in a technically advantageous embodiment of the method, it can be provided that measures to rectify the fault are initiated by means of the fault message, wherein in particular mechanical actuators within an individual process are triggered to resolve the fault and / or wherein in particular mechanical actuators within the conveying technology are triggered to resolve the fault and / or wherein in particular process parameters of individual processes are changed and / or wherein in particular transport parameters of the conveying technology are changed and / or wherein in particular material flow within the sorting system is changed.

[0035] For example, mechanical actuators can be triggered by moving components such as flaps, stamps or nozzle bars, or by adaptively controlling conveyor technology.

[0036] For example, it may be possible to change the process parameters by: changing the air classifier speed and / or the screen drum speed and / or activating a nozzle bar and / or activating sensor-based sorting systems and / or changing the direction of rotation and / or the speed of dosing technology and / or comminution technology and / or changing the speed and / or direction of rotation and / or the angle of a ballistic separator.

[0037] For example, to change the transport parameters, it may be necessary to reverse conveyor belts and / or change a conveying speed and / or a vibration intensity and / or a vibration frequency.

[0038] To change the material flow, for example, individual process steps can be bypassed and / or a division across several process lines can be planned.

[0039] Furthermore, in a technically advantageous embodiment of the method, it may be provided that the sensor data are temporarily stored in a memory in the manner of a shift register or are permanently stored in a memory.

[0040] Furthermore, in a technically advantageous embodiment of the method, it may be provided that the sensor data stored before the timestamp are evaluated by means of the pattern recognition routine in such a way that a multi-level confidence value, in particular a significance level, preferably a multi-level significance level, is determined.

[0041] Furthermore, in a technically advantageous embodiment of the method, motor current data, in particular motor current data of a belt conveyor, can be continuously recorded. For the purposes of the invention, "continuously recorded motor current data" can also be understood to mean that motor current data is recorded automatically at intervals and / or, in particular, in real time. Specifically, it can be understood that, without a manual request, for example by a user, changes in the measured values ​​are measured promptly or at the moment they occur, and not only at a later time or accumulated over a period of time.

[0042] This can advantageously improve the process further in order to eliminate time delays in detection beyond at least one automated cycle.

[0043] Furthermore, in an advantageous embodiment of the method, it is provided that a normal state or a target state is detected by forming a long-term average of the sensor data, and a state deviating from the normal state or the target state is detected, in particular continuously, by forming a short-term average of the sensor data and comparing it with the long-term average of the sensor data.

[0044] For the purposes of the invention, the continuous detection of a state deviating from the normal or target state can also be understood to mean that this occurs automatically at regular intervals and / or, in particular, in real time. Specifically, it can be understood that changes in the measured values ​​are measured promptly, i.e., at the moment they occur, without a manual request, for example, by a user, and not only at a later time or cumulatively over a period of time.

[0045] The advantages already mentioned also apply here, and the method can be further improved because average values ​​from different time periods replace individual measured values, which can filter out particularly short-term fluctuations.

[0046] Furthermore, in an advantageous embodiment of the method, a confidence value, in particular a confidence value table, is generated by calculating the difference between sensor data, specifically between long-term and short-term sensor data, and the pattern recognition routine is then generated using the confidence value and / or the confidence value table. For the purposes of the invention, long-term sensor data can be understood to mean data obtained, for example, over several hours, in particular several days, preferably several weeks of trouble-free operation, or data from model data of a system operating without problems.

[0047] This allows, for example, the collection of optimal data for each plant under specific local conditions and operating requirements, which can further improve the process.

[0048] Furthermore, in an advantageous embodiment of the method, it is provided that a pre-detection of a fault corresponding to the fault message, in particular a detection during the development of the fault, is enabled preferably by means of a pre-trained pattern of the pattern recognition routine and / or by means of defined criteria, in particular by means of a confidence value table.

[0049] The advantages already mentioned still apply, and the process can be further improved.

[0050] Furthermore, in an advantageous embodiment of the method, it is provided that the method, in particular the sorting plant, is preferably exclusively suitable for processing and / or preparing waste and / or valuable materials and / or recycled materials.

[0051] Exemplary embodiments of the invention are shown in the figure and are described in more detail below.

