• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    Summary The present invention relates to a method and apparatus for predicting the state of a single machine or a component of a machine based on measurements of a state monitoring parameter. The apparatus comprises - a data store (6) for saving a movable window containing measured values for the condition monitoring parameter, - a monitoring module (7) arranged to determine when a measured value exceeds a potential damage level at which damage to the machine or component is assumed to have been initiated, and to determine the time (tp) when the monitoring parameter exceeds the potential damage level, - a data cleaning module (8) arranged to throw a new measured value if it is less than the last saved measured value in the moving window, to determine an expected value of the measurement based on a defined pattern for the condition monitoring pattern, reflects an expected occurrence of the monitoring parameter after damage has been initiated and the time elapsed since the time when the monitoring parameter exceeded the potential damage level, and to discard the new measurement value if the deviations are too large from the expected value, and to save non-discarded measured value the moving window, and - a prediction module (10) arranged to predict the level of the monitoring parameter at the future time based on the measured values stored in the moving window. (Figure 3)
    公开号:SE1300126A1
    申请号:SE1300126
    申请日:2013-02-19
    公开日:2014-08-20
    发明作者:Basim Al-Najjar
    申请人:Basim Al-Najjar；
    IPC主号:
    专利说明:

    The article "Mechanistic model for predicting the CM parameter value: a case study" by Basim Al-Najjar and lmad Alsyouf, Conference Modeling in industrial maintenance and reliability, MIMAR 2004, Salford University, US pages 7-12, shows a method for predicting condition monitoring parameters for a rotating machine based on measurements of condition monitoring parameters.The method detects when a damage to the components has been initiated.
    The time that has elapsed since the damage was initiated is estimated. A wear factor is estimated based on current machine load, future load and previous change in wear speed and the gradient of wear. A value in the near future, e.g. at the next planned stop or the next measurement, predicted for at least one CM parameter. The prediction is performed using a dynamic mathematical model shown in the article. The predicted value of the CM parameter is used to determine the condition of the machine or its components in the near future, which can be used for planning future maintenance of the machine. However, the result of the prediction depends on the measurements used as input to the dynamic model. Raw measurement data are often affected by disturbances that affect the result of the prediction. In order to achieve a satisfactory result of the prediction, it is necessary to identify a set of measured values whose values reliably represent the wear process, i.e. an injury under development. Therefore, measurement data must be processed before being used as input to the dynamic model.
    US2004 / 0098233A1 discloses a system for tracking small changes in technical processes, machines or the like based on measurement data. Measurement data is processed before it is used for tracking. Validation conditions are defined for measurement data and only measurement data that meets the validation conditions are used. A validation condition for measurement data is e.g. a value that is higher or lower than a particular measured value. The purpose of the processing of measurement data is to increase the chances of detecting small changes in the condition of the process or machine, which are otherwise difficult to detect due to. noise and other disturbances. OBJECTS AND SUMMARY OF THE INVENTION An object of the present invention is to improve the prediction of a condition monitoring parameter and thereby improve the estimate of the most efficient time to replace or perform maintenance on a machine or important components of the machine.
    According to one aspect of the invention, this object is achieved by a method as described in claim 1.
    Measurement data used for the prediction is purified in such a way that an optimal prediction is achieved. If any part of the machine is damaged, the measurements should increase because the damage to the machine component is irreversible, i.e. it cannot repair itself, which should mean that the values of the measurements cannot decrease over time. During the first phase, measurement data is purified by selecting only measurement values with increasing value, i.e. values that are greater than the values that have been selected during previous measurement occasions. In this way, measured values that are contaminated by internal and external disturbance factors are discarded. The selected values are saved in a moving window adapted to save a predetermined number of measured values. When the movable window is full, the oldest values are always discarded when a new value is saved in the movable window. In this way, the number of measured values saved always remains the same.
    The prediction begins when the value of the measured condition monitoring parameter has exceeded a potential damage level, i.e. damage has been initiated, in which it is assumed that a potential damage to the machine or component is initiated. The potential level of damage can e.g. be a predefined value based on previous measurements of the parameter. During the second phase, in addition to selecting only the measured value with increasing values, the measured values are purified by throwing the monitoring parameter. The expected value is determined based on a predefined pattern of measurement values that deviate too much from an expected value of which reflects an expected behavior of the monitored parameter after the damage has been initiated, and the time that has elapsed between the time the measurement was performed and the time of the monitoring parameter. the potential level of damage. In this way, incorrect measured values are thrown away. The measured values must thus meet two different conditions in order to be allowed to be saved in the moving window; the values must be increasing and they are not allowed to deviate too much, i.e. no more than a certain value from their expected values. This value can e.g. be a predetermined value, or it can be determined dynamically.
