Screw machine fault detection method and device and water chilling unit

Abstract

The invention discloses a screw machine fault detection method and device and a water chilling unit, and belongs to the field of water chilling units. The screw machine fault detection method comprises the steps of performing trainability detection and diagnosability detection on actual working parameters of a to-be-detected screw machine based on candidate screw machine types in an online model, and obtaining trainable parameters and diagnosable parameters; based on at least one of the trainable parameters and the diagnosable parameters, a target mode is controlled to be entered, and the target mode comprises at least one of a precise diagnosis mode, a model training mode and a basic diagnosis mode. According to the screw machine fault detection method, the obtained real-time working parameters are subjected to trainability detection and diagnosability detection, the optimal detection mode is selected based on the obtained actual data, the method is suitable for fault detection of multiple types of screw machines, high universality is achieved, multiple fault types can be detected, and the fault detection efficiency is improved. The method has high detection precision and detection accuracy.

Description

Technical Field

[0001] The present application belongs to the field of water chillers, and in particular, relates to a screw machine fault detection method, device and water chiller. Background Art

[0002] In chillers, refrigeration screw compressors are mainly without slide valve position indicators. In the related art, the compressor current model is generally used to predict the compressor current, thereby determining the position of the slide valve. However, due to the complex and changeable actual working conditions, the prediction accuracy of the existing compressor current model is poor, and the position of the compressor slide valve cannot be correctly specified, resulting in misdiagnosis or missed diagnosis of faults, and the accuracy of the detection results is poor. In addition, the existing compressor fault detection method has poor versatility and cannot be applied to a variety of different types of screw machines. Summary of the invention

[0003] The present application aims to solve at least one of the technical problems existing in the prior art. To this end, the present application proposes a screw machine fault detection method, device and chiller, which are suitable for fault detection of various types of screw machines, have high versatility, can detect various fault types, and have high detection precision and accuracy.

[0004] In a first aspect, the present application provides a screw machine fault detection method, the method comprising:

[0005] Perform trainability detection and diagnosability detection on the actual working parameters of the screw machine to be tested based on the candidate screw machine types in the online model to obtain trainable parameters and diagnosable parameters;

[0006] Based on at least one of the trainable parameters and the diagnosable parameters, control is performed to enter a target mode, wherein the target mode includes at least one of a precise diagnosis mode, a model training mode, and a basic diagnosis mode.

[0007] According to the screw machine fault detection method of the present application, by performing trainability tests and diagnosability tests on the acquired real-time working parameters, the best detection mode is selected based on the actual data acquired. It is suitable for fault detection of various types of screw machines, has high versatility, and can detect various fault types, with high detection precision and accuracy.

[0008] According to an embodiment of the present application, the trainability detection and diagnosability detection of the actual working parameters of the screw machine to be tested based on the candidate screw machine type in the online model to obtain the trainable parameters and the diagnosable parameters include:

[0009] Perform multi-dimensional coverage calculation based on the candidate screw machine types to obtain the diagnosable parameters;

[0010] The trainable parameter is obtained based on at least two of the candidate screw machine type, the density of the actual operating parameter, and the duration of the actual operating parameter.

[0011] According to one embodiment of the present application, the step of controlling entry into a target mode based on at least one of the trainable parameter and the diagnosable parameter includes:

[0012] When the diagnosable parameter is greater than a first threshold, controlling entry into the precise diagnosis mode;

[0013] When the diagnosable parameter is less than a second threshold, controlling entry into the precise diagnosis mode and the model training mode;

[0014] When the trainable parameter is greater than a third threshold and the diagnosable parameter is less than a fourth threshold, controlling entry into the model training mode;

[0015] In a case where the diagnosable parameter is not greater than the first threshold or the trainable parameter is not greater than the third threshold, control enters the basic diagnostic mode.

[0016] According to one embodiment of the present application, at least one of the third threshold and the fourth threshold is a Pareto optimal value determined based on the initial normal operating parameters corresponding to the type of screw machine to be tested and based on the fault diagnosis time and fault diagnosis accuracy.

[0017] According to one embodiment of the present application, controlling entry into the precise diagnosis mode includes:

[0018] Based on the actual working parameters, obtaining oil pressure fault information and control action fault information;

[0019] When it is determined that the oil pressure and the control action are normal, the actual working parameters are input into the online model to obtain the predicted current value output by the online model;

[0020] determining that the solenoid valve is faulty when a second deviation between the actual current value and the predicted current value exceeds a maximum current error tolerance;

[0021] When the second deviation between the actual current value and the predicted current value does not exceed the maximum current error tolerance, it is determined that the solenoid valve is normal.

[0022] According to one embodiment of the present application, the maximum current error tolerance is determined based on the type of the screw machine to be tested, the model training accuracy and the sampling frequency corresponding to the actual working parameters.

[0023] According to one embodiment of the present application, controlling entry into the basic diagnostic mode includes:

[0024] Based on the actual working parameters, oil pressure fault information and control action fault information are obtained.

[0025] According to an embodiment of the present application, the obtaining of oil pressure fault information and control action fault information based on the actual working parameters includes:

[0026] The actual working parameters are subjected to at least one of state range analysis, outlier detection, dynamic characteristic analysis and neighborhood analysis to determine oil pressure abnormality.

[0027] According to an embodiment of the present application, the obtaining of oil pressure fault information and control action fault information based on the actual working parameters includes:

[0028] Obtaining a controlled parameter controlled by a controller of the screw machine from the actual working parameter;

[0029] Based on the controlled parameter, it is determined that the control action is abnormal.

