Range extender fault detection model construction method, fault detection and diagnosis method and maintenance method

Abstract

The present invention discloses a range extender fault detection and diagnosis method. According to the characteristics of multi-channel original signals collected during fault detection, multiple nonlinear processing modules are set in series in the range extender fault detection model to deeply explore potential range extender fault characteristics, so that the model output results are constantly approaching the real results, and finally the detection and diagnosis of range extender faults are realized; after the range extender fault detection model is embedded in the ECU of the automobile, during driving, multi-channel original signals can be collected in real time, and the range extender fault detection model in the ECU can be used to perform real-time detection on the current working state of the range extender. According to the working state expression result of the range extender output by the range extender fault detection model, it can be concluded whether the range extender is in a normal working state or a faulty working state. The present invention also discloses a range extender fault detection and diagnosis method and a range extender maintenance method.

Description

Technical Field

[0001] The present invention belongs to the technical field of fault detection and diagnosis, and specifically relates to a range extender fault detection model construction method, a fault detection and diagnosis method, and a maintenance method. Background Art

[0002] In order to cope with the challenges of environmental pollution and energy shortage, the transformation of automobiles to clean energy is being actively promoted. New energy vehicles have become one of the important means to achieve this goal. As a type of new energy vehicle, electric vehicles have many advantages, such as not needing to use traditional petroleum fuels and not producing harmful exhaust emissions. However, the range of pure electric vehicles is restricted by factors such as low battery energy density, large low-temperature energy loss, short battery life, and long charging time, which makes many people question the practicality and reliability of electric vehicles. In order to overcome the above problems, extended-range hybrid electric vehicles came into being. This type of vehicle uses a range extender and a power battery pack as energy and is driven by an electric motor. Compared with pure electric vehicles, extended-range hybrid electric vehicles have the advantages of long range, fast energy replenishment, and good power smoothness.

[0003] The range extender is the core energy supply component of the extended-range hybrid electric vehicle. It consists of two parts: the engine and the generator. The engine is coupled with the generator to drive the generator to rotate and generate electricity to continue to supply energy to the car. When the power battery is insufficient, the range extender can provide additional power to the battery pack, thereby extending the range of the electric vehicle. Therefore, the development of range extender technology will promote the popularization and promotion of extended-range hybrid electric vehicles and provide more possibilities for the development of clean energy vehicles.

[0004] However, as a complex power system composed of an engine and a generator, the complex mechanical, electrical and liquid coupling relationships result in that once a range extender fails, it is usually a coupling failure of different components. When a coupling failure occurs, the signal collected by the sensor contains the characteristics of a single fault and the fault characteristics after coupling, which brings great difficulties to the fault detection and diagnosis of the range extender. Traditional statistical analysis methods, pattern analysis methods, and modeling analysis methods are difficult to detect and diagnose range extender faults efficiently and accurately, resulting in the system being unable to timely reflect the true working status of the range extender during driving, posing potential safety hazards. In addition, the fault type and location are also difficult to determine, which brings additional costs to subsequent maintenance work. Therefore, it is extremely necessary to develop a range extender fault detection and diagnosis method that can overcome the impact of coupling failures and accurately identify the fault type and location.

[0005] With the development of sensor technology and computer technology, more and more parameters can be obtained during vehicle driving, and the ability of computers to process these parameters is also constantly improving. This provides the possibility for data-driven range extender fault detection and diagnosis. Some emerging technologies such as machine learning and deep learning are being widely used in the field of fault detection and diagnosis. These technologies can efficiently analyze and process large amounts of data, and can automatically extract features, so that faults can be diagnosed more accurately. Compared with traditional methods, these emerging technologies can handle more complex fault modes and can automatically learn new fault modes, making fault detection and diagnosis more intelligent and efficient. It has been proved in practice that these emerging technologies have achieved good results in the field of industrial fault detection and diagnosis, and are expected to further promote the development of range extender fault detection and diagnosis technology. Summary of the invention

[0006] In view of this, the object of the present invention is to provide a range extender fault detection model construction method, a fault detection and diagnosis method and a maintenance method, which can perform real-time fault detection and diagnosis of the range extender, ensure the safety and stability of the vehicle during driving and reduce subsequent maintenance costs.

