An electromagnetic fryer circuit fault detection system based on artificial intelligence
By constructing an electromagnetic fryer circuit fault detection system with a fully connected feedforward neural network and a global adaptive activation function, the problem of poor fault detection accuracy in the existing technology is solved, and timely detection of minor faults and suppression of false alarms are achieved, thereby improving the effect of electromagnetic fryer circuit fault detection.
Patent Information
- Application Number
- CN202510766646.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-10
- Publication Date
- 2025-09-19
- Estimated Expiration
- 2045-06-10
AI Technical Summary
The existing electromagnetic fryer circuit fault detection system has an increased rate of missed detection or false detection when the operating conditions or environment change. It is difficult to distinguish between ordinary noise and small fault signals, resulting in poor fault detection accuracy. It is also unable to capture slight fluctuations caused by slight changes in resistance and over-responds to sensor spike noise, resulting in poor fault detection effect.
A healthy operating condition current prediction model based on a fully connected feedforward neural network is constructed. A loss function is designed to target parameter drift and spike noise caused by small resistance changes and temperature drift. The loss sensitivity is adaptively adjusted through a global adaptive activation function to achieve timely detection of small faults and suppress false alarms.
The accuracy and flexibility of fault detection are improved, and it is possible to detect small changes in coil turn-to-turn resistance and medium-amplitude changes caused by temperature drift within the noise level, reducing false alarms and improving the ability to detect early-stage minor faults.
Smart Images
Figure CN120275811B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of circuit fault detection, and in particular to an electromagnetic fryer circuit fault detection system based on artificial intelligence. Background Art
[0002] Electromagnetic fryer circuit fault detection systems monitor parameters such as current, voltage, and temperature in real time, automatically diagnosing fault types and locations and providing timely warnings. However, typical electromagnetic fryer circuit fault detection systems suffer from inflexible fault judgment criteria, leading to increased rates of missed or false detections when operating conditions or the environment change. They also struggle to distinguish between ordinary noise and subtle fault signals, resulting in poor fault detection accuracy. Furthermore, these systems fail to detect slight fluctuations caused by minute changes in resistance and overreact to sensor spike noise, leading to poor fault detection. Summary of the Invention
[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides an electromagnetic fryer circuit fault detection system based on artificial intelligence. The fault judgment criteria of the general electromagnetic fryer circuit fault detection system are not flexible enough. Once the working conditions or environment change, the missed detection or false detection rate increases; it is difficult to distinguish between ordinary noise and small fault signals, which leads to poor fault detection accuracy. This solution constructs a healthy working condition current prediction model based on a fully connected feedforward neural network, designs loss functions for small resistance changes, parameter drift caused by temperature drift, and spike noise, and uses underlying variable mapping to generate parameters, adaptively adjusts the sensitivity of each loss; it can detect small coil turn-to-turn resistance within the noise level. Changes, it is sensitive to medium-amplitude changes caused by temperature drift, and can detect and suppress sensor spike noise, ensuring timely detection of early and minor faults while effectively avoiding false alarms. To address the problem that general electromagnetic fryer circuit fault detection systems are unable to capture slight fluctuations caused by slight resistance changes and over-respond to sensor spike noise, resulting in poor fault detection, this solution constructs a global adaptive activation function based on the loss functions designed for small resistance changes, parameter drift caused by temperature drift, and spike noise. The parameter adaptability is consistent with the sensitivity of the three types of losses, and the forward activation and backward loss are jointly amplified or suppressed at the same fault feature scale, thereby improving the fault detection effect.
[0004] The technical solution adopted by the present invention is as follows: the present invention provides an electromagnetic fryer circuit fault detection system based on artificial intelligence, including a data acquisition module, a pre-establishment module, a loss function design module, a joint optimization parameter module, a global adaptive activation function module and a fault detection module;
[0005] The data acquisition module acquires characteristic data of the electromagnetic fryer in a healthy state;
[0006] The pre-established module constructs a feedforward neural network model based on the standardized feature vector in the healthy state;
[0007] The loss function design module introduces three loss functions, one for small resistance change, one for medium fluctuation, and one for spike noise.