[0052] It shows:

[0053] Figure 1 shows a schematic representation of a method for detecting and evaluating an event during the operation of a sorting plant.

[0054] Figure 1 shows a schematic representation of a method for detecting and evaluating an event during the operation of a sorting system. The method steps are executed on a monitoring system for machine parameters of the sorting system such that the monitoring system continuously acquires sensor data from machine components and / or material flows. After a detection criterion, defined for the respective sensor, is created from the sensor data (a timestamp is set in memory), an event criterion is generated, the existing sensor data is stored and associated with the event criterion, and a pattern recognition routine is started.Using the pattern recognition routine, the sensor data recorded from the timestamp onwards within a time interval of a specified duration are evaluated in such a way that, based on the currently determined confidence value, a decision is made as follows: either a fault message is assigned to the event criterion and this is sent to at least one fault reporting point, or an event message is assigned to the event criterion and this is sent to a reporting point and the pattern recognition routine is executed again, or the pattern recognition routine is aborted.

[0055] In particular, Figure 1 shows the start of the method (1), whereby machine and / or material data streams are received by way of example (2). For example, engine load data is received as sensor data from an engine load sensor via a network protocol for machine-to-machine communication. Then, a first statistical measure of the sensor data is calculated over a first period (3). Then, a second statistical measure of the sensor data is calculated over a second period (4). Finally, a recognition criterion defined for a respective sensor is updated (5).

[0056] If a detection criterion is met (6, 8), it is checked whether a continuous event is present (10). This check uses, for example, pattern recognition methods. If a continuous event is present, information indicating that a continuous and active event exists is transmitted to an application programming interface, e.g., a reporting center (8, 16), and the procedure is then terminated (9). If, on the other hand, no continuous event is present, information indicating that a new event has been detected is transmitted to an application programming interface (7, 15), and the procedure is then terminated (9).

[0057] If the recognition criterion is not met (6, 7), it is also checked whether an existing event exists (11). If not, the procedure is terminated (7, 9). If so, a pattern recognition routine determines whether the event can be classified as terminated. If not, the information that an event has ended is transmitted to a programming interface and the procedure is terminated (9). If not, information that an ongoing and inactive event exists is transmitted to a programming interface (8, 13) and the procedure is terminated (9).

[0058] The advantages of this method, which can be implemented, for example, at least partially as a computer program or computer program product, lie in the fact that, indirectly and compared to the prior art, malfunctions such as material blockages, conveyor belt misalignment, and / or impending maintenance events can be detected much earlier. Furthermore, upcoming maintenance needs can be identified early, allowing for timely preparations and thus reducing and preventing downtime.

[0059] Reference symbol list:

[0060] 1. Start of the procedure

[0061] 2. Receiving machine and / or material data streams

[0062] 3. Calculation of an initial statistical indicator from the sensor data over an initial period

[0063] 4. Calculation of a second statistical indicator from the sensor data over a second period

[0064] 5. Updating a detection criterion defined for a particular sensor

[0065] 6. Determining whether a recognition criterion is met

[0066] 7 true

[0067] 8 incorrect

[0068] 9. Ending the procedure

[0069] 10. There is an ongoing event.

[0070] 11. An existing event exists.

[0071] 12. Determining whether the event is complete

[0072] 13. Transmission of information that a continuous and inactive event is occurring to a programming interface

[0073] 14. Transmission of information that an event has ended to a programming interface

[0074] 15. Transmission of the information that a new event has been detected to a programming interface

[0075] 16. Transmission of information that a continuous and active event is occurring to a programming interface

Claims

Claims:

1. A method for detecting and evaluating an event during the operation of a sorting plant, wherein the method steps are performed on a monitoring system for machine parameters and / or material data streams of the sorting plant such that the monitoring system continuously records sensor data from aggregates and / or sensor data from machines and / or sensor data from machine components and / or sensor data from material streams (2); such that, after the sensor data of at least one sensor fulfills a detection criterion (5) defined for the respective sensor, in particular a defined threshold value, a timestamp is set in the memory, an event criterion is created, the existing sensor data remain stored assigned to the event criterion, and a pattern recognition routine is started,wherein the sensor data acquired from the timestamp onwards in a time interval of a predetermined duration are evaluated by means of the pattern recognition routine such that a confidence value (5) is repeatedly determined, and / or such that in the time interval a decision is repeatedly made on the basis of the sensor data and / or on the basis of the currently determined confidence value (5) such that a fault message is assigned to the event criterion and this is sent to at least one fault location (13, 15, 16) or such that an event message is assigned to the event criterion and this is sent to a reporting location and the pattern recognition routine is executed again or such that the pattern recognition routine is aborted (14).