[0030] According to an embodiment of the present application, determining that the control action is abnormal based on the controlled parameter includes:

[0031] Obtaining a first deviation between the controlled parameter and a first target parameter;

[0032] When the first deviation continues to maintain a consistent deviation within a first time period, it is determined that the control action is abnormal.

[0033] In a second aspect, the present application provides a screw machine fault detection device, the device comprising:

[0034] A first processing module is used to perform trainability detection and diagnosability detection on actual working parameters of the screw machine to be tested based on the candidate screw machine types in the online model, and obtain trainable parameters and diagnosable parameters;

[0035] The second processing module is used to control entering into a target mode based on at least one of the trainable parameters and the diagnosable parameters, wherein the target mode includes at least one of a precise diagnosis mode, a model training mode and a basic diagnosis mode.

[0036] According to the screw machine fault detection device of the present application, by performing trainability tests and diagnosability tests on the acquired real-time working parameters, the best detection mode is selected based on the actual data acquired. It is suitable for fault detection of various types of screw machines, has high versatility, and can detect various fault types, with high detection precision and accuracy.

[0037] In a third aspect, the present application provides a chiller, comprising:

[0038] Screw machine;

[0039] As described in the screw machine fault detection device of the second aspect, the screw machine fault detection device is electrically connected to the screw machine.

[0040] In a fourth aspect, the present application provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the screw machine fault detection method as described in the first aspect above is implemented.

[0041] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the screw machine fault detection method as described in the first aspect above.

[0042] The above one or more technical solutions in the embodiments of the present application have at least one of the following technical effects:

[0043] By performing trainability tests and diagnosability tests on the acquired real-time working parameters, the best detection mode is selected based on the actual data obtained. It is suitable for fault detection of various types of screw machines, has high versatility, can detect various fault types, and has high detection precision and accuracy.

[0044] Furthermore, by determining the maximum current error tolerance based on the type of screw machine to be tested, the model training accuracy and the sampling frequency corresponding to the actual working parameters, the value of the maximum current error tolerance can be automatically adjusted to the optimal range based on the change of the type of screw machine to be tested, effectively reducing the risk of missed detection and false detection, and improving calculation efficiency, thereby effectively improving the accuracy of the prediction results and prediction efficiency.

[0045] Furthermore, adaptively adjusting the third and fourth thresholds based on the type of screw machine can significantly improve the current prediction accuracy of the online model, with strong generalization and scalability, and is suitable for all screw compressor energy adjustment failure diagnosis.

[0046] Furthermore, by combining the operating status of the controller and the equipment, steady-state and dynamic characteristic analysis of the current, intake and exhaust pressure, and solenoid valve action can be performed, which can accurately identify various faults such as solenoid valve failure, control action failure, and insufficient oil pressure, thereby improving the fault diagnosis rate.

[0047] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0049] Figure 1 It is one of the flow diagrams of the screw machine fault detection method provided in the embodiment of the present application;

[0050] Figure 2 This is the second flow chart of the screw machine fault detection method provided in the embodiment of the present application;

[0051] Figure 3 This is the third flow chart of the screw machine fault detection method provided in the embodiment of the present application;

[0052] Figure 4 This is one of the schematic diagrams of the effect of the screw machine fault detection method provided in the embodiment of the present application;

[0053] Figure 5 This is the second schematic diagram of the effect of the screw machine fault detection method provided in the embodiment of the present application;

[0054] Figure 6 This is the third schematic diagram of the effect of the screw machine fault detection method provided in the embodiment of the present application;

[0055] Figure 7 is a structural schematic diagram of a screw machine fault detection device provided in an embodiment of the present application;

[0056] Figure 8 is a schematic diagram of the structure of a chiller provided in an embodiment of the present application;

[0057] Fig. 9 It is a schematic diagram of the principle of the online model provided in the embodiment of the present application. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the present application belong to the scope of protection of this application.

[0059] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are generally of one type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0060] The screw machine fault detection method, screw machine fault detection device, electronic device and readable storage medium provided in the embodiments of the present application are described in detail below in combination with the accompanying drawings through specific embodiments and their application scenarios.

[0061] Among them, the screw machine fault detection method can be applied to a terminal, and can be specifically executed by hardware or software in the terminal.

[0062] The screw machine fault detection method provided in the embodiment of the present application may be executed by a chiller or a functional module or functional entity in the chiller that can implement the screw machine fault detection method. The screw machine fault detection method provided in the embodiment of the present application is described below using the chiller as an example of the execution body.

[0063] like Figure 1 As shown, the screw machine fault detection method includes: step 110 and step 120.

[0064] Step 110: Perform trainability detection and diagnosability detection on actual working parameters of the screw machine to be tested based on the candidate screw machine types in the online model to obtain trainable parameters and diagnosable parameters;

[0065] In this step, the actual working parameters are the actual working condition data of the screw machine.

[0066] The actual working parameters may include, but are not limited to: actual current, suction pressure of the screw machine, exhaust pressure of the screw machine, start / stop status of the solenoid valve, and controlled parameters related to the controller.

[0067] For example, the controlled parameter may be the evaporator outlet water temperature or the like.

[0068] The online model is a pre-calibrated model.

[0069] Online models are used to diagnose faults on different types of screw machines.

[0070] Trainability testing is used to quantify the extent to which the actual working parameters currently obtained can effectively train the online model.