[0008] The joint optimization parameter module generates the parameters of the loss term through a learnable fractional exponential mapping, and automatically adjusts them together with the network parameters through back propagation during training;
[0009] The global adaptive activation function module calculates the global activation parameters based on the fractional exponential of the three losses and constructs an adaptive activation function;
[0010] The fault detection module performs fault detection on new input data.
[0011] Furthermore, the data acquisition module selects the key nodes at the power supply side and both ends of the inductor coil from the electromagnetic fryer to collect the voltage and current time series signals in a healthy state; performs short-time Fourier transform on the time series to extract the frequency domain and time domain feature vectors; and performs short-time Fourier transform on the feature vectors of all dimensions. Perform zero mean and unit variance standardization.
[0012] Furthermore, the pre-established module is to establish a model in a healthy state, with the standardized feature vector As input, a healthy operating current prediction model is constructed based on a fully connected feedforward neural network, which is expressed as: ; residual Expressed as: ;in, is the predicted value of the current in the kth frame; is the actual value; F(·) is the fully connected feedforward neural network mapping, are network parameters.
[0013] Furthermore, the loss function design module constructs a fractional-order loss function; the short memory characteristics of the fractional-order loss are used to characterize the residual, and the loss driven by the residual is divided into three items, corresponding to different fault modes; for the resistance change of the tiny coil, the residual sensitive detection loss is defined , expressed as: ; ; Among them, x is the input of the loss function, and the residual is taken ; c is the scaling constant; is the fractional order index; is the gamma function; is the gradient; is a sign function; for the electrical parameter drift caused by temperature, it defines the medium amplitude fluctuation loss , expressed as: ; ; where h is the step constant; M is the series truncation upper limit; n is the memory component series index; tanh(·) is the hyperbolic tangent function; for occasional short-circuit spikes in the sensor, the robust loss is defined as , expressed as: ; .
[0014] Furthermore, the joint optimization parameter module takes the three losses Independent of each other, let the fractional order index of each loss term By an underlying variable Generate, expressed as: ;in, and are the fractional exponent and underlying variable of the i-th loss respectively; and They are the upper and lower limits of the index respectively; indirect and direct updates are used , updates are updated simultaneously through back propagation and ;pass Mapping Update , defining the overall loss of the healthy current prediction model , expressed as: ; The joint optimization is expressed as: ; Where N is the total number of training samples; is the sample of the kth frame; and It is the final setting of the initialization parameters.
[0015] Furthermore, the global adaptive activation function module is to convert the three losses into Use 、 and Indicates that global activation parameters are defined , expressed as: ; And define the global adaptive activation function of the healthy working condition current prediction model, expressed as: ;in, is the global adaptive activation function; u is the activation function input.
[0016] Furthermore, the fault detection module is to detect the new input after the training is completed. Calculate prediction residuals , and instantly evaluate the total score index , expressed as: , the residual distribution under healthy conditions is used to determine the alarm threshold, which is expressed as: ;when , it is determined that a fault has occurred; among them, and are the mean and standard deviation under healthy working conditions; is the safety factor.
[0017] The beneficial effects achieved by the present invention using the above scheme are as follows:
[0018] (1) In view of the fact that the fault judgment criteria of the general electromagnetic fryer circuit fault detection system are not flexible enough, once the working conditions or environment change, the missed detection or false detection rate increases; it is difficult to distinguish between ordinary noise and small fault signals, which leads to poor fault detection accuracy. This scheme constructs a healthy working condition current prediction model based on a fully connected feedforward neural network, designs loss functions for small resistance changes, parameter drift caused by temperature drift, and spike noise, and uses the underlying variable mapping to generate parameters, and adaptively adjusts the sensitivity of each loss; it can detect small coil turn resistance changes within the noise level, is sensitive to medium amplitude changes caused by temperature drift, and realizes the detection and suppression of sensor spike noise, which not only ensures the timely detection of early and small faults, but also effectively avoids false alarms.