    The prediction is performed based on measured values that are saved in the moving window, ie. which are purified from disturbances. In this way, the prediction is improved.
    Due to the fact that measurement data is also purified during the first phase, it is possible to start the prediction based on the measurement values saved in the moving window as soon as the value of the monitoring parameter has passed the potential damage level. In this way, the prediction is based on measured values saved during the first phase as well as during the second phase.
    It is advantageous to start the prediction as early as possible in order to provide early warnings and maintenance instructions to the user.
    According to an embodiment of the invention, the movable window contains at least three saved measured values. Thanks to the purification of the measured values, the number of necessary measurements is reduced to achieve a sufficiently good prediction. Three measured values which have been purified according to the invention are sufficient to obtain a reliable result.
    According to one embodiment of the invention, the predefined pattern of the condition monitoring parameter is a curve that begins to increase at the time when the monitoring parameter exceeds the potential damage level. The curve describes the value of the monitoring parameter in relation to the time that has elapsed since the time when the monitoring parameter exceeded the potential damage level. The shape of the pattern and thus of the curve depends on e.g. on the type of parameter, the type of damage, e.g. a breakage or wear, and the type of wear process, the load and speed of the machine, and the type and quality of the machine or component. For many types of errors, the curve is exponentially increasing. The curve is determined e.g. based on the previous value of the same condition monitoring parameter for the same or similar types of machines or components. The curve begins at the same time as it has been detected that an injury has been initiated, i.e. at the time when the monitoring parameter exceeds the potential damage level, and the curve increases throughout the second phase. The expected value of the measurement is determined based on the curve and the time of the measurement. In this way, it is easy to determine the expected value of the measurement.
    According to an embodiment of the invention, the new measured value is discarded if the measured value deviates from the expected value by more than one limit value, the limit value can be predetermined or calculated dynamically during the second phase.
    According to an embodiment of the invention, the measurement method further comprises: a message to the user is generated to check the measurement and provide a new measurement value when it has been discovered that the measurement value deviates too much from the defined pattern, a new measurement value is received and the new measurement value is compared with the expected the value that has been determined in accordance with the defined pattern of the condition monitoring parameter, and a warning is generated if the new measured value also deviates too much from the expected value. Otherwise, the oldest of the saved measured values in the movable window is replaced with the new measured value. A new measurement is performed if the value of the measurement deviates too much from the expected value during the second phase. Alternatively, a new measurement value is taken automatically when it has been detected that a measurement deviates too much from the expected value. With this embodiment, an incorrect measurement value is discarded and replaced with a new wear-relevant measurement.
    If the new measurement also deviates too much from the expected value, a warning is generated to the user. This can either mean that something is wrong with the measuring equipment or that the damage has developed much faster than expected and that a maintenance operation needs to be carried out urgently.
    According to an embodiment of the invention, the method further comprises: the predicted level of the condition monitoring parameter is compared with at least one limit value of the condition monitoring parameter, the previous steps are repeated at least until the predicted level of the condition monitoring parameter exceeds the limit level and a message is generated. the monitoring parameter exceeds the limit level. The limit level is e.g. an alert level that is below the replacement level of the component or machine. It is also possible to have several boundary levels between the potential damage level and a replacement level. This embodiment makes it possible to propose a maintenance measure with sufficient lead time at a cost-effective time.
    According to an embodiment of the invention, the machine is a rotating machine and the measured values are any of vibration measurements, temperature measurements, sound measurements and shock pulse measurements.
    According to another aspect of the invention, this object is achieved by a computer software product directly downloadable in the internal memory of a computer, or by a remote computer, web service, or cloud service comprising software for performing the steps of the method when said program is run on the computer.
    According to another aspect of the invention, this object is achieved by a data readable medium having a program recorded, where the program causes the computer to perform the steps of the method when the program is run on the computer.
    According to another aspect of the invention, this object is achieved by an apparatus as defined in claim 8. General Description of the Drawings The invention will now be explained in more detail by describing various embodiments of the invention and with reference to the accompanying figures.
    Figure 1 shows an example of how measurement of a condition monitoring parameter can vary over the time before and after injury has been initiated.
    Figure 2 shows an example of how the measurement of a condition monitoring parameter can vary over time in relation to a defined potential damage level and limit levels for generating maintenance instructions.
