A method and system for fault monitoring of electrical components of a mixing station

Abstract

The present application relates to the technical field of industrial monitoring and intelligent fault diagnosis. A kind of fault monitoring method and system for mixing station electrical element are provided. The steps are as follows: the operating signal of mixing station electrical element is collected;Based on the operating signal, signal anti-interference processing is carried out to obtain preprocessed signal;According to the preprocessed signal, the fault characteristic parameters of mixing station electrical element are extracted, including the step-down starting conversion logic feature of contactor and the overload transient feature of motor;Based on the fault characteristic parameters, fault trend analysis is carried out to identify component performance degradation trend and potential failure risk;According to the result of the fault trend analysis, generate multi-level early warning signal, for triggering preventive maintenance operation. Solve the problem of failure feature extraction failure, signal distortion and potential risk warning loss caused by insufficient pertinence, weak anti-interference ability and low intelligent degree of existing mixing station electrical element monitoring technology.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial monitoring and intelligent fault diagnosis, in particular to a fault monitoring method and system for electrical elements of a mixing station. BACKGROUND

[0002] In the field of construction engineering, the mixing station is a key facility for producing materials such as concrete, and its stable and efficient operation depends on the normal work of numerous electrical elements (such as motors, contactors, relays, sensors, etc.). However, these elements are long-term exposed to harsh working conditions of large dust, strong vibration, and complex electromagnetic interference, and faults occur frequently.

[0003] Existing industrial monitoring technologies (such as SCADA systems) have been applied to the monitoring of some equipment, but there are still obvious limitations in the fault monitoring of electrical elements of mixing stations:

[0004] Lack of pertinence, general systems do not fully consider the operating characteristics and fault modes of specific electrical elements of mixing stations (such as AC contactor step-down starting conversion logic and motor overload transient characteristics), making it difficult to extract effective fault feature parameters;

[0005] Weak anti-interference ability, strong electromagnetic environment of the mixing station easily causes the collected voltage, current and other signals to be polluted by power frequency interference and high frequency noise, and traditional single filtering method is difficult to ensure signal quality, affecting the accuracy of subsequent analysis;

[0006] Low degree of intelligence, existing solutions are mostly based on simple threshold alarms, lack of intelligent analysis of running data trends and multi-level warning mechanisms, and cannot effectively identify potential fault risks (such as slow degradation of element performance), making it difficult to provide data support for preventive maintenance. SUMMARY

[0007] The purpose of the present application is to provide a fault monitoring method and system for electrical elements of a mixing station, aiming to solve the problems of failure of fault feature extraction, signal distortion and lack of potential risk warning caused by lack of pertinence, weak anti-interference ability and low degree of intelligence in existing mixing station electrical element monitoring technologies.

[0008] The present application is achieved by the following technical solutions:

[0009] A fault monitoring method for electrical elements of a mixing station, comprising the following steps:

[0010] Collecting the operating signals of the electrical elements of the mixing station, the operating signals including voltage signals and current signals;

[0011] Based on the operating signals, performing signal anti-interference processing to suppress power frequency interference and high frequency noise pollution, and obtaining preprocessed signals;

[0012] According to the preprocessed signal, a fault feature parameter of the electrical element of the mixing station is extracted, and the fault feature parameter includes a step-down starting conversion logic feature of a contactor and an overload transient feature of a motor.

[0013] Based on the fault feature parameter, a fault trend analysis is performed to identify a performance degradation trend and a potential fault risk of the element.

[0014] According to the result of the fault trend analysis, a multi-level warning signal is generated to trigger a preventive maintenance operation.

[0015] Optionally, the operation signal of the electrical element of the mixing station is collected, and the specific process of the operation signal includes a voltage signal and a current signal.

[0016] For the core electrical elements of the AC contactor and the motor in the mixing station, corresponding data collection modules are configured.

[0017] The voltage signal of the electrical element is isolated and sensed by a voltage transformer, and a true RMS measurement chip is used to continuously collect the true RMS value and the peak value of the voltage waveform in real time.

[0018] The current signal is collected by a current sensor, and the voltage signal and the current signal are transmitted to the signal processing end through the RS-485 bus interface according to the MODBUS-RTU protocol.

[0019] Optionally, based on the operation signal, signal anti-interference processing is performed to suppress power frequency interference and high-frequency noise pollution to obtain a preprocessed signal, and the specific process is as follows:

[0020] The voltage signal and the current signal from the RS-485 bus interface are received.

[0021] The voltage signal and the current signal are signal-conditioned, including amplification and baseline correction, to eliminate the offset error introduced by the sensor.

[0022] An adaptive band-pass filter is used to suppress power frequency interference, and the center frequency is dynamically adjusted to the power frequency of the power grid.

[0023] A wavelet threshold denoising filter is used to suppress high-frequency noise pollution, and high-frequency noise components are removed through wavelet decomposition and reconstruction.

[0024] The filtered signal is subjected to a moving average process to smooth transient interference and obtain a stable preprocessed signal.

[0025] Optionally, the specific process of extracting the fault feature parameter of the electrical element of the mixing station according to the preprocessed signal is as follows:

[0026] In the contactor starting stage, the current waveform in the pretreatment signal is divided into time windows to identify the action timing of the step-down starting contactor and the running contactor; the current conversion time difference between the step-down starting stage and the full-pressure running stage is calculated, and the logic interlocking state of the contactor contact action signal in the two stages is detected; if the conversion time difference exceeds the preset threshold or the logic interlocking failure is detected, the conversion timeout feature parameter or the logic conflict feature parameter is extracted respectively;

[0027] The real-time monitoring of the motor current effective value in the pretreatment signal is performed; the current effective value instantaneous surge event is captured, and the surge amplitude and duration are recorded; the energy integral or equivalent thermal effect of the surge event is calculated; if the surge amplitude exceeds the rated current set multiple, the duration is greater than the minimum overload maintenance time, or the energy integral exceeds the safety threshold, the transient overload amplitude feature parameter, the overload duration feature parameter, or the overload energy feature parameter is extracted respectively.