[0019] (2) Aiming at the problem that the fault detection system of general electromagnetic fryer circuit cannot capture slight fluctuations caused by slight changes in resistance and over-responds to sensor spike noise, which leads to poor fault detection effect, this scheme constructs a global adaptive activation function based on the loss function designed for small resistance changes, parameter drift caused by temperature drift and spike noise; the parameter adaptability is consistent with the sensitivity of the three types of losses, and the forward activation and backward loss are linked to amplify or suppress at the same fault feature scale, thereby improving the fault detection effect. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 This is a flow chart of an artificial intelligence-based electromagnetic fryer circuit fault detection system provided by the present invention.
[0021] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION
[0022] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0023] In the description of the present invention, it should be understood that terms such as "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, they should not be understood as limiting the present invention.
[0024] Example 1, see Figure 1 , the present invention provides an electromagnetic fryer circuit fault detection system based on artificial intelligence, including a data acquisition module, a pre-establishment module, a loss function design module, a joint optimization parameter module, a global adaptive activation function module and a fault detection module;
[0025] The data acquisition module acquires characteristic data of the electromagnetic fryer in a healthy state; and sends the data to the pre-establishment module;
[0026] The pre-established module constructs a feedforward neural network model based on the standardized feature vector in the healthy state; and sends the data to the loss function design module;
[0027] The loss function design module introduces three loss functions, one for small resistance change, one for medium fluctuation, and one for spike noise; and sends the data to the joint optimization parameter module;
[0028] The joint optimization parameter module generates the parameters of the loss term through a learnable fractional exponential mapping, and automatically adjusts them together with the network parameters through back propagation during training; and sends the data to the global adaptive activation function module;
[0029] The global adaptive activation function module calculates the global activation parameters based on the fractional exponents of the three losses and constructs an adaptive activation function; and sends the data to the fault detection module;
[0030] The fault detection module performs fault detection on new input data.
[0031] Example 2, see Figure 1 This embodiment is based on the above embodiment. The data acquisition module selects the key nodes at the power supply side and both ends of the inductor from the electromagnetic fryer to collect the voltage and current time series signals in a healthy state; performs short-time Fourier transform on the time series to extract the frequency domain and time domain feature vectors; the time domain statistical features include mean, variance, skewness, kurtosis, impulse factor and form factor; the frequency domain statistical features include spectrum centroid, spectrum bandwidth, spectrum entropy, spectrum flatness and spectrum flux; the feature vectors of all dimensions are calculated. Perform zero mean and unit variance standardization to eliminate dimension and magnitude differences.
[0032] Example 3, see Figure 1 This embodiment is based on the above embodiment. The pre-established module is to establish a model in a healthy state to compare with the actual signal and extract the fault characteristics; the standardized feature vector As input, a healthy operating current prediction model is constructed based on a fully connected feedforward neural network, which is expressed as: ; residual Expressed as: ;in, is the predicted value of the current in the kth frame; is the actual value; F(·) is the fully connected feedforward neural network mapping, is a network parameter; under normal operating conditions, the residual should be close to zero; faults such as coil short circuit, open circuit, and inter-turn short circuit will produce significant outlier residuals.
[0033] Example 4, see Figure 1 This embodiment is based on the above embodiment. The loss function design module is to construct a fractional-order loss function. The short memory characteristic of the fractional-order loss is used to more finely characterize the residual, which neither over-amplifies ordinary noise nor suppresses extreme transient spikes. The loss driven by the residual is divided into three items, corresponding to different fault modes. For the resistance change of the tiny coil, the residual sensitive detection loss is defined. , expressed as: ; ; Among them, x is the input of the loss function, and the residual is taken ; c is the scaling constant; is the fractional order index; is the gamma function; is the gradient; is a sign function; for the electrical parameter drift caused by temperature, it defines the medium amplitude fluctuation loss , expressed as: ; ; where h is the step constant; M is the series truncation upper limit; n is the memory component series index; tanh(·) is the hyperbolic tangent function; for occasional short-circuit spikes of the sensor, the robust loss is defined as , expressed as: ; .