    Figure 3 shows a block diagram of an apparatus for predicting the condition of a machine or a component of the machine according to an embodiment of the invention.
    Figure 4 shows three examples of patterns for the condition monitoring parameter, which reflect an expected behavior of the monitoring parameter after the injury has been initiated.
    Figure 5 shows a flow chart of a method for predicting the condition of a machine or a component of the machine according to an embodiment of the invention.
    Detailed Description of Preferred Embodiments of the Invention Figure 2 shows an example of how a measurement of a condition monitoring parameter can vary over time, before and after the injury has been initiated. A potential damage level Xp of a condition monitoring parameter has been defined, at which level a damage to the machine or component has been initiated. Two additional limit levels A1, A2 for generating maintenance instructions have been defined. An exchange level RL is also shown in the figure. When the replacement level approaches or is reached, the component or machine must be replaced. The number of limit levels for generating maintenance instructions may vary. In another embodiment, only one boundary level can be defined. It is also possible to have several limit levels, and to generate a special message for each limit level and component in a machine, identify the machine (if several machines are included), show the location of the component, how serious the component damage is, number of days left until the predicted level of the condition monitoring parameter, and increasing maintenance alerts.
    Figure 3 shows an example of an apparatus 1 for predicting the condition of a component 2, such as a roller bearing of a machine, according to an embodiment of the invention. The device contains hardware, e.g. a computer containing a processor, and software modules running the computer. A condition monitoring parameter, such as a vibration level, temperature level, noise level or SPM level is measured continuously or periodically by means of a sensor 3. The condition monitoring parameter is measured e.g. monthly. The measured values and information about the time when these were measured are stored in the company's database 4. The apparatus 1 contains a communication unit 5 configured to communicate with the database 4 and to receive measured values and the time for the measurement. Alternatively, the machine sends data containing measured values and the time of the measurement to the device, or the measured values and the time of the measurement are sent to the device using special files. In this case, the communication unit receives measurement data.
    The apparatus contains a data store 6 arranged to store a movable window containing a plurality of measured values for the condition monitoring parameter. The movable window is adapted to save a predetermined number of measured values, e.g. three measured values. It predicted the values of the condition monitoring parameter, the time of damage initiation, i.e. The time when the monitoring parameter exceeded the potential damage level Xp is shown in Figure 2. The time of the predicted value, the discarded values of the condition monitoring parameter and other analysis results can also be saved in the data layer 6. When the moving window is full, the oldest measured value is removed from the moving window. a new value is saved in the moving window. In this way, the number of measured values saved remains the same over time. Preferably, the movable window is adapted to save at least three measured values. The data warehouse is e.g. a register.
    The apparatus further comprises a monitoring module 7 arranged to monitor the measured values and determine which measured values exceed the potential damage level Xp, at which a damage of the component has been initiated, and to determine the time tp when the monitoring parameter exceeds the potential damage level. The monitoring module 7 compares new measured values with the potential damage level Xp, and when a measured value exceeds the potential damage level, one damage has been initiated and the second phase has been initiated. The time when the monitoring parameter exceeds the potential damage level is the time when the potential damage starts. This time can be calculated based on the measurement values and the time for the measurements.
    The apparatus further contains a data purification module 8 for purifying obtained measured values.
    The data purification module is arranged to compare a new measured value for the condition monitoring parameter with the last saved measured value in the moving window and to throw the measured value if it is less than the last saved measured value. Discarded values are not saved in the moving window. However, it is possible to save discarded values in the data store 6 for later use in performing analyzes of the measured values.
    In this way, measured values that are contaminated by internal or external disturbance factors, i.e. they are not saved in the moving window and are not used in prediction. The data purification module is arranged to save the measured values in the moving window if they are greater than the last saved measured value. The measured values are saved in the moving window until the window is full. If the movable window e.g. is adapted to save three measured values, the moving window is full when three measured values have been saved in the moving window. If the movable window is full, the oldest measured value is moved away from the movable window and instead the new value is saved in the movable window, i.e. the oldest measured value stored in the movable window is replaced by the new measured value.
    The data cleaning module is arranged to continue saving measured values in the moving window as they are larger than the last saved measured values, as long as a damage has not been initiated, i.e. as long as the monitoring parameter has not yet exceeded the potential level of damage.