[0028] Optionally, the specific process of performing fault trend analysis based on the fault feature parameter to identify the element performance degradation trend and the potential fault risk is:

[0029] Periodically collect the conversion timeout feature parameter, the logic conflict feature parameter, the transient overload amplitude feature parameter, the overload duration feature parameter, and the overload energy feature parameter to construct a time sequence feature database;

[0030] The trigger frequency of the conversion timeout feature parameter in a preset period is counted, and the sliding change rate of the time difference average of the trigger frequency is calculated; the duration of the continuous appearance of the logic conflict feature parameter is detected, and the action number of the corresponding contactor is associated; if the monthly change rate of the conversion time difference average exceeds the monthly change threshold or the weekly increase amplitude of the logic conflict frequency is greater than the weekly increase amplitude threshold, the contactor contact wear degradation is determined;

[0031] The transient overload amplitude feature parameter is subjected to exponential smoothing processing to extract the long-term rising slope of the overload event amplitude; the overload duration feature parameter and the overload energy feature parameter are aggregated to calculate the equivalent cumulative thermal stress coefficient; if the amplitude rising slope is higher than 2 times the standard deviation of the baseline value for 3 consecutive periods, or the cumulative thermal stress coefficient breaks through the insulation aging threshold, the motor winding performance degradation is determined.

[0032] Optionally, the specific process of generating a multi-level warning signal according to the result of the fault trend analysis for triggering the preventive maintenance operation is:

[0033] Based on the contactor contact wear degradation state or the motor winding performance degradation state identified in the fault trend analysis, a hierarchical warning decision is made:

[0034] When the contactor contact wear degradation state is identified, a first-level early warning signal is generated, triggering a contactor contact lubrication maintenance instruction; if the contact wear degradation state continues to deteriorate and the logic conflict frequency breaks through the safety threshold, a second-level early warning signal is upgraded, triggering a contactor replacement operation instruction;

[0035] When the motor winding performance degradation state is identified, a first-level early warning signal is generated, triggering a motor load optimization and heat dissipation system inspection instruction; if the equivalent cumulative thermal stress coefficient continues to exceed the insulation aging threshold and the amplitude rising slope reaches the critical slope, a third-level early warning signal is upgraded, triggering a motor winding insulation reinforcement or replacement operation instruction;

[0036] The multi-level early warning signals are pushed to the remote operation and maintenance platform through the industrial Internet of Things gateway, and the local audible and visual alarm devices of the mixing station are synchronously activated.

[0037] Optionally, the fault trend analysis includes: based on the fault characteristic parameters, using a pre-trained machine learning model to intelligently predict the element performance degradation trend and potential fault risk; wherein the machine learning model is trained by historical fault characteristic parameter data, used to optimize the fault identification accuracy, and outputs the prediction result to undertake the generation of the multi-level early warning signal.

[0038] Based on the same inventive concept, the application also provides a fault monitoring system for electrical elements of a mixing station, for realizing the fault monitoring method for electrical elements of a mixing station, comprising:

[0039] The signal acquisition module is used for configuring corresponding data acquisition modules for the AC contactor and the motor core electrical elements in the mixing station, isolating and sensing the voltage signals of the electrical elements through the voltage transformer, and using the true RMS measurement chip to continuously collect the true RMS value and peak value of the voltage waveform in real time, collecting the current signal through the current sensor, and transmitting the voltage signal and the current signal to the signal processing end through the RS-485 bus interface according to the MODBUS-RTU protocol;

[0040] The signal anti-interference processing module is used for receiving the voltage signal and the current signal from the RS-485 bus interface, and performing signal conditioning on the voltage signal and the current signal, including amplification and baseline correction, eliminating the offset error introduced by the sensor, using an adaptive bandpass filter to suppress power frequency interference, with the center frequency dynamically adjusted to the power frequency of the power grid, using a wavelet threshold denoising filter to suppress high-frequency noise pollution, removing high-frequency noise components through wavelet decomposition and reconstruction, and performing sliding average processing on the filtered signal to smooth transient interference, obtaining a stable preprocessed signal;

[0041] The fault feature parameter extraction module is configured to divide the current waveform in the pretreatment signal into time windows during the contactor starting stage, identify the action timing of the step-down starting contactor and the running contactor, calculate the current conversion time difference between the step-down starting stage and the full-pressure running stage, and detect the logic interlocking state of the contactor contact action signal in the two stages. If the conversion time difference exceeds the preset threshold or the logic interlocking failure is detected, the conversion timeout feature parameter or the logic conflict feature parameter is extracted, respectively. Meanwhile, the RMS of the motor current in the pretreatment signal is monitored in real time, the current RMS instantaneous surge event is captured, the surge amplitude and duration are recorded, the energy integral or equivalent thermal effect of the surge event is calculated, and if the surge amplitude exceeds the rated current set multiple, the duration is greater than the minimum overload maintenance time, or the energy integral exceeds the safety threshold, the transient overload amplitude feature parameter, the overload duration feature parameter, or the overload energy feature parameter is extracted, respectively.