2. Method according to claim 1, characterized in that the pattern recognition routine comprises creating a first statistical identifier of the sensor data over a first period (3), as well as a comparison with a second statistical identifier over a second period (4), wherein in particular the second period and the first period refer to time intervals of the same or different lengths.

3. Method according to one of the preceding claims, characterized in that during the acquisition of the sensor data (2), Machine data, in particular from individual processes, and / or material data, in particular from individual processes, and / or conveyor technology data, and / or level sensor data, and / or vibration sensors, and / or temperature sensors, and / or inline sensor data, and / or online sensor data, and / or imaging inline sensor data, and / or event data, in particular from the plant control system, and / or error messages, in particular from the plant control system, and / or warning messages, in particular from the plant control system, are recorded.

4. Method according to one of the preceding claims, characterized in that, for the pattern recognition routine, the sensor data, in particular as sensor data streams, are evaluated individually and / or in combination using algorithms and / or models, wherein in particular an algorithm and / or a model used supplements predictions with a statement of the confidence value and / or with the statement of meta-information, wherein the meta-information is derived in particular by means of the acquisition of sensor data of the machine components or material flows.

5. Method according to one of the preceding claims, characterized in that measures to remedy the fault are initiated by means of the fault report, wherein in particular operating personnel, preferably operating personnel who are in close proximity to the fault, are informed.

6. Method according to one of the preceding claims, characterized in that the confidence value and / or the meta-information are taken into account in a targeted communication and / or an initiation of countermeasures, in particular by means of the operating personnel and / or automatically by means of the plant control, and in particular the fault messages are prioritized.

7. Method according to one of the preceding claims, characterized in that the fault message includes information regarding the time of the fault and / or regarding the location of the fault.

8. A method according to one of the preceding claims, characterized in that measures to rectify the fault are initiated by means of the fault message, wherein in particular mechanical actuators within an individual process are triggered to resolve the fault and / or wherein in particular mechanical actuators within the conveyor technology are triggered to resolve the fault and / or in particular, process parameters of individual processes are changed and / or in particular, transport parameters of the conveyor technology are changed and / or in particular, material flow within the sorting plant is changed.

9. Method according to one of the preceding claims, characterized in that the sensor data are temporarily stored in a memory in the manner of a shift register or are permanently stored in a memory.

10. Method according to one of the preceding claims, characterized in that the pattern recognition routine is used to identify the data stored prior to the timestamp. Sensor data are evaluated in such a way that a multi-level confidence value, in particular a significance level, preferably a multi-level significance level, is determined.

11. Method according to one of the preceding claims, characterized in that motor current data, in particular motor current data of a belt conveyor, are continuously recorded.

12. Method according to one of the preceding claims, characterized in that a normal state or a target state is detected by forming a long-term average of the sensor data and a state deviating from the normal state or the target state is detected, in particular continuously, by forming a short-term average of the sensor data and comparing it with the long-term average of the sensor data.

13. Method according to one of the preceding claims, characterized in that a confidence value, in particular a confidence value table is formed by means of a difference value calculation of sensor data, in particular the calculation of a difference value between the long-term sensor data and the short-term sensor data, whereby the pattern recognition routine is generated using the confidence value and / or the confidence value table.

14. Method according to one of the preceding claims, characterized in that a pre-detection of a fault corresponding to the fault message, in particular a detection during the development of the fault, is enabled preferably by means of a pre-trained pattern of the pattern recognition routine and / or by means of defined criteria, in particular by means of a confidence value table.

15. Method according to one of the preceding claims, characterized in that the method, in particular the sorting plant, is preferably exclusively suitable for processing and / or preparing waste and / or valuable materials and / or recycled materials.

Images