    The data cleaning module is further arranged to purify measured values obtained after the time when the damage has been initiated, i.e. after the monitoring parameter has exceeded the potential damage level. If the damage has been initiated, the cleaning module is arranged to determine an expected value for the measurement based on a defined pattern of the condition monitoring parameter, which reflects an expected behavior of the monitoring parameter after the damage has been initiated and the time elapsed since the monitoring potential parameter. The time that has elapsed since the time when the monitoring parameter exceeded the potential damage level is determined based on the time of the measurement. The pattern is e.g. a curve that begins to increase at the time tp when the monitoring parameter exceeded the potential damage level. Figure 4 shows three examples of suitable patterns A, B, C for the condition monitoring parameter after an injury has been initiated.
    Pattern A is a fast exponentially increasing curve. Pattern B is a medium exponentially increasing curve and pattern C is a slow exponentially increasing curve. The shape of the pattern and thus of the curve depends on e.g. on the type of parameter, the type of damage e.g. if it is a crack or if it is wear, the type of damage process, machine load, speed, and the type and quality of the machine or component. An appropriate pattern can be defined based on previous measurements of the condition monitoring parameter from the time the injury was initiated until the component was replaced. Previous measurements may belong to the machine / component during monitoring, or previous data from other identical or similar machines / components.
    As shown in the example in Figure 2, the measured values increase exponentially after the damage has been initiated and thus the defined pattern should also preferably be exponentially increasing.
    If the pattern has been defined as curve C and a measurement is performed at time t1, an expected value X (t1) of the measurement is determined based on curve C and the time somt that has elapsed since time tp when the monitoring parameter exceeded the potential damage level, as shown in Figure 4.
    The data purification module is arranged to discard new measured values if they deviate too much from the expected value, i.e. if the new measurement deviates from the expected value by more than one limit value, and to replace the oldest of the measured values that have been saved in the moving window with the new measured value if the measured value does not deviate too much from the expected value.
    Thus, if the new measured value does not deviate from the expected value by more than the limit value, the oldest measured value stored in the moving window is discarded and the new measured value is saved in the moving window. This means that after the time when the monitoring parameter has exceeded the potential damage level, measured values are only allowed to be saved in the moving window if the measurement meets two criteria; the value must be greater than the value of the last saved measurement and the value must not deviate too much from the defined pattern. If both of these criteria are met, the measured value is saved in the moving window. At abnormal values, i.e. if the measured value deviates too much from the expected value, the device generates a message to the user, e.g. to request that the measurement be repeated. The abnormal values for a measurement may be different based on the level of the condition monitoring measured value, and the time that has elapsed since the time when the monitoring parameter exceeded the potential damage level, e.g. it can be four times the average increase of the last three measured values, or it can be 100% of the last measured value. The abnormal values can also be different for different machines and components and therefore abnormality can be determined dynamically.
    The apparatus includes a prediction module 10 arranged to predict the level of the condition monitoring parameter at a future time based on the measured values stored in the moving window. The prediction module 10 is arranged to start the prediction when the condition monitoring parameter has exceeded the potential damage level Xp and the movable window is full of measurements of the condition monitoring parameter. If both criteria mentioned earlier are met, the measured value is saved in the moving window and prediction of the value of the condition monitoring parameter is performed.
    Modules 5, 6, 7, 8 and 10 can be implemented in software, hardware or a combination thereof. The units and modules 5-8, 10 and 11 can be connected to each other, for example via an internal bus 9.
    A detailed example of how the prediction can be performed is described in the article "Mechanistic module for predicting the CM parameter value: a case study" by Basim Al-Najjar and lmad Alysyouf. In this example, a wear factor is determined based on the current machine load, future load and previous changes in the degree of wear and the gradient of the wear.
    The prediction module 10 predicts the value of the condition monitoring parameter in the near future, e.g. at the next planned stop or measurement time, based on the measurements stored in the moving window, the estimated time elapsed from the time when the damage was initiated and the estimated wear factor using a mathematical model.
    The model is dynamic because all model parameters and constants are recalculated for each measurement (and prediction), and means that the prediction of the condition monitoring level follows the changes in the behavior of the component wear. The model also responds to changes in production speed and load. Thanks to the prediction being based on measured values in the moving window, it is ensured that the prediction uses recently taken measured values and always excludes older measured values when new measured values have been taken.