[0042] The fault trend analysis module is configured to periodically collect the conversion timeout feature parameter, the logic conflict feature parameter, the transient overload amplitude feature parameter, the overload duration feature parameter, and the overload energy feature parameter, construct a time sequence feature database, count the trigger frequency of the conversion timeout feature parameter in a preset period, calculate the sliding change rate of the time difference average of the trigger frequency, detect the duration of the continuous appearance of the logic conflict feature parameter, associate the action number of the corresponding contactor, and determine the contactor contact wear degradation if the monthly change rate of the conversion time difference average exceeds the monthly change threshold or the weekly increase amplitude of the logic conflict frequency is greater than the weekly increase amplitude threshold. The transient overload amplitude feature parameter is subjected to exponential smoothing processing, the long-term rising slope of the overload event amplitude is extracted, the overload duration feature parameter and the overload energy feature parameter are aggregated, the equivalent cumulative thermal stress coefficient is calculated, and the motor winding performance degradation is determined if the amplitude rising slope is higher than 2 times the standard deviation of the baseline value for 3 consecutive periods or the cumulative thermal stress coefficient breaks through the insulation aging threshold. Furthermore, based on the fault feature parameters, the element performance degradation trend and potential fault risk are intelligently predicted by using a pre-trained machine learning model.

[0043] The multi-stage early warning signal generation module is used for executing a hierarchical early warning decision based on the contactor contact wear degradation state or the motor winding performance degradation state identified in the fault trend analysis, generating a first-stage early warning signal when the contactor contact wear degradation state is identified, triggering a contactor contact lubrication maintenance instruction, upgrading to a second-stage early warning signal if the contact wear degradation state continues to deteriorate and the logic conflict frequency breaks through a safety threshold, triggering a contactor replacement operation instruction, generating a first-stage early warning signal when the motor winding performance degradation state is identified, triggering a motor load optimization and heat dissipation system inspection instruction, upgrading to a third-stage early warning signal if the equivalent cumulative thermal stress coefficient continues to exceed an insulation aging threshold and the amplitude rising slope reaches a critical slope, triggering a motor winding insulation reinforcement or replacement operation instruction, pushing the multi-stage early warning signal to a remote operation and maintenance platform through an industrial Internet of Things gateway, and synchronously activating a local audible and visual alarm device of the mixing station.

[0044] Based on the same inventive concept, the present application also provides an electronic device comprising a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the above-mentioned fault monitoring method for electrical elements of a mixing station.

[0045] Based on the same inventive concept, the present application also provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the above-mentioned fault monitoring method for electrical elements of a mixing station.

[0046] The technical solution of the present application has at least the following advantages and beneficial effects:

[0047] For the operation characteristics and fault modes of specific electrical elements of a mixing station, the fault feature parameters such as the step-down starting conversion logic feature of a contactor and the overload transient feature of a motor are accurately extracted, which can better fit the actual operation of electrical elements of a mixing station compared with general industrial monitoring systems, effectively overcomes the problem of insufficient monitoring pertinence in the prior art, and significantly improves the effectiveness and accuracy of fault feature extraction.

[0048] Through special anti-interference processing of the collected voltage, current and other operation signals, the power frequency interference and high-frequency noise pollution in the strong electromagnetic environment of a mixing station can be effectively suppressed, the signal quality can be better guaranteed compared with a traditional single filtering method, analysis errors caused by signal interference can be avoided, and a more reliable data basis is provided for subsequent fault analysis.

[0049] Break through the limitations of the existing scheme simple threshold alarm, based on the extracted fault characteristic parameters for fault trend analysis, can effectively identify the performance degradation trend and potential failure risk of the element, and generate multi-level warning signals, which can realize the deep intelligent analysis of the running state of the electrical element, provide strong data support for preventive maintenance, change the previous difficulty to find potential failure in advance, reduce the probability of equipment sudden failure, ensure the stable and efficient operation of the mixing station, and reduce the economic loss and time delay caused by failure downtime. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 A flowchart of a fault monitoring method for electrical elements of a mixing station according to an embodiment of the present application is shown in the figure.

[0051] Figure 2 A structural diagram of a fault monitoring system for electrical elements of a mixing station according to an embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0052] The following is a specific embodiment combined with the drawings.

[0053] Reference Figure 1 A fault monitoring method for electrical elements of a mixing station, comprising the following steps:

[0054] Step one, collecting the running signals of the electrical elements of the mixing station, the running signals including voltage signals and current signals.

[0055] In some embodiments, the specific process of collecting the running signals of the electrical elements of the mixing station, the running signals including voltage signals and current signals, is as follows:

[0056] For the AC contactor and the core electrical elements of the motor in the mixing station, corresponding data acquisition modules are configured;

[0057] The voltage signals of the electrical elements are isolated and sensed by a voltage transformer, and a true RMS measurement chip is used to continuously collect the true RMS value and peak value of the voltage waveform in real time;

[0058] The current signals are collected by a current sensor, and the voltage signals and current signals are transmitted to the signal processing end according to the MODBUS-RTU protocol through the RS-485 bus interface.

[0059] Step two, based on the running signals, signal anti-interference processing is performed to suppress power frequency interference and high frequency noise pollution, and preprocessed signals are obtained.