[0071] The trainable parameters are used to characterize whether the actual working parameters obtained can be used to train the online model.

[0072] Diagnosability testing is used to quantify the extent to which faults can be accurately and effectively identified by current online models.

[0073] The diagnosable parameters are used to characterize whether the online model can perform effective fault diagnosis on the screw machine under test.

[0074] In some embodiments, the diagnosable parameter may be expressed as coverage of the data.

[0075] Trainable parameters can be represented as grid density.

[0076] The following uses coverage and grid density as examples to illustrate how to determine diagnosable parameters and trainable parameters.

[0077] In some embodiments, step 110 may include:

[0078] Perform multi-dimensional coverage calculation based on candidate screw machine types to obtain diagnosable parameters;

[0079] Trainable parameters are obtained based on at least two of the candidate screw machine types, the density of the actual operating parameters, and the duration of the actual operating parameters.

[0080] In this embodiment, the diagnosable parameter may be expressed as the coverage of the data.

[0081] Trainable parameters can be represented as grid density.

[0082] The density of the actual working parameters is the density of the input samples, and the duration of the actual working parameters is the duration of the input samples.

[0083] For example, during the actual execution process, a multi-dimensional coverage calculation can be performed based on the candidate screw machine types in the current online model and the applicable scope of the model operation to obtain the coverage rate.

[0084] A comprehensive judgment is made based on the candidate screw machine types in the current online model, the density of the input samples, and the duration of the input samples to obtain the corresponding grid density.

[0085] It is understandable that different types of screw machines have different operating ranges, and the threshold values ​​of grid density requirements for input variables of different dimensions may also be different.

[0086] like Figure 2 As shown, in some embodiments, step 110 may also include:

[0087] Preprocessing the actual working parameter to obtain the first working parameter;

[0088] Based on the candidate screw machine types in the online model, a trainability test and a diagnosability test are performed on the first working parameter to obtain a trainable parameter and a diagnosable parameter.

[0089] In this embodiment, preprocessing is used to eliminate abnormal data, including but not limited to: screening shutdown data and performing missing value interpolation.

[0090] By preprocessing the actual working parameters, abnormal data in the initial data can be effectively eliminated, and the accuracy of the data set used for subsequent analysis and calculation can be improved, thereby improving the accuracy and precision of the prediction results.

[0091] Step 120: Control entering a target mode based on at least one of a trainable parameter and a diagnosable parameter.

[0092] In this step, the target mode may include at least one of a precise diagnosis mode, a model training mode, and a basic diagnosis mode.

[0093] Among them, the precise diagnosis mode is used to diagnose multiple types of faults such as oil pressure fault type, control action fault type and solenoid valve fault type.

[0094] The basic diagnosis mode is used to diagnose multiple types of faults such as oil pressure fault type and control action fault type.

[0095] The model training mode is used to calibrate, train, and update online models.

[0096] After obtaining the trainable parameters and the diagnosable parameters, trainability detection and diagnosability judgment can be performed based on the trainable parameters and the diagnosable parameters, and entry into a corresponding mode can be controlled based on the judgment result.

[0097] It should be noted that in the present application, the precise diagnosis mode needs to involve an online model, and the precise diagnosis mode can only be entered if the diagnosability test is passed.

[0098] The basic diagnostic mode is routine testing and does not involve online models.

[0099] The model training mode can only be entered if the trainability test passes.

[0100] The specific online model training method will be described in the following examples and will not be elaborated here.

[0101] During the research and development process, the inventors discovered that the models used for screw machine fault diagnosis in the relevant technology are often matched with the type of screw machine. For different types of chillers or screw machines with different performances, the models need to be retrained, learned and verified, resulting in cumbersome steps and poor versatility. In addition, for the same type of screw machine, when the actual working parameters obtained are very different from the original training sample set, the accuracy of the final prediction results will be greatly reduced.

[0102] In the present application, by establishing an online training, diagnosis and qualitative model driven by both physics and data, the optimal fault detection mode is selected based on the results of trainability detection and diagnosability detection of real-time working parameters. The model can be adaptively adjusted for different screw machine types or different sample sets, has strong generalization and scalability, and significantly improves the accuracy of fault prediction results and the efficiency of fault diagnosis.

[0103] According to the screw machine fault detection method provided in the embodiment of the present application, by performing trainability tests and diagnosability tests on the acquired real-time working parameters, the best detection mode is selected based on the actual data acquired. The method is suitable for fault detection of various types of screw machines, has high versatility, can detect various fault types, and has high detection precision and accuracy. In addition, the online model can be updated in real time based on the actual working parameters acquired, and has strong generalization ability.

[0104] In some embodiments, step 120 may include:

[0105] When the diagnosable parameter is greater than a first threshold, the control enters a precise diagnosis mode;

[0106] When the diagnosable parameter is less than the second threshold, the control enters the precise diagnosis mode and the model training mode;

[0107] When the trainable parameter is greater than the third threshold and the diagnosable parameter is less than the fourth threshold, controlling entry into the model training mode;

[0108] In the event that the diagnosable parameter is not greater than the first threshold or the trainable parameter is not greater than the third threshold, control enters the basic diagnostic mode.

[0109] In this embodiment, the first threshold is a threshold that satisfies diagnosability detection. For example, the first threshold may be set to 0.

[0110] Continue to refer Figure 2When the diagnosable parameter is not greater than the first threshold, it indicates that there is no data available for fault diagnosis in the actual working parameters obtained, that is, the current online model cannot perform effective fault diagnosis on the screw machine under test, and the precise diagnosis mode cannot be entered. In this case, the basic diagnosis mode can be entered.