    The apparatus further includes a man-machine interface (HM1) 11 arranged to provide messages of maintenance measures and warnings to a user. The man-machine interface 11 is e.g. a monitor. The human-machine interface 11 may also be configured to generate a message regarding whether a measurement value deviates too much from an expected value of the measurement based on a defined pattern of the condition monitoring parameter and to generate a message regarding information to a user that no prediction of the condition monitoring parameter value has been performed. pga. that the movable window is not full, ie. the number of measurements is less than the predetermined number, e.g. three or four measurements. The monitoring module 7 is designed to compare the predicted level of the condition monitoring parameter with at least one limit value A1, A2 for the condition monitoring parameter and to generate a message regarding a maintenance measure when the predicted level of the condition monitoring parameter exceeds the limit value. It is also possible to have several levels with different degrees of sharpness. The monitoring module 7 assesses new predicted values regarding whether maintenance is required or not by comparing the new predicted value with already set levels for safety.
    Figure 5 is a flow chart showing a method and a computer program product according to an embodiment of the present invention. It is understood that each block in the flowchart can be implemented with computer program instructions.
    A potential damage level Xp for the condition monitoring parameter has been defined in advance. One or more limit values that lie between the replacement level RL and the potential damage level Xp may be in advance. the monitoring parameter at which a potential damage has been initiated. From the time also defined in The potential damage level Xp is a level of the condition monitoring parameter exceeds the potential damage level is the damage of the component or machine under development. A pattern for the condition monitoring parameter, which reflects an expected behavior of the monitoring parameter after the damage has been initiated, has been defined in advance or is defined dynamically.
    A new measured value Xi (t) for the condition monitoring parameter is obtained, block 20. The measured value can be retrieved from or received from the company database, the machine or a special file. The measurement values also contain information about the times when the measurement was taken. The new measured value Xi (t) is compared with the last saved measured value in the moving window, block 22. If the new measured value is less than the last saved measured value, the measured value is discarded, blocks 24, 26. If the new measured value is greater than the last saved the measured value in the moving window compares the new measured value Xi (t) with the potential damage level Xp of the condition monitoring parameter, block 28. If the measured value Xilt) is below the potential damage level, the measured value in the moving window is saved together with the time the measurement was performed, block 30. If the movable window is full, the oldest of the saved measured values in the movable window is replaced with the new measured value. Blocks 20-30 are repeated until the new measured value exceeds the potential damage level of the condition monitoring parameter.
    When the measured value is greater than the potential damage level, the time when the condition monitoring parameter exceeded the potential damage level Xp is determined, if this has not already been done for a previous measured value, block 32. The potential damage level Xp may be different for different machines and components. Therefore, it can always be saved for each machine or components in the machine. The time that has elapsed since the time when the monitoring parameter exceeded the potential damage level is determined for the new measurement value based on the difference in time between the time when the measurement was performed and the time when the monitoring parameter exceeded the potential damage level. An expected value for the measurement is determined in accordance with the defined pattern of the condition monitoring parameter and the time that has elapsed since the time when the monitoring parameter exceeded the potential damage level, block 34. The new measured value is compared with the determined expected value and it is determined whether the measured from the expected value, ie whether the measured value deviates from the expected value by more than one limit value, block 34. 11 If the new measured value Xi (t) deviates too much from the pattern, the measurement is discarded and the user is asked to check the measurement and / or repeat the measurement, blocks 36, 38 If the new measurement value Xi (t) deviates too much from the pattern, the measurement may have been performed incorrectly. If the measurement was performed incorrectly, a new correctly measured value Xi + 1 (t) will probably not deviate too much from the expected value. However, if the value was correctly measured and the new value Xi + 1 (t) still deviates too much from the expected value, it may be necessary to immediately perform a check of the measuring system or to perform a maintenance operation, and a warning is generated and displayed to the user.
    If the measured value Xi (t) does not deviate too much from the pattern, the measured value is saved in the moving window together with the time when the measurement was taken, by replacing the oldest measured value saved in the moving window with the new measured value, blocks 36, 40.
    When the condition monitoring parameter exceeds the potential damage level Xp, the prediction of the condition monitoring parameter, block 42, begins. The prediction is repeated for each new measured value that does not deviate too much from the defined pattern. The prediction is performed based on measured values saved in the moving window. The prediction determines the value of the condition monitoring parameter at a future time, e.g. at the next measurement.
    The predicted value is compared with at least one limit value A1 for the condition monitoring parameter, block 44. The limit value is e.g. a replacement level for components. If the predicted value is close to or exceeds the limit value, maintenance instructions are generated that are displayed to the user. If e.g. the predicted value approaches or exceeds the replacement level of the component, an instruction is generated that prompts the user to replace the component immediately. Even when the predicted value exceeds any of the limit values that lie between the replacement level and the potential damage level of the component, an instruction is generated that prompts the user to either wait, begin planning a maintenance operation or to urgently replace the component. The method is repeated until the maintenance has been performed, e.g. that the component has been replaced, block 46. When the maintenance has been performed, measurement data is removed from the moving window and the program automatically restarts for the new replaced machine or component, block 48.