[0060] In some embodiments, the specific process of performing signal anti-interference processing based on the running signals to suppress power frequency interference and high frequency noise pollution to obtain preprocessed signals is as follows:

[0061] Receiving voltage and current signals from RS-485 bus interface;

[0062] Signal conditioning of the voltage and current signals, including amplification and baseline correction, to eliminate offset errors introduced by the sensor;

[0063] Using an adaptive band-stop filter to suppress power frequency interference, with the center frequency dynamically adjusted to the power grid power frequency;

[0064] Using a wavelet threshold denoising filter to suppress high-frequency noise pollution, removing high-frequency noise components through wavelet decomposition and reconstruction;

[0065] Performing a moving average process on the filtered signal to smooth transient interference and obtain a stable preprocessed signal.

[0066] Step three, extracting fault feature parameters of the electrical elements of the mixing station according to the preprocessed signal, the fault feature parameters including the step-down start transition logic feature of the contactor and the overload transient feature of the motor.

[0067] In some embodiments, the specific process of extracting fault feature parameters of the electrical elements of the mixing station according to the preprocessed signal is as follows:

[0068] In the contactor start-up phase, the current waveform in the preprocessed signal is divided into time windows to identify the action timing of the step-down start contactor and the running contactor; the current transition time difference between the step-down start phase and the full-pressure running phase is calculated, and the logic interlocking state of the contactor contact action signal in the two phases is detected; if the transition time difference exceeds the preset threshold or the logic interlocking fails is detected, the transition timeout feature parameter or the logic conflict feature parameter is extracted respectively;

[0069] Real-time monitoring of the motor current effective value in the preprocessed signal; capturing the instantaneous surge event of the current effective value, recording the surge amplitude and duration; calculating the energy integral or equivalent thermal effect of the surge event; if the surge amplitude exceeds the rated current set multiple, the duration is greater than the minimum overload maintenance time, or the energy integral exceeds the safety threshold, the transient overload amplitude feature parameter, the overload duration feature parameter, or the overload energy feature parameter is extracted respectively.

[0070] Step four, based on the fault feature parameters, performing fault trend analysis to identify the performance degradation trend and potential fault risk of the elements.

[0071] In some embodiments, the specific process of performing fault trend analysis based on the fault feature parameters to identify the performance degradation trend and potential fault risk of the elements is as follows:

[0072] Periodically collect the conversion timeout feature parameters, the logical conflict feature parameters, the transient overload amplitude feature parameters, the overload duration feature parameters and the overload energy feature parameters, and construct a time sequence feature database;

[0073] Statistically count the trigger frequency of the conversion timeout feature parameters in a preset period, calculate the sliding change rate of the time difference mean of the trigger frequency, detect the duration of the continuous appearance of the logical conflict feature parameters, associate the action times of the corresponding contactors, and determine the contactor contact wear degradation if the monthly change rate of the conversion time difference mean exceeds a monthly change threshold or the weekly increase amplitude of the logical conflict frequency is greater than a weekly increase amplitude threshold;

[0074] Exponentially smooth the transient overload amplitude feature parameters, extract the long-term rising slope of the overload event amplitude, aggregate the overload duration feature parameters and the overload energy feature parameters, and calculate the equivalent cumulative thermal stress coefficient; if the amplitude rising slope is higher than 2 times the standard deviation of the baseline value for 3 consecutive periods, or the cumulative thermal stress coefficient breaks through the insulation aging threshold, the motor winding performance degradation is determined.

[0075] Step five, according to the result of the fault trend analysis, a multi-level warning signal is generated to trigger a preventive maintenance operation.

[0076] In some embodiments, the specific process of generating a multi-level warning signal according to the result of the fault trend analysis for triggering a preventive maintenance operation is:

[0077] Based on the contactor contact wear degradation state or the motor winding performance degradation state identified in the fault trend analysis, a hierarchical warning decision is made:

[0078] When the contactor contact wear degradation state is identified, a first-level warning signal is generated to trigger a contactor contact lubrication maintenance instruction; if the contact wear degradation state continues to deteriorate and the logical conflict frequency breaks through the safety threshold, it is upgraded to a second-level warning signal to trigger a contactor replacement operation instruction;

[0079] When the motor winding performance degradation state is identified, a first-level warning signal is generated to trigger a motor load optimization and heat dissipation system inspection instruction; if the equivalent cumulative thermal stress coefficient continuously exceeds the insulation aging threshold and the amplitude rising slope reaches the critical slope, it is upgraded to a third-level warning signal to trigger a motor winding insulation reinforcement or replacement operation instruction;

[0080] The multi-level warning signal is pushed to the remote operation and maintenance platform through the industrial Internet of Things gateway, and the local sound and light alarm device of the mixing station is synchronously activated.

[0081] In some embodiments, the fault trend analysis comprises: based on the fault characteristic parameters, using a pre-trained machine learning model to intelligently predict the element performance degradation trend and potential fault risk; wherein the machine learning model is trained by historical fault characteristic parameter data, is used to optimize the fault identification accuracy, and outputs a prediction result to undertake the generation of the multi-level early warning signal.

[0082] The historical fault characteristic parameter data can be collected from multiple mixing station sites for a long time to form a time series data set, which includes:

[0083] The contactor-related features include a conversion timeout characteristic parameter (denoted as , in seconds) and a logic conflict characteristic parameter (denoted as , which is a binary value or a frequency count);

[0084] The motor-related features include a transient overload amplitude characteristic parameter (denoted as , in amperes), an overload duration characteristic parameter (denoted as , in seconds), and an overload energy characteristic parameter (denoted as , in joules);

[0085] Label data: based on actual maintenance records or expert diagnosis, the element performance degradation state at each time point is labeled (as a label for supervised learning); the label adopts discrete levels:

[0086] Contactors: 0 (normal), 1 (slight wear degradation), and 2 (severe wear degradation);

[0087] Motors: 0 (normal), 1 (slight insulation degradation), and 2 (severe insulation degradation).