[0111] When the diagnosable parameter is greater than the first threshold, it indicates that at least part of the acquired actual working parameters can be used for fault diagnosis, that is, the existing online model can effectively diagnose the faults of the screw machine to be tested, and the precise diagnosis mode can be entered.

[0112] The second threshold may be set to 100%.

[0113] When the diagnosable parameter is less than the second threshold, it indicates that the actual working parameters obtained all cover the applicable scope of the online model, and there is new data entering that exceeds the applicable scope of the original online model; these new data can be used to update the online model; that is, the control simultaneously enters the precise diagnosis mode and the model training mode, so as to update the online model while performing precise diagnosis, so as to improve the applicable scope and calculation accuracy of the online model.

[0114] The third threshold is a grid density threshold that satisfies the trainability test.

[0115] It is understandable that different types of screw machines may have different operating ranges, and thus the third thresholds corresponding to different types of screw machines may also be different.

[0116] The fourth threshold is a specific update threshold.

[0117] The fourth threshold is greater than the first threshold and less than the second threshold. The fourth threshold can be user-defined, such as set to 80% or 75%, etc., which is not limited in this application.

[0118] For example, when the obtained trainable parameter is greater than the third threshold, it indicates that the input sample for online model updating is trainable; in this case, the online model needs to be updated in combination with the coverage in the diagnosability detection. If the coverage is less than the fourth threshold, the control enters the model training mode, and the model is updated to enter the model training mode.

[0119] When the trainable parameter is not greater than the third threshold, or when the trainable parameter is greater than the third threshold but the coverage is not less than the fourth threshold, the model training mode is not entered.

[0120] When the trainable parameters and controllable parameters are 0, it indicates that there is no available online model (it may be that the screw machine type does not match, or it may be that the applicable scope does not match, or there may be no online model, etc.), and the quality of the actual working parameters input is not high, then the control enters the basic diagnosis mode, that is, in addition to fault diagnosis through the online model, data analysis, control deviation detection, pressure difference detection and other fault analysis are performed.

[0121] According to the screw machine fault detection method provided in the embodiment of the present application, by establishing an online training, diagnosis and qualitative model driven by both physics and data, the optimal fault detection mode is selected based on the results of trainability detection and diagnosability detection of real-time working parameters. The model can be adaptively adjusted for different screw machine types or different sample sets, and has strong generalization and scalability.

[0122] In some embodiments, the third threshold and the fourth threshold may be adaptive thresholds.

[0123] In some embodiments, the third threshold and the fourth threshold may be Pareto optimal values ​​determined based on the initial normal operating parameters corresponding to the type of screw machine to be tested and based on the fault diagnosis time and the fault diagnosis accuracy.

[0124] In this embodiment, the grid density threshold (ie, the third threshold) and the coverage update threshold (ie, the fourth threshold) are both optimal values ​​determined based on initial partial normal operation data collected from different types of screw machines.

[0125] Among them, the optimality refers to finding the Pareto optimality based on fault diagnosis time and fault diagnosis accuracy.

[0126] like Figure 4 The following example shows the trend of fault diagnosis time and current prediction accuracy of a screw machine under different coverage thresholds for the whole month of July. Figure 4 It can be seen that the higher the fourth threshold is, the more frequently the training model is updated, and the higher the current prediction accuracy is; however, with the change of the fourth threshold, the diagnosis time does not change monotonically, which depends on the quality of the collected data and the operating conditions, and depends on whether the current mode is dominated by the training mode or the diagnosis mode. Figure 4 It can be known that the optimal fourth threshold value corresponding to this type of screw machine may be 60% or 100%.

[0127] The inventor has determined through multiple experiments that the current prediction accuracy of the online training model optimized by the grid density provided by the method provided in this application reaches more than 90%. Figure 5 and Figure 6 As shown, the application can quickly identify through the online model that the operating status from August to September 25 was good and no abnormalities were detected, while adjustable valve failures were detected on September 26 and 27.

[0128] According to the screw machine fault detection method provided in the embodiment of the present application, the third threshold and the fourth threshold are adaptively adjusted based on the type of the screw machine, which can significantly improve the current prediction accuracy of the online model, has strong generalization ability and scalability, and is suitable for all screw compressor energy adjustment failure diagnosis.

[0129] The specific implementation methods of the three modes are described below.

[0130] 1. Accurate diagnosis mode

[0131] like Figure 3 As shown, in some embodiments, controlling entry into the precision diagnosis mode may include:

[0132] Based on the actual working parameters, obtaining at least one of oil pressure fault information and control action fault information;

[0133] When it is determined that the oil pressure and control action are normal, the actual working parameters are input into the online model to obtain the predicted current value output by the online model;

[0134] determining that the solenoid valve is faulty when a second deviation between the actual current value and the predicted current value exceeds a maximum current error tolerance;

[0135] When the second deviation between the actual current value and the predicted current value does not exceed the maximum current error tolerance, it is determined that the solenoid valve is normal.

[0136] In this embodiment, the solenoid valve is an adjustable solenoid valve.

[0137] Oil pressure fault information includes abnormal oil pressure and normal oil pressure.

[0138] Among them, abnormal oil pressure can be manifested as a poor adjustment problem caused by insufficient oil pressure.

[0139] In the present application, the actual working parameters of the screw machine may be subjected to time series data feature analysis to determine whether the oil pressure is abnormal.