    For example, all levels involved in the device, such as the potential damage level Xp and warning levels, such as limit values, are interchangeable with respect to the user's experience, the machine or component's significance, and the consequences of an injury. The number of measurements needed to fill the movable window can be reduced and increased depending on the need.
    If the new measurement Xi (t) deviates too much from the pattern during the period when the value of the monitoring parameter is still below the potential damage level Xp, the measurement is discarded and the user is asked to check the measurement and / or to repeat the measurement. If the 12 new measurement value deviates too much from the pattern, it may be because the measurement was performed incorrectly. If the measurement was performed incorrectly, a new correct measurement value will probably not deviate too much from the expected value. However, if the new value was correctly measured and the new measured value still deviates too much from the expected value, it may be necessary to perform an immediate check of the measuring system or perform a maintenance operation, and a warning is generated and displayed to the user. If the value of the measurement deviates too much, but is still below the exchange level, the value of the condition monitoring parameter is predicted in the near future and a new warning message is generated.
    By throwing a measured value is meant that the measured value is thrown with respect to the prediction. The discarded measured value can be saved and used for other purposes.
    The steps defined in the requirements can be performed in different order. For example, the prediction can be performed as soon as the new measured value has exceeded the potential damage level of the condition monitoring parameter, i.e. step h can be performed between steps b and e. Measurement values can also be collected from additional condition monitoring parameters e.g. acoustic emission, thermographic, and debris particles of oil.
    权利要求:
    Claims (10)
    [1]
    1. 0 15 20 25 30 35 <13
    [2]
    Method for predicting the condition of a machine or component of the machine based on measurements of a condition monitoring parameter, wherein a potential damage level (Xp) of the condition monitoring parameter has been defined, at which level it is assumed that a damage to the machine or component has been initiated, characterized that a pattern for the condition monitoring parameter has been defined, which pattern reflects an expected behavior of the monitoring parameter after the damage has been initiated, and a plurality of measured values for the condition monitoring parameter have been stored in a moving window, the method comprising: a) a new measurement value for the new measured value for the condition monitoring parameter is compared with the last saved measured value, and the new measured value is discarded if it is less than the last saved measured value, and otherwise the oldest measured value saved in the moving window is replaced with the new measured value; the new a the measured value damage level of the condition monitoring parameter, and steps a - c are repeated until the new measured value exceeds the potential damage level of the condition monitoring parameter, and is compared with the potential d) time (tp) when the monitoring parameter exceeds the potential measurement value level) f, the new measured value is compared with the last saved measured value, and the new measured value is discarded if it is less than the last saved measured value, g) the new measured value is compared with an expected value (X (t1)) which has been determined in accordance with said defined pattern for the condition monitoring parameter and the time (At) that has elapsed since the time when the monitoring parameter exceeded the potential damage level, and the new measurement value is discarded if it deviates too much from the pattern, otherwise the oldest of the measured values is replaced in the moving the window out towards the new measured value, and h) the level of the condition the monitoring parameter at a future time is predicted based on the measured values saved in the moving window.
    [3]
    The method of claim 1, wherein the movable window contains at least three saved measurement values.
    [4]
    The method of claim 1 or 2, wherein said defined pattern of the condition monitoring parameter is a curve (A, B, C) that begins to rise at the time when the monitoring parameter exceeds the potential damage level. 10 15 20 25 30 35 14
    [5]
    The method according to any one of the preceding claims, wherein the new measured value deviates too much from the expected value if the measurement deviates from the expected value by more than one limit value.
    [6]
    The method according to any of the preceding claims, wherein if the measurement deviates too much from the expected value, step g further comprises: - a message to the user is generated to check the measurement and provide a new measurement, - a new measurement value is received, - it the new measured value is compared with an expected value (X (t1)) determined in accordance with the said defined pattern the condition monitoring parameter and the time that has elapsed (At) since the time when the monitoring parameter exceeded the potential damage level n, and - a warning is generated if the new measurement value also deviates too much from the expected value, and otherwise the oldest measured value stored in the moving window is replaced by the new measured value.
    [7]
    A method according to any one of the preceding claims, wherein the method further comprises: i) the predicted level of the condition monitoring parameter is compared with at least one limit level of the condition monitoring parameter, steps e -i are repeated at least until the predicted level of the condition monitoring parameter exceeds a limit level, and j) regarding a maintenance measure when the predicted level of the condition monitoring parameter exceeds said limit level.