[0088] The characteristic parameters are collected at a fixed period (such as daily or weekly) to construct a time series.

[0089] The time series data is converted into a supervised learning format. For each time point , a sliding window is used to extract the feature sequence of the past periods as the model input. For example, the window size (representing data from the past 30 days), the input sequence is , where is the feature vector at time (for contactors, ; for motors, ). To simplify, the contactor and the motor can be trained independently to avoid feature coupling.

[0090] The characteristic parameters are z-score standardized to eliminate the dimensional influence. The historical data are divided into training set (70%), validation set (15%), and test set (15%) in chronological order to ensure the integrity of the time sequence.

[0091] The long short-term memory network (LSTM) is used as the core machine learning model. LSTM is good at processing long-term dependencies of time series data and suitable for capturing performance degradation trends (such as progressive wear of contactor contacts or cumulative thermal stress effects of motor windings). The model structure includes an input layer, an LSTM layer, a fully connected layer, and an output layer. The input layer receives a sliding window sequence with a dimension of , where represents the feature dimension. In this example, the contactor model and the motor model ; the LSTM layer includes multiple LSTM units to process sequences and extract time series features. Each LSTM unit contains a forget gate, an input gate, and an output gate to dynamically learn important information. The fully connected layer is used to convert the LSTM output into a high-level representation. The output layer uses the softmax activation function to output the predicted probability distribution of the degradation state, which includes normal, slight degradation, and severe degradation.

[0092] The categorical cross-entropy loss is used to optimize the difference between the predicted label and the true label, as shown in equation (1):

[0093]

[0094] where represents the number of samples; represents the true label (one-hot encoding) of sample ; and represents the probability that the model predicts that sample belongs to class .

[0095] The Adam optimizer is used with a learning rate of 0.001, and early stopping is used to prevent overfitting. The training is stopped when the validation set loss does not decrease for 10 consecutive epochs. The loss function is minimized while monitoring the validation set accuracy. After training is complete, the model can predict the degradation state probability for the next period (e.g., the next day).

[0096] After step three is completed, the system obtains the latest fault characteristic parameters; updates the sliding window sequence by removing the oldest data and adding the latest data; inputs the pre-trained LSTM model and outputs the predicted probability distribution, i.e., the predicted class that the sample belongs to. the probability of ; based on the probability to generate the degradation state prediction, take as the degradation level, that is, from the three degradation state probabilities predicted by the model, select the state with the highest probability value as the final determined element performance degradation level.

[0097] The output of the machine learning model can be used as a supplement or optimization of existing rules:

[0098] When the model predicts the degradation level as 1 or 2, it is preferred to use the model result because it is based on data-driven and has higher accuracy;

[0099] When the model prediction is uncertain, such as the highest probability being less than 0.6, fall back to the original rule analysis in step four, such as calculating the monthly change rate;

[0100] The output prediction result (degradation level and probability) is directly input to step five for generating multi-level warning signals.

[0101] For example, the processing process of a single LSTM unit is as follows:

[0102] Input: time step feature vector (normalized), hidden state vector of the previous time step , cell state vector of the previous time step .

[0103] The gating mechanism is calculated as shown in equation (2):

[0104]

[0105] wherein is the forget gate output vector; is the input gate output vector; is the candidate cell state vector; is the current cell state vector; is the output gate output vector; is the current hidden state vector; denotes the sigmoid function; denotes the hyperbolic tangent function; denotes element-wise multiplication; , , and denote the forget gate weight matrix, the input gate weight matrix, the candidate cell state weight matrix, and the output gate weight matrix, respectively; , , and respectively represent the forget gate bias vector, the input gate bias vector, the candidate cell state bias vector, and the output gate bias vector.

[0106] corresponding to the entire input sequence The LSTM layer iteratively computes all time steps to obtain the final hidden state .

[0107] The output layer is calculated as shown in equation (3) as follows:

[0108]

[0109] wherein, represents the weight matrix of the output layer; represents the bias vector of the output layer; represents the logits vector after linear transformation; represents the predicted probability distribution vector; represents the probability of belonging to the class; represents the softmax activation function; represents the class index; represents the i-th element of the logits vector.

[0110] In step five, the prediction result is directly used to trigger the early warning:

[0111] Contactor early warning decision:

[0112] If the predicted degradation level is 1 (slight wear degradation, probability greater than 0.6), a first-level warning signal (triggering lubrication maintenance) is generated;

[0113] If the predicted degradation level is 2 (severe wear degradation, probability greater than 0.8), and the logic conflict feature parameter breaks through the safety threshold (such as frequency greater than 5 times / week), it is upgraded to a second-level warning signal (triggering replacement operation).

[0114] Motor early warning decision:

[0115] If the predicted degradation level is 1 (slight insulation degradation, probability greater than 0.6), a first-level warning signal (triggering load optimization and heat dissipation inspection) is generated;

[0116] If the predicted degradation level is 2 (severe insulation degradation, probability greater than 0.8), and the overload energy feature parameter continues to exceed the threshold, it is upgraded to a third-level warning signal (triggering insulation reinforcement or replacement).

[0117] ​Compared with the pure rule method, the model provides more refined risk quantification through the probability output (for example, the probability value can be directly used as the basis for adjusting the early warning threshold), and can issue an early warning at an early stage of degradation (for example, when the probability slowly rises), thereby improving the timeliness of preventive maintenance.