[0140] In the actual implementation process, the oil pressure fault information can be obtained through conventional data processing such as table lookup, or by using a pre-trained neural network model, or by any other feasible method, which is not limited in this application.

[0141] Control action fault information is used to characterize control-related fault conditions.

[0142] The control action fault information includes abnormal control action and normal control action. In the case of abnormal control action, the control action fault information may also include a specific control fault type.

[0143] It is understandable that sensors and actuators, as the most frequently used parts, have a higher failure rate. In some embodiments, rapid fault screening can be performed based on control deviation detection of a controller related to the compressor.

[0144] Of course, in other embodiments, a pre-trained neural network model may be collected to determine control action fault information, and in actual application, a selection may be made based on actual conditions, which is not limited in this application.

[0145] The online model is a pre-calibrated model used to predict the current value of the screw machine under normal working conditions.

[0146] Fig. 9 An online model is illustrated, whose input is the actual operating parameters and output is the predicted current value.

[0147] The online model can predict the current value corresponding to one or more types of screw machines, which is determined based on the actual calibration situation.

[0148] The second deviation may be an absolute value of a difference between the predicted current value and the actual current value.

[0149] In this embodiment, the predicted current value obtained by the current prediction model is compared with the actual current value, and when the second deviation between the predicted current value and the actual current value is greater than the maximum current error tolerance, it is determined that the solenoid valve is faulty.

[0150] When the second deviation between the predicted current value and the actual current value is not greater than the maximum current error tolerance, it is approximately considered that the solenoid valve is normal.

[0151] According to the screw machine fault detection method provided in the embodiment of the present application, combined with the controller and the equipment operating status, steady-state and dynamic characteristic analysis of the current, suction and exhaust pressure, and solenoid valve action is performed, which can accurately identify various faults such as solenoid valve failure, control action failure, and insufficient oil pressure, thereby improving the fault diagnosis rate.

[0152] In some embodiments, the maximum current error tolerance may be an adaptive threshold.

[0153] In some embodiments, the maximum current error tolerance can be determined based on the type of screw machine to be tested, the model training accuracy, and the sampling frequency corresponding to the actual working parameters.

[0154] In this embodiment, the maximum current error tolerance can be determined based on the type of screw machine, model training accuracy, and the sampling frequency of the operating data itself. It can be understood that if the maximum current error tolerance is too large, the diagnosis speed will be accelerated, but it is too large to cause missed detection, that is, the FAB in the confusion matrix of Table 1 increases; if the maximum current error tolerance is too small, the diagnosis speed will be slowed down and it is easy to cause false detection, that is, the FBA increases;

[0155] That is, if the maximum current error tolerance is too large or too small, it will affect the accuracy of fault diagnosis.

[0156] Table 1

[0157]

[0158] In this application, the current error tolerance is dynamically adjusted for different screw machine types (adjusted as needed when the sampling frequency changes significantly), which is the maximum training error ε obtained by training the model with the initial normal operation data. max .

[0159] According to the screw machine fault detection method provided in the embodiment of the present application, by determining the maximum current error tolerance based on the type of screw machine to be tested, the model training accuracy and the sampling frequency corresponding to the actual working parameters, the value of the maximum current error tolerance can be automatically adjusted to the optimal range based on the change of the type of screw machine to be tested, thereby effectively reducing the risk of missed detection and false detection and improving calculation efficiency, thereby effectively improving the accuracy of the prediction results and the prediction efficiency.

[0160] 2. Basic Diagnostic Mode

[0161] Continue to refer Figure 3 In some embodiments, controlling entry into the basic diagnostic mode may include: obtaining oil pressure fault information and control action fault information based on actual working parameters.

[0162] In the basic diagnosis mode, the oil pressure fault information can be obtained based on the actual working parameters to determine whether the oil pressure is abnormal. The diagnosis is terminated after the oil pressure is determined to be abnormal. After the oil pressure is determined to be normal, further processing is performed to obtain the control action fault information to determine whether the control action is abnormal.

[0163] Continue to refer Figure 3 In some embodiments, obtaining oil pressure fault information and control action fault information based on actual working parameters may include: performing at least one of state range analysis, outlier detection, dynamic characteristic analysis and neighborhood analysis on the actual working parameters to determine oil pressure abnormality.

[0164] In this embodiment, the state range analysis is to analyze the range of the actual working parameters to determine whether it exceeds the reasonable range. If it exceeds the reasonable range, it can be approximately considered that the oil pressure is abnormal.

[0165] Outlier detection is used to detect the degree of dispersion of actual working parameters in order to detect abnormal data that is far different from normal data.

[0166] Dynamic characteristics analysis is used to output the response characteristics to input quantities that vary with time.

[0167] Neighborhood analysis is to conduct neighborhood analysis with the target phase element as the center.

[0168] In the actual implementation process, conventional means can be used to implement state range analysis, outlier detection, dynamic characteristic analysis and neighborhood analysis, and this application does not limit this.

[0169] According to the screw machine fault detection method provided in the embodiment of the present application, by performing at least one of state range analysis, outlier detection, dynamic characteristic analysis and neighborhood analysis on the actual working parameters, it is possible to effectively diagnose energy regulation failure problems caused by insufficient oil pressure, with high diagnostic accuracy and simple and convenient operation.

[0170] Continue to refer Figure 3 In some embodiments, obtaining oil pressure fault information and control action fault information based on actual working parameters may include:

[0171] When the oil pressure is determined to be normal, the controlled parameters controlled by the controller of the screw machine are obtained from the actual working parameters;

[0172] Based on the controlled parameters, determine that the control action is abnormal.