[0034] Example 5, see Figure 1 This embodiment is based on the above embodiment, and the joint optimization parameter module takes the three losses Independent of each other, to avoid offline grid search, let the fractional order index of each loss term By an underlying variable Generate, expressed as: ;in, and are the fractional exponent and underlying variable of the i-th loss respectively; and They are the upper and lower limits of the index respectively; indirect and direct updates are used , updates are updated simultaneously through back propagation and ;pass Mapping Update , ensure the correctness of constrained optimization; define the overall loss of the healthy operating current prediction model , expressed as: ; The joint optimization is expressed as: ; Where N is the total number of training samples; is the sample of the kth frame; and It is the final setting of the initialization parameters.
[0035] By performing the above operations, we can address the problems of general electromagnetic fryer circuit fault detection systems, such as insufficient flexibility in fault judgment criteria. Once the operating conditions or environment change, the missed detection or false detection rate increases; it is difficult to distinguish between ordinary noise and minor fault signals, which leads to poor fault detection accuracy. This solution constructs a healthy operating condition current prediction model based on a fully connected feedforward neural network, designs loss functions for minor resistance changes, parameter drift caused by temperature drift, and spike noise, and uses underlying variable mapping to generate parameters, adaptively adjusting the sensitivity of each loss. It can detect minor changes in coil turn resistance within the noise level, is sensitive to medium-amplitude changes caused by temperature drift, and detect and suppress sensor spike noise, ensuring timely detection of early and minor faults while effectively avoiding false alarms.
[0036] Example 6, see Figure 1 This embodiment is based on the above embodiment. The global adaptive activation function module is to enable the network to have adaptive expression capabilities between different fault feature scales. Use 、 and Indicates that global activation parameters are defined , expressed as: ; And define the global adaptive activation function of the healthy working condition current prediction model, expressed as: ;in, is the global adaptive activation function; u is the activation function input; when When it gets bigger, Also increases, the inflection point of the global adaptive activation function becomes steeper, and the network is better at learning high-frequency spike features; on the contrary, when When it becomes smaller, The network is more inclined to fit large-scale steady changes, which enables the network to adaptively adjust its nonlinear expression ability to different feature scales when detecting small resistance changes, temperature drifts and spike short circuits.
[0037] By performing the above operations, we address the problem that general electromagnetic fryer circuit fault detection systems cannot capture slight fluctuations caused by small changes in resistance, over-respond to sensor spike noise, and thus have poor fault detection results. This solution constructs a global adaptive activation function based on loss functions designed for small resistance changes, parameter drift caused by temperature drift, and spike noise. The parameter adaptability is consistent with the sensitivity of the three types of losses. Forward activation and backward loss are jointly amplified or suppressed at the same fault feature scale, thereby improving fault detection effectiveness.
[0038] Example 7, see Figure 1 This embodiment is based on the above embodiment. The fault detection module is to detect the new input after the training is completed. Calculate prediction residuals , and instantly evaluate the total score index , expressed as: , the residual distribution under healthy conditions is used to determine the alarm threshold, which is expressed as: ;when , it is determined that a fault has occurred; among them, and are the mean and standard deviation under healthy working conditions; is the safety factor.
[0039] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.
[0040] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.
[0041] The present invention and its embodiments are described above. This description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if a person skilled in the art is inspired by this and, without departing from the purpose of the present invention, designs structures and embodiments similar to this technical solution without inventiveness, they shall fall within the scope of protection of the present invention.