[0118] Based on the same inventive concept, corresponding to any of the above embodiments, with reference to Figure 2 The application provides a fault monitoring system for electrical elements of a mixing station, which is used to implement the above-mentioned fault monitoring method for electrical elements of a mixing station, and comprises:

[0119] The operation signal acquisition module is configured to, for the AC contactor and the motor core electrical element in the mixing station, configure a corresponding data acquisition module, isolate and sense the voltage signal of the electrical element through a voltage transformer, and continuously collect the true RMS value and the peak value of the voltage waveform in real time by using a true RMS measurement chip, collect the current signal by using a current sensor, and transmit the voltage signal and the current signal to the signal processing end through an RS-485 bus interface according to a MODBUS-RTU protocol.

[0120] The signal anti-interference processing module is configured to receive the voltage signal and the current signal from the RS-485 bus interface, and perform signal conditioning on the voltage signal and the current signal, including amplification and baseline correction, elimination of offset errors introduced by the sensor, suppression of power frequency interference by using an adaptive bandpass filter, dynamic adjustment of the center frequency to the power frequency of the power grid, suppression of high-frequency noise pollution by using a wavelet threshold denoising filter, removal of high-frequency noise components by wavelet decomposition and reconstruction, and smoothing of transient interference by performing sliding average processing on the filtered signal to obtain a stable preprocessed signal.

[0121] The fault feature parameter extraction module is configured to, in the contactor starting stage, divide the current waveform in the preprocessed signal into time windows, identify the action timing of the step-down starting contactor and the running contactor, calculate the current conversion time difference between the step-down starting stage and the full-pressure running stage, and detect the logic interlocking state of the contactor contact action signal in the two stages, and if the conversion time difference exceeds a preset threshold or the logic interlocking fails is detected, respectively extract a conversion timeout feature parameter or a logic conflict feature parameter, and simultaneously monitor the motor current effective value in the preprocessed signal in real time, capture a current effective value instantaneous surge event, record the surge amplitude and duration, calculate the energy integral or equivalent thermal effect of the surge event, and if the surge amplitude exceeds a set multiple of the rated current, the duration is greater than the minimum overload maintenance time, or the energy integral exceeds a safety threshold, respectively extract a transient overload amplitude feature parameter, an overload duration feature parameter, or an overload energy feature parameter.

[0122] The fault trend analysis module is configured to periodically collect the conversion timeout feature parameter, the logic conflict feature parameter, the transient overload amplitude feature parameter, the overload duration feature parameter and the overload energy feature parameter, construct a time sequence feature database, count the trigger frequency of the conversion timeout feature parameter in a preset period, calculate the sliding change rate of the time difference mean value of the trigger frequency, detect the duration of the continuous occurrence of the logic conflict feature parameter, correlate the action number of the corresponding contactor, and determine that the contactor contact point is worn and degraded if the monthly change rate of the conversion time difference mean value exceeds a monthly change threshold or the weekly increase amplitude of the logic conflict frequency is greater than a weekly increase amplitude threshold. The transient overload amplitude feature parameter is subjected to exponential smoothing processing, the long-term rising slope of the overload event amplitude is extracted, the overload duration feature parameter and the overload energy feature parameter are aggregated, and the equivalent cumulative thermal stress coefficient is calculated. If the amplitude rising slope is higher than 2 times the standard deviation of the baseline value for 3 consecutive periods or the cumulative thermal stress coefficient breaks through an insulation aging threshold, it is determined that the motor winding performance is degraded. The element performance degradation trend and the potential fault risk can also be intelligently predicted based on the fault feature parameters by using a pre-trained machine learning model.

[0123] The multi-level early warning signal generation module is configured to execute a hierarchical early warning decision based on the contactor contact point wear degradation state or the motor winding performance degradation state identified in the fault trend analysis. When the contactor contact point wear degradation state is identified, a first-level early warning signal is generated to trigger a contactor contact point lubrication maintenance instruction. If the contact point wear degradation state continues to deteriorate and the logic conflict frequency breaks through a safety threshold, a second-level early warning signal is upgraded to trigger a contactor replacement operation instruction. When the motor winding performance degradation state is identified, a first-level early warning signal is generated to trigger a motor load optimization and heat dissipation system inspection instruction. If the equivalent cumulative thermal stress coefficient continuously exceeds the insulation aging threshold and the amplitude rising slope reaches a critical slope, a third-level early warning signal is upgraded to trigger a motor winding insulation reinforcement or replacement operation instruction. The multi-level early warning signals are pushed to a remote operation and maintenance platform through an industrial Internet of Things gateway to synchronously activate a local audible and visual alarm device of the mixing station.

[0124] Based on the same inventive concept, the present application provides an electronic device comprising a memory and a processor. The memory is configured to store a computer program, and the processor is configured to run the computer program to enable the electronic device to execute the fault monitoring method for electrical elements of a mixing station according to any of the above embodiments.

[0125] Optionally, the electronic device can be a server.

[0126] In addition, the present embodiment also provides a computer readable storage medium having a computer program stored thereon. The computer program is executed by a processor to implement the fault monitoring method for electrical elements of a mixing station according to the embodiments.

[0127] It is appreciated that the processor in the embodiments of the present application can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. The general-purpose processor can be a microprocessor or any conventional processor.

[0128] The method steps in the embodiments of the present application can be realized by hardware or by the processor executing software instructions. The software instructions can be composed of corresponding software modules, and the software modules can be stored in a random access memory (RAM), a flash memory, a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically-erasable programmable read-only memory (EEPROM), a register, a hard disk, a mobile hard disk, a CD-ROM or any other form of storage medium well known in the art. An exemplary storage medium is coupled to the processor, so that the processor can read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an ASIC.