[0173] In this embodiment, the controlled parameters are parameters obtained from actual working parameters and controlled by the controller of the screw machine.

[0174] For example, the controlled parameter may be the evaporator outlet water temperature or the like.

[0175] The control action is a related action for controlling the controlled parameter, such as a PID control action or other industrial control action, etc., which is not limited in this application.

[0176] The control action abnormality may be an abnormality related to the control action, including but not limited to: a controller abnormality related to the compressor, an adjustable valve control action abnormality, other solenoid valve control action abnormality, and a control parameter setting abnormality.

[0177] In the actual execution process, whether the control action is abnormal can be determined based on the change trend of the controlled parameter or whether the controlled parameter is within the normal range. In some embodiments, determining that the control action is abnormal based on the controlled parameter may include:

[0178] Obtaining a first deviation between a controlled parameter and a first target parameter;

[0179] When the first deviation continues to maintain a consistent deviation within the first time period, it is determined that the control action is abnormal.

[0180] In this embodiment, the first target parameter is a preset parameter and can be customized by the user.

[0181] The first deviation may be expressed as a difference or a ratio.

[0182] The first duration is a longer period of time, and the specific value can be based on user customization, such as setting it to 30 minutes or 45 minutes, etc., which is not limited in this application.

[0183] The following explanation is given by taking the evaporator outlet water temperature as an example where the controlled parameter is used.

[0184] When the controlled parameter is the evaporator outlet water temperature T, the first target parameter can be a preset given amount of water outlet temperature TS. By comparing the changing trend of the first deviation between the evaporator outlet water temperature T and the given amount of water outlet temperature TS, it can be determined whether a control abnormality occurs; for example, when the first deviation between the evaporator outlet water temperature T and the given amount of water outlet temperature TS maintains a consistent positive / negative deviation within a first time period, it can be approximately considered that the controlled parameter is not within the normal range, and a preliminary judgment is made as to the control abnormality area.

[0185] In some embodiments, determining that the control action is abnormal when the first deviation continues to maintain a consistent deviation within a first time period may also include: determining a target control fault when the rate of change of the controlled parameter is greater than a fault detection threshold corresponding to the target control fault.

[0186] In this embodiment, the value of the fault detection threshold may be user-defined.

[0187] Target control failures may include abnormal control action of the adjustable valve or abnormal control action of the solenoid valve.

[0188] Continuing with the example of the controlled parameter being the evaporator outlet water temperature, when the first deviation between the evaporator outlet water temperature T and the given outlet water temperature TS maintains a consistent positive / negative deviation within the first time period, the control abnormality area can be approximately judged; then the change rate of the evaporator outlet water temperature T is further judged. If the change rate of the outlet water temperature T is greater than the fault detection threshold, the control action failure is determined.

[0189] In this application, by combining the functions of monitoring and diagnosing the operating status of the controller and the equipment, the cause of the fault can be further refined. For example, the abnormal index Exception Flag = 1 represents a poor adjustment fault. When checking for insufficient oil pressure or controller-related faults, the index is 2 and 3 respectively.

[0190] Continue to refer Figure 3 In some embodiments, after step 120, the method may further include: outputting at least one of a fault diagnosis report and an alarm message based on the diagnosis result.

[0191] In this embodiment, the alarm information may include: variable out of bounds, shutdown, and NAN, etc.

[0192] By outputting fault diagnosis reports and alarm information, it is convenient to provide guidance for subsequent fault-tolerant control.

[0193] 3. Model Training Mode

[0194] In some embodiments, the online model may be a neural network model or an n-coefficient model.

[0195] This application does not limit the form of the online model, which can be set based on actual needs.

[0196] For example, the online model can be a fixed frequency four-slide valve section 15 coefficient model, which is expressed as follows:

[0197] Current=(C1+C2*Psuc+C3*Pdis+C4*Psuc*Psuc+C5*Psuc*Pdis+C6*Pdis*Pdis+C7*Psu c*Psuc*Psuc+C8*Pdis*Psuc*Psuc+C9*Psuc*Pdis*Pdis+C10*Pdis*Pdis*Pdis)+(C11*SV1+C12*SV2+C13*SV3+C14*SV4+C15*SV1*SV2)

[0198] In some embodiments, the online model can be determined based on model training time and prediction accuracy.

[0199] In this embodiment, for example, for integer 0-1 Boolean inputs such as slide valve switches, the n-coefficient model is often likely to have accuracy deviations, especially when the capacity-adjusting solenoid valve is just actuated and the transient characteristics are relatively strong. In this case, a neural network model is required for prediction and estimation. However, it is difficult for general feedforward artificial neural networks to accurately predict such integer inputs, so a recurrent neural network RNN ​​can be selected.

[0200] It should be noted that the recurrent neural network model here is universal for different inputs and outputs, and can be directly applied to different types of fixed-frequency and variable-frequency compressors; while the n-coefficient model training module is universal, but the model form needs to be input in advance, and it is also applicable to different types of compressors.

[0201] Continue to refer Figure 3 In some embodiments, for each online learning model, in addition to storing the applicable scope, normalization parameters, model form, and optimal threshold required for the model, an appropriate amount of data can also be saved for model updating. For example, stratified sampling can be used to obtain data for model updating.

[0202] The screw machine fault detection method provided in the embodiment of the present application has high flexibility by providing multiple types of online models.