Claims
1. An electromagnetic fryer circuit fault detection system based on artificial intelligence, characterized by: The system includes a data acquisition module, a pre-establishment module, a loss function design module, a joint optimization parameter module, a global adaptive activation function module and a fault detection module; The data acquisition module acquires characteristic data of the electromagnetic fryer in a healthy state; The pre-established module constructs a feedforward neural network model based on the standardized feature vector in the healthy state; The loss function design module introduces three loss functions, one for small resistance change, one for medium fluctuation, and one for spike noise. The joint optimization parameter module generates the parameters of the loss term through a learnable fractional exponential mapping, and automatically adjusts them together with the network parameters through back propagation during training; The global adaptive activation function module calculates the global activation parameters based on the fractional exponential of the three losses and constructs an adaptive activation function; The fault detection module performs fault detection on new input data; The loss function design module constructs a fractional-order loss function; uses the short memory characteristics of the fractional-order loss to characterize the residual, and divides the residual-driven loss into three items, corresponding to different fault modes; defines the residual sensitive detection loss for the resistance change of the tiny coil , expressed as: ; ; Among them, x is the input of the loss function, and the residual is taken ; c is the scaling constant; is the fractional order index; is the gamma function; is the gradient; is a sign function; for the electrical parameter drift caused by temperature, it defines the medium amplitude fluctuation loss , expressed as: ; ; where h is the step constant; M is the series truncation upper limit; n is the memory component series index; tanh(·) is the hyperbolic tangent function; for occasional short-circuit spikes in the sensor, the robust loss is defined as , expressed as: ; .
2. The artificial intelligence-based electromagnetic fryer circuit fault detection system according to claim 1, characterized in that: The joint optimization parameter module takes the three losses Independent of each other, let the fractional order index of each loss term By an underlying variable Generate, expressed as: ;in, and are the fractional exponent and underlying variable of the i-th loss respectively; and They are the upper and lower limits of the index respectively; indirect and direct updates are used , updates are updated simultaneously through back propagation and ;pass Mapping Update , defining the overall loss of the healthy current prediction model , expressed as: ; The joint optimization is expressed as: ; Where N is the total number of training samples; is the sample of the kth frame; and It is the final setting of the initialization parameters.
3. The artificial intelligence-based electromagnetic fryer circuit fault detection system according to claim 2, characterized in that: The global adaptive activation function module is to convert the three losses into Use 、 and Indicates that global activation parameters are defined , expressed as: ; And define the global adaptive activation function of the healthy working condition current prediction model, expressed as: ;in, is the global adaptive activation function; u is the activation function input.
4. The artificial intelligence-based electromagnetic fryer circuit fault detection system according to claim 3, characterized in that: The data acquisition module selects the key nodes at the power supply side and both ends of the inductor coil from the electromagnetic fryer to collect the voltage and current time series signals in the healthy state; performs short-time Fourier transform on the time series to extract the frequency domain and time domain feature vectors; and performs short-time Fourier transform on the feature vectors of all dimensions. Perform zero mean and unit variance standardization.
5. The artificial intelligence-based electromagnetic fryer circuit fault detection system according to claim 4, characterized in that: The pre-built module is to build a model in a healthy state with the normalized feature vector As input, a healthy operating current prediction model is constructed based on a fully connected feedforward neural network, which is expressed as: ; residual Expressed as: ;in, is the predicted value of the current in the kth frame; is the actual value; F(·) is the fully connected feedforward neural network mapping, are network parameters.
6. The artificial intelligence-based electromagnetic fryer circuit fault detection system according to claim 5, characterized in that: The fault detection module is to detect new input Calculate prediction residuals , and instantly evaluate the total score index , expressed as: , the residual distribution under healthy conditions is used to determine the alarm threshold, which is expressed as: ;when , it is determined that a fault has occurred; among them, and are the mean and standard deviation under healthy working conditions; is the safety factor.
Citation Information
Patent Citations
Intelligent early warning car washer
CN119851449A