    [8]
    A method according to any one of the preceding claims, wherein the machine is a rotating machine and the condition monitoring measurement values are any of vibration measurements, temperature measurements, sound measurements and shock pulse measurements.
    [9]
    An apparatus for predicting the state of a machine or a component of a machine based on measurements of a state monitoring parameter, comprising: - a communication unit (5) arranged to receive measurement values for the state monitoring parameter, - a data store (6) for save a movable window containing measured values for the condition monitoring parameter, - a monitoring module (7) arranged to monitor the measured values and to determine when a measured value exceeds a potential damage level (Xp) at which it is assumed that a damage of 10 15 20 25
    [10]
    The machine or component has been initiated, and determining the time (tp) when the monitoring parameter exceeds the potential damage level n, - a data cleaning module (8) arranged to throw a new measured value if it is less than the last saved measured value in the moving window, determining an expected value for the measurement based on a defined pattern of the condition monitoring parameter, which pattern reflects an expected behavior of the monitoring parameter after the damage has been initiated, and the time that has elapsed since the monitoring parameter exceeded the potential damage level. too large from the expected value, to save discarded measured values in the moving window, and to replace the oldest of the measured values stored in the moving window with the new measured value if the moving window is full, and - a prediction module (10) arranged to predict the level of the condition monitoring parameter at a future time based on the measured values saved in the moving window. The apparatus of claim 8, wherein the apparatus further includes a man-machine interface (11) arranged to provide messages with maintenance actions to a user, and the monitoring module is designed to compare the predicted level of the condition monitoring parameter with at least one limit value for the condition monitoring parameter, maintenance measure when the predicted level of the condition monitoring parameter exceeds the limit value. A computer-readable medium having a program recorded, wherein the program causes the computer to perform the steps of the method according to any one of claims 1 - 7 when the program is run on the computer.
    类似技术:
    公开号 | 公开日 | 专利标题
    SE1300126A1|2014-08-20|A method and apparatus for predicting the condition of a machine or component of the machine
    JP6404893B2|2018-10-17|Tool life estimation device
    JP6008070B1|2016-10-19|Operation management apparatus, operation management method, and recording medium on which operation management program is recorded
    JP2017188030A|2017-10-12|Machine learning device for learning failure prediction of main axis or motor driving main axis and machine learning method, and failure prediction device having machine learning device and failure prediction system
    KR20190030763A|2019-03-22|Accurate predictive maintenance method of driving part
    JPWO2018230645A1|2020-05-21|Abnormality detection device, abnormality detection method, and program
    TWI566065B|2017-01-11|Communication anomaly detection device, communication anomaly detection method and program
    KR20190029749A|2019-03-20|Accurate predictive maintenance method of driving part
    KR101893744B1|2018-09-04|Preventive maintenance method of driving device
    CN104238491B|2017-12-12|Information platform for the data processing based on industrial automation stream
    CN107391335B|2021-09-03|Method and equipment for checking health state of cluster
    JP2017102765A|2017-06-08|Abnormality sign detection system and abnormality sign detection method
    JP6301902B2|2018-03-28|Maintenance time prediction system and maintenance time prediction device
    JP2007213194A|2007-08-23|Situation analytical system and situation analytical method
    JP2018040756A|2018-03-15|Oil mist concentration management device, oil mist management system, and method for managing oil mist
    CN104184123A|2014-12-03|Motor, motor motion protection method and device
    KR101857393B1|2018-06-20|Preventive maintenance method of driving device
    KR102108975B1|2020-05-13|Apparatus and method for condition based maintenance support of naval ship equipment
    JP2018169769A5|2020-02-27|
    KR101893745B1|2018-10-04|Preventive maintenance method of driving device
    KR20190108273A|2019-09-24|Predictive maintenance method of driving device