[0203] The screw machine fault detection method provided in the embodiment of the present application can be executed by a screw machine fault detection device. In the embodiment of the present application, the screw machine fault detection device executing the screw machine fault detection method is taken as an example to illustrate the screw machine fault detection device provided in the embodiment of the present application.

[0204] The embodiment of the present application also provides a screw machine fault detection device.

[0205] like Figure 7 As shown, the screw machine fault detection device includes: a first processing module 710 and a second processing module 720.

[0206] A first processing module 710 is used to perform trainability detection and diagnosability detection on actual working parameters of the screw machine to be tested based on the candidate screw machine types in the online model, and obtain trainable parameters and diagnosable parameters;

[0207] The second processing module 720 is used to control entering into a target mode based on at least one of the trainable parameters and the diagnosable parameters, where the target mode includes at least one of a precise diagnosis mode, a model training mode, and a basic diagnosis mode.

[0208] According to the screw machine fault detection device provided in the embodiment of the present application, by performing trainability tests and diagnosability tests on the acquired real-time working parameters, the optimal detection mode is selected based on the actual data acquired. The device is suitable for fault detection of various types of screw machines, has high versatility, can detect various fault types, and has high detection precision and accuracy.

[0209] In some embodiments, the first processing module 710 may also be used to:

[0210] Perform multi-dimensional coverage calculation based on candidate screw machine types to obtain diagnosable parameters;

[0211] Trainable parameters are obtained based on at least two of the candidate screw machine types, the density of the actual operating parameters, and the duration of the actual operating parameters.

[0212] In some embodiments, the second processing module 720 may also be used to:

[0213] When the diagnosable parameter is greater than a first threshold, the control enters a precise diagnosis mode;

[0214] When the diagnosable parameter is less than the second threshold, the control enters the precise diagnosis mode and the model training mode;

[0215] When the trainable parameter is greater than the third threshold and the diagnosable parameter is less than the fourth threshold, controlling entry into the model training mode;

[0216] In the event that the diagnosable parameter is not greater than the first threshold or the trainable parameter is not greater than the third threshold, control enters the basic diagnostic mode.

[0217] In some embodiments, the second processing module 720 may also be used to:

[0218] acquiring at least one of oil pressure fault information and control action fault information based on actual working parameters;

[0219] When it is determined that the oil pressure and control action are normal, the actual working parameters are input into the online model to obtain the predicted current value output by the online model;

[0220] In the case where the absolute value of the difference between the actual current value and the predicted current value exceeds the maximum current error tolerance, determining that the solenoid valve is faulty;

[0221] When the second deviation between the actual current value and the predicted current value does not exceed the maximum current error tolerance, it is determined that the solenoid valve is normal.

[0222] In some embodiments, the second processing module 720 may also be used to:

[0223] Based on the actual working parameters, obtain oil pressure fault information and control action fault information.

[0224] In some embodiments, the second processing module 720 may also be used to:

[0225] At least one of state range analysis, outlier detection, dynamic characteristic analysis and neighborhood analysis is performed on the actual working parameters to determine oil pressure fault information.

[0226] In some embodiments, the second processing module 720 may also be used to:

[0227] When the oil pressure is determined to be normal, the controlled parameters controlled by the controller of the screw machine are obtained from the actual working parameters;

[0228] Based on the controlled parameters, determine that the control action is abnormal.

[0229] In some embodiments, the second processing module 720 may also be used to:

[0230] Obtaining a first deviation between a controlled parameter and a first target parameter;

[0231] When the first deviation continues to maintain a consistent deviation within the first time period, it is determined that the control action is abnormal.

[0232] The screw machine fault detection device in the embodiment of the present application can be an electronic device, or a component in the electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal, or it can be other devices other than the terminal. Exemplarily, the electronic device can be a mobile phone, a tablet computer, a laptop computer, a PDA, a vehicle-mounted electronic device, a mobile Internet device (Mobile Internet Device, MID), an augmented reality (augmented reality, AR) / virtual reality (virtual reality, VR) device, a robot, a wearable device, an ultra-mobile personal computer (ultra-mobile personal computer, UMPC), a netbook or a personal digital assistant (personal digital assistant, PDA), etc., and can also be a server, a network attached storage (Network Attached Storage, NAS), a personal computer (personal computer, PC), a television (television, TV), a teller machine or a self-service machine, etc., and the embodiment of the present application is not specifically limited.

[0233] The screw machine fault detection device in the embodiment of the present application can be a device with an operating system. The operating system can be an Android operating system, an IOS operating system, or other possible operating systems, which are not specifically limited in the embodiment of the present application.

[0234] The screw machine fault detection device provided in the embodiment of the present application can achieve Figures 1 to 6 To avoid repetition, the various processes implemented by the method embodiment are not described here.

[0235] An embodiment of the present application further provides a chiller, comprising: a screw machine and a screw machine fault detection device as described in any of the above embodiments.

[0236] Wherein, the screw machine fault detection device is electrically connected to the screw machine.

[0237] The screw machine fault detection device is used to execute the screw machine fault detection method as described in any of the above embodiments.

[0238] According to the chiller provided in the embodiment of the present application, by performing trainability tests and diagnosability tests on the acquired real-time working parameters, the best detection mode is selected based on the actual data acquired. It is suitable for fault detection of various types of screw machines, has high versatility, can detect various fault types, and has high detection precision and accuracy.