[0129] In the embodiments described above, all or some of the embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or some of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded into and executed by a computer, all or some of the procedures or functions according to the embodiments of the present application are generated. The computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions can be stored in or transmitted from a storage medium. The computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through a wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) manner. The storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. integrated with one or more available media. The available media can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, a solid state disk (SSD)), etc.

Claims

1. A fault monitoring method for electrical components of a mixing plant, characterized in that: The following steps are involved: Collecting operating signals of electrical components of the mixing plant, wherein the operating signals include voltage signals and current signals; Based on the operating signal, signal anti-interference processing is performed to suppress power frequency interference and high-frequency noise pollution to obtain a preprocessed signal; Extracting fault characteristic parameters of electrical components of the mixing plant based on the preprocessed signals, wherein the fault characteristic parameters include the voltage reduction start-up conversion logic characteristics of the contactor and the overload transient characteristics of the motor; The specific process of extracting the fault characteristic parameters of the mixing station electrical components based on the pre-processed signals is as follows: During the contactor startup phase, the current waveform in the pre-processed signal is divided into time windows to identify the action sequence of the reduced-voltage starting contactor and the running contactor. The current conversion time difference between the reduced-voltage startup phase and the full-voltage operation phase is calculated, and the logical interlocking state of the contactor contact action signal in the two phases is detected. If the conversion time difference exceeds a preset threshold or a logic interlock failure is detected, it is extracted as a conversion timeout characteristic parameter or a logic conflict characteristic parameter respectively; Real-time monitoring of the effective value of the motor current in the pre-processed signal; capturing the instantaneous sudden increase of the effective value of the current, and recording the sudden increase magnitude and duration; Calculate the energy integral of the sudden increase event; if the sudden increase exceeds the set multiple of the rated current, the duration is longer than the minimum overload maintenance time, or the energy integral exceeds the safety threshold, then extract it as the transient overload amplitude characteristic parameter, overload duration characteristic parameter, or overload energy characteristic parameter respectively; Based on the fault characteristic parameters, perform fault trend analysis to identify component performance degradation trends and potential failure risks; The specific process of performing fault trend analysis based on the fault characteristic parameters and identifying component performance degradation trends and potential fault risks is as follows: Periodically collecting the conversion timeout characteristic parameters, logic conflict characteristic parameters, transient overload amplitude characteristic parameters, overload duration characteristic parameters, and overload energy characteristic parameters to construct a time series characteristic database; Count the triggering frequency of the conversion timeout characteristic parameter within a preset period and calculate the sliding change rate of the time difference mean of the triggering frequency; detect the continuous occurrence period of the logic conflict characteristic parameter and associate it with the number of corresponding contactor operations; if the monthly change rate of the time difference mean exceeds the monthly change threshold or the weekly increase in the logic conflict frequency exceeds the weekly increase threshold, it is determined that the contactor contacts are worn and degraded; Exponential smoothing is performed on the transient overload amplitude characteristic parameters to extract the long-term rising slope of the overload event amplitude; Aggregate the overload duration characteristic parameters and the overload energy characteristic parameters to calculate the equivalent cumulative thermal stress coefficient. If the amplitude rising slope is higher than the baseline value by 2 standard deviations for three consecutive cycles, or the cumulative thermal stress coefficient exceeds the insulation aging threshold, the motor winding performance is determined to be degraded. Based on the results of the fault trend analysis, a multi-level early warning signal is generated to trigger a preventive maintenance operation.

2. The fault monitoring method for electrical components of a mixing plant according to claim 1, characterized in that: The specific process of collecting the operating signals of the electrical components of the mixing station, which include voltage signals and current signals, is as follows: Configure corresponding data acquisition modules for the AC contactors and motor core electrical components in the mixing plant; The voltage signal of the electrical component is isolated and sensed through a voltage transformer, and a true RMS measurement chip is used to continuously collect the true RMS value and peak value of the voltage waveform in real time; The current signal is collected by the current sensor, and the voltage signal and the current signal are transmitted to the signal processing end via the RS-485 bus interface according to the MODBUS-RTU protocol.

3. The fault monitoring method for electrical components of a mixing plant according to claim 2, characterized in that: The specific process of performing signal anti-interference processing based on the operating signal to suppress power frequency interference and high-frequency noise pollution to obtain the preprocessed signal is as follows: Receive voltage and current signals from the RS-485 bus interface; Performing signal conditioning on the voltage signal and the current signal, including amplification and baseline correction, to eliminate offset errors introduced by the sensor; Adaptive band-stop filters are used to suppress power frequency interference, with the center frequency dynamically adjusted to the power frequency of the grid; Use wavelet threshold denoising filter to suppress high-frequency noise pollution, and remove high-frequency noise components through wavelet decomposition and reconstruction; The filtered signal is processed by sliding average to smooth transient interference and obtain a stable preprocessed signal.