    JP6369895B2|2018-08-08|Motor abnormality detection system, motor abnormality detection method, and motor abnormality detection program
    KR20180106701A|2018-10-01|Device management system and method based on Internet Of Things
    JP2016031568A|2016-03-07|Abnormality diagnosis apparatus, abnormality diagnosis method, and abnormality diagnosis program
    JP2016004298A|2016-01-12|Pump abnormality detecting system, pump abnormality detecting method, and pump abnormality detecting program
    同族专利:
    公开号 | 公开日
    US9476803B2|2016-10-25|
    SE536922C2|2014-10-28|
    EP2959347A1|2015-12-30|
    US20150377745A1|2015-12-31|
    WO2014127937A1|2014-08-28|
    EP2959347B1|2017-04-26|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US5561610A|1994-06-30|1996-10-01|Caterpillar Inc.|Method and apparatus for indicating a fault condition|
    WO1996013764A1|1994-10-26|1996-05-09|Siemens Aktiengesellschaft|Process for analysing a measurement and measurement analyser for implementing it|
    GB9608953D0|1996-04-29|1996-07-03|Pulp Paper Res Inst|Automatic control loop monitoring and diagnostics|
    US5852793A|1997-02-18|1998-12-22|Dme Corporation|Method and apparatus for predictive diagnosis of moving machine parts|
    US6363332B1|1998-12-22|2002-03-26|Caterpillar Inc.|Method and apparatus for predicting a fault condition using non-linear curve fitting techniques|
    US6370957B1|1999-12-31|2002-04-16|Square D Company|Vibration analysis for predictive maintenance of rotating machines|
    US7233886B2|2001-01-19|2007-06-19|Smartsignal Corporation|Adaptive modeling of changed states in predictive condition monitoring|
    US7539597B2|2001-04-10|2009-05-26|Smartsignal Corporation|Diagnostic systems and methods for predictive condition monitoring|
    AU2002353097A1|2001-12-07|2003-07-09|Battelle Memorial Institute|Methods and systems for analyzing the degradation and failure of mechanical systems|
    EP1403750A1|2002-09-27|2004-03-31|Peter Renner|System for the detection and the evaluation of changements in technical processes|
    US7657399B2|2006-07-25|2010-02-02|Fisher-Rosemount Systems, Inc.|Methods and systems for detecting deviation of a process variable from expected values|
    JP5852793B2|2010-05-21|2016-02-03|株式会社半導体エネルギー研究所|Method for manufacturing liquid crystal display device|
    US20140074345A1|2012-09-13|2014-03-13|Chanan Gabay|Systems, Apparatuses, Methods, Circuits and Associated Computer Executable Code for Monitoring and Assessing Vehicle Health|JP6564363B2|2013-04-22|2019-08-21|ボルボトラックコーポレーション|Method for monitoring the health status of a vehicle system|
    JP6190343B2|2014-09-12|2017-08-30|株式会社神戸製鋼所|Rotating machine abnormality diagnosis device, rotating machine abnormality diagnosis method, and rotating machine|
    US10037026B2|2014-09-25|2018-07-31|General Electric Company|Systems and methods for fault analysis|
    JP6365233B2|2014-10-24|2018-08-01|富士ゼロックス株式会社|Failure prediction device, failure prediction system, and program|
    EP3151072B1|2015-09-29|2020-07-29|Siemens Aktiengesellschaft|Method and system for error detection and monitoring in an electronically controlled or regulated machine part|
    WO2018119845A1|2016-12-29|2018-07-05|深圳配天智能技术研究院有限公司|State detection method and system for numerical control machine tool|
    CN110082103B|2018-04-26|2020-12-15|广东电网有限责任公司|Instability fault early warning method for steam turbine nozzle steam distribution shaft system|
    FR3094421A1|2019-03-29|2020-10-02|Wilo Intec|PREDICTIVE MAINTENANCE PROCEDURE FOR A FLUID CIRCULATION PUMP|
    CN110605615A|2019-07-30|2019-12-24|青岛市中心医院|Computer-controlled intelligent production line with automatic repair function|
    CN112943595A|2021-02-07|2021-06-11|三一重工股份有限公司|Hydraulic pump fault prediction method, hydraulic pump fault prediction device, electronic equipment and storage medium|
    法律状态:
    优先权:
    申请号 | 申请日 | 专利标题
    SE1300126A|SE536922C2|2013-02-19|2013-02-19|A method and apparatus for predicting the condition of a machine or component of the machine|SE1300126A| SE536922C2|2013-02-19|2013-02-19|A method and apparatus for predicting the condition of a machine or component of the machine|
    US14/768,605| US9476803B2|2013-02-19|2014-01-20|Method and an apparatus for predicting the condition of a machine or a component of the machine|
    PCT/EP2014/051006| WO2014127937A1|2013-02-19|2014-01-20|A method and an apparatus for predicting the condition of a machine or a component of the machine|
    EP14701029.2A| EP2959347B1|2013-02-19|2014-01-20|A method and an apparatus for predicting the condition of a machine or a component of the machine|
    [返回顶部]