[0239] Figure 8An example of a structure of a chiller 800 is provided, including a processor 801, a memory 802, and a computer program stored in the memory 802 and executable on the processor 801. When the program is executed by the processor 801, each process of the above-mentioned screw machine fault detection method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0240] It should be noted that the electronic devices in the embodiments of the present application include the mobile electronic devices and non-mobile electronic devices mentioned above.

[0241] An embodiment of the present application also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the various processes of the above-mentioned screw machine fault detection method embodiment are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0242] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.

[0243] An embodiment of the present application also provides a computer program product, including a computer program, which implements the above-mentioned screw machine fault detection method when executed by a processor.

[0244] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.

[0245] An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned screw machine fault detection method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0246] It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0247] It should be noted that, in this article, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0248] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, a disk, or an optical disk), and includes a number of instructions for a terminal (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0249] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

[0250] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0251] Although the embodiments of the present application have been shown and described, those skilled in the art will appreciate that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present application, and that the scope of the present application is defined by the claims and their equivalents.

Claims

1. A screw machine fault detection method, characterized in that: include: Perform trainability detection and diagnosability detection on the actual working parameters of the screw machine to be tested based on the candidate screw machine types in the online model to obtain trainable parameters and diagnosable parameters; Based on at least one of the trainable parameters and the diagnosable parameters, control is performed to enter a target mode, wherein the target mode includes at least one of a precise diagnosis mode, a model training mode, and a basic diagnosis mode.

2. The screw machine fault detection method according to claim 1, characterized in that: The trainability detection and diagnosability detection are performed on the actual working parameters of the screw machine to be tested based on the candidate screw machine types in the online model to obtain the trainable parameters and the diagnosable parameters, including: Perform multi-dimensional coverage calculation based on the candidate screw machine types to obtain the diagnosable parameters; The trainable parameter is obtained based on at least two of the candidate screw machine type, the density of the actual operating parameter, and the duration of the actual operating parameter.

3. The screw machine fault detection method according to claim 1, characterized in that: The controlling entering into the target mode based on at least one of the trainable parameter and the diagnosable parameter comprises: When the diagnosable parameter is greater than a first threshold, controlling entry into the precise diagnosis mode; When the diagnosable parameter is less than a second threshold, controlling entry into the precise diagnosis mode and the model training mode; When the trainable parameter is greater than a third threshold and the diagnosable parameter is less than a fourth threshold, controlling entry into the model training mode; In a case where the diagnosable parameter is not greater than the first threshold or the trainable parameter is not greater than the third threshold, control enters the basic diagnostic mode.

4. The screw machine fault detection method according to claim 3, characterized in that: At least one of the third threshold and the fourth threshold is a Pareto optimal value determined based on the initial normal operating parameters corresponding to the type of the screw machine to be tested and based on the fault diagnosis time and the fault diagnosis accuracy.

5. The screw machine fault detection method according to any one of claims 1 to 4, characterized in that: Controlling entry into the precise diagnosis mode includes: Based on the actual working parameters, obtaining oil pressure fault information and control action fault information; When it is determined that the oil pressure and the control action are normal, the actual working parameters are input into the online model to obtain the predicted current value output by the online model; determining that the solenoid valve is faulty when a second deviation between the actual current value and the predicted current value exceeds a maximum current error tolerance; When a second deviation between the actual current value and the predicted current value does not exceed the maximum current error tolerance, it is determined that the solenoid valve is normal.

6. The screw machine fault detection method according to claim 5, characterized in that: The maximum current error tolerance is determined based on the type of the screw machine to be tested, the model training accuracy and the sampling frequency corresponding to the actual working parameters.

7. The screw machine fault detection method according to any one of claims 1 to 4, characterized in that: Controlling entry into the basic diagnostic mode includes: Based on the actual working parameters, oil pressure fault information and control action fault information are obtained.

8. The screw machine fault detection method according to claim 7, characterized in that: The obtaining of oil pressure fault information and control action fault information based on the actual working parameters includes: The actual working parameters are subjected to at least one of state range analysis, outlier detection, dynamic characteristic analysis and neighborhood analysis to determine oil pressure abnormality.

9. The screw machine fault detection method according to claim 8, characterized in that: The obtaining of oil pressure fault information and control action fault information based on the actual working parameters includes: When it is determined that the oil pressure is normal, obtaining a controlled parameter controlled by a controller of the screw machine from the actual working parameter; Based on the controlled parameter, it is determined that the control action is abnormal.

10. The screw machine fault detection method according to claim 9, characterized in that: The determining, based on the controlled parameter, that the control action is abnormal, includes: Obtaining a first deviation between the controlled parameter and a first target parameter; When the first deviation continues to maintain a consistent deviation within a first time period, it is determined that the control action is abnormal.

11. A screw machine fault detection device, characterized in that: include: A first processing module is used to perform trainability detection and diagnosability detection on actual working parameters of the screw machine to be tested based on the candidate screw machine types in the online model, and obtain trainable parameters and diagnosable parameters; The second processing module is used to control entering into a target mode based on at least one of the trainable parameters and the diagnosable parameters, wherein the target mode includes at least one of a precise diagnosis mode, a model training mode and a basic diagnosis mode.

12. A chiller, characterized in that: include: Screw machine; The screw machine fault detection device according to claim 11, wherein the screw machine fault detection device is electrically connected to the screw machine.

13. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the screw machine fault detection method according to any one of claims 1 to 10 is implemented.

14. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the screw machine fault detection method according to any one of claims 1 to 10 is implemented.

Images