4. The fault monitoring method for electrical components of a mixing plant according to claim 1, characterized in that: The specific process of generating a multi-level warning signal based on the results of the fault trend analysis to trigger a preventive maintenance operation is as follows: Based on the wear and degradation of the contactor contacts or the performance degradation of the motor windings identified in the fault trend analysis, a hierarchical early warning decision is made: When the contactor contact wear and degradation state is identified, a first-level warning signal is generated, triggering the contactor contact lubrication maintenance instruction; If the contact wear and degradation continues to worsen and the logic conflict frequency exceeds the safety threshold, it will be upgraded to a secondary warning signal, triggering the contactor replacement operation instruction; When motor winding performance degradation is detected, a level 1 warning signal is generated, triggering instructions for motor load optimization and cooling system inspection. If the equivalent cumulative thermal stress coefficient continues to exceed the insulation aging threshold and the amplitude increase slope reaches the critical slope, it is upgraded to a level 3 warning signal, triggering instructions for motor winding insulation reinforcement or replacement. The multi-level early warning signals are pushed to the remote operation and maintenance platform through the industrial Internet of Things gateway, and the local sound and light alarm devices of the mixing station are activated simultaneously.

5. The fault monitoring method for electrical components of a mixing plant according to claim 1, characterized in that: The fault trend analysis includes: based on the fault characteristic parameters, using a pre-trained machine learning model to intelligently predict the component performance degradation trend and potential failure risks; wherein, the machine learning model is trained through historical fault characteristic parameter data, used to optimize fault identification accuracy, and output prediction results to undertake the generation of the multi-level warning signal.

6. A fault monitoring system for electrical components of a mixing plant, used to implement the fault monitoring method for electrical components of a mixing plant according to any one of claims 1 to 5, characterized in that: include: The signal acquisition module is used to configure corresponding data acquisition modules for the AC contactors and motor core electrical components in the mixing plant. The voltage signals of the electrical components are isolated and sensed through voltage transformers. The true RMS measurement chip is used to continuously collect the true RMS value and peak value of the voltage waveform in real time. The current signal is collected through the current sensor. The voltage and current signals are then transmitted to the signal processing end via the RS-485 bus interface according to the MODBUS-RTU protocol; A signal anti-interference processing module is used to receive voltage and current signals from the RS-485 bus interface, perform signal conditioning on the voltage and current signals, including amplification and baseline correction, eliminate offset errors introduced by the sensor, use an adaptive band-stop filter to suppress power frequency interference, and dynamically adjust the center frequency to the power grid frequency. A wavelet threshold denoising filter is used to suppress high-frequency noise pollution, and high-frequency noise components are removed through wavelet decomposition and reconstruction. A sliding average process is performed on the filtered signal to smooth transient interference and obtain a stable pre-processed signal. A fault characteristic parameter extraction module is used to divide the current waveform in the preprocessing signal into time windows during the contactor startup phase, identify the action timing of the step-down starting contactor and the running contactor, calculate the current conversion time difference between the step-down starting phase and the full-voltage operation phase, and detect the logical interlocking state of the contactor contact action signal in the two phases. If the conversion time difference exceeds a preset threshold or a logic interlock failure is detected, it is extracted as a conversion timeout characteristic parameter or a logic conflict characteristic parameter. At the same time, the effective value of the motor current in the preprocessing signal is monitored in real time, the instantaneous sudden increase event of the effective value of the current is captured, the sudden increase amplitude and duration are recorded, and the energy integral of the sudden increase event is calculated. If the sudden increase amplitude exceeds the set multiple of the rated current, the duration is greater than the minimum overload maintenance time, or the energy integral exceeds the safety threshold, it is extracted as a transient overload amplitude characteristic parameter, an overload duration characteristic parameter, or an overload energy characteristic parameter. A fault trend analysis module is used to periodically collect the conversion timeout characteristic parameters, logic conflict characteristic parameters, transient overload amplitude characteristic parameters, overload duration characteristic parameters, and overload energy characteristic parameters, build a time series characteristic database, count the triggering frequency of the conversion timeout characteristic parameters within a preset period, calculate the sliding change rate of the time difference mean of the triggering frequency, detect the continuous occurrence period of the logic conflict characteristic parameters, and associate the number of actions of the corresponding contactor. If the monthly change rate of the time difference mean exceeds the monthly change threshold or the weekly increase in the logic conflict frequency is greater than the weekly increase threshold, it is determined that the contactor contact is worn and degraded. The transient overload amplitude characteristic parameters are exponentially smoothed to extract the long-term rising slope of the overload event amplitude, aggregate the overload duration characteristic parameters and the overload energy characteristic parameters, and calculate the equivalent cumulative thermal stress coefficient. If the amplitude rising slope is higher than 2 times the standard deviation of the baseline value for three consecutive cycles, or the cumulative thermal stress coefficient exceeds the insulation aging threshold, it is determined that the motor winding performance is degraded. Based on the fault characteristic parameters, a pre-trained machine learning model is used to intelligently predict the component performance degradation trend and potential failure risk; The multi-level warning signal generation module is used to execute hierarchical warning decisions based on the contactor contact wear and degradation status or motor winding performance degradation status identified in the fault trend analysis. When the contactor contact wear and degradation status is identified, a first-level warning signal is generated to trigger the contactor contact lubrication maintenance instruction. If the contact wear and degradation status continues to deteriorate and the frequency of logical conflicts exceeds the safety threshold, it is upgraded to a second-level warning signal to trigger the contactor replacement operation instruction. When the motor winding performance degradation status is identified, a first-level warning signal is generated to trigger the motor load optimization and heat dissipation system inspection instruction. If the equivalent cumulative thermal stress coefficient continues to exceed the insulation aging threshold and the amplitude rising slope reaches the critical slope, it is upgraded to a third-level warning signal to trigger the motor winding insulation reinforcement or replacement operation instruction. The multi-level warning signal is pushed to the remote operation and maintenance platform through the industrial Internet of Things gateway, and the local sound and light alarm device of the mixing station is simultaneously activated.

Images