[0007] In order to achieve the above object, the present invention provides the following technical solutions:

[0008] The present invention first proposes a method for constructing a range extender fault detection model, comprising the following steps:

[0009] S1: Simulate the working conditions of the range extender, including normal working conditions and various fault working conditions, which may occur during driving. Use sensors arranged at key positions of the range extender to collect and obtain multi-channel original signals of the range extender under various working conditions, and use the multi-channel original signals to make a range extender working condition detection data set;

[0010] S2: Construct a multi-layer neural network as the information encoder of the multi-channel signal to encode the multi-channel original signal into a latent space feature matrix;

[0011] S3: construct multiple nonlinear processing modules connected in series and obtain a nonlinear processing module group, each of which includes a convolution layer, a pooling layer and a nonlinear activation function layer; input the latent space feature matrix into the nonlinear processing module group, and use the nonlinear processing module group to deeply mine the range extender operating condition features hidden in the multi-channel original signal and integrate them into a high-dimensional semantic feature matrix;

[0012] S4: Construct a multi-layer neural network as an information decoder of the high-dimensional semantic feature matrix, aggregate and reduce the dimension of the high-dimensional semantic feature matrix layer by layer, and finally output the expression result of the range extender working condition;

[0013] S5: comparing the expression result of the range extender operating condition with the actual result recorded in the range extender operating condition detection data set and calculating the cross entropy;

[0014] S6: Use the cross entropy as the loss function, use the gradient descent optimizer to iteratively optimize the loss function, and update the parameters of each nonlinear processing module;

[0015] S7: Loop through steps S2 to S6, use the range extender operating condition detection data set to train the model, iteratively perform forward operations and optimize model parameters, so that the model gradually learns the potential correspondence between multi-channel original signals and various range extender operating conditions, and obtains the range extender fault detection model after training is completed.

[0016] Further, in step S3, in the nonlinear processing module group, at least one nonlinear processing module is a denoising module, at least one nonlinear processing module is a time processing module, and at least one nonlinear processing module is a channel processing module; the denoising module adopts a soft threshold denoising strategy, uses the set multi-layer depooling convolution layer to act as a filter, generates a reasonable soft threshold to perform denoising operations on the input multi-channel original signal; the time processing module constructs a multi-layer convolution layer with different kernel sizes and a global pooling layer according to the timing characteristics of the signal to perform deep feature mining on long and short range time information; the channel processing module performs a mixed pooling operation after convolution to fully extract the correlation characteristics between multi-channel signals.

[0017] Furthermore, in step S3, the convolutional layer is used to extract the latent space feature matrix X i Potential information:

[0018]

[0019] in, represents the feature matrix extracted by the convolution layer; k represents the size of the convolution kernel, W represents the convolution kernel weight vector, and p represents the latent space feature matrix X i The pth element in , s represents the feature matrix obtained by convolution The sth element in ;

[0020] The pooling layer is used to aggregate the high-dimensional features extracted by the convolutional layer:

[0021]

[0022] in, represents the feature matrix extracted by the pooling layer; k′ is the size of the pooling kernel, and pool(·) represents the pooling operation;

[0023] The nonlinear activation function σ(·) is used to change the linear relationship of feature data to make the model learn nonlinearly.

[0024] Furthermore, the latent space feature matrix X i In the nonlinear processing module, the convolution layer, pooling layer and nonlinear activation function layer are sequentially passed to output a high-dimensional feature matrix.

[0025]

[0026] in, A high-dimensional feature matrix representing the output of the nonlinear processing module.

[0027] Furthermore, all the nonlinear processing modules are connected in series, and the output of the first nonlinear processing module As the input of the second nonlinear processing module, after being processed by the second nonlinear processing module, a higher-dimensional feature matrix is ​​output ... and so on, the n nonlinear processing modules connected in series gradually extract the latent space feature matrix X i The high-dimensional features hidden in the final result are the high-dimensional semantic feature matrix

[0028] Furthermore, in step S5, the cross entropy is calculated as follows:

[0029]

[0030] Where m represents the sample size; u(·) represents the true distribution of the sample; v(·) represents the expression distribution of the model; and the cross entropy H(u,v) is used to measure the difference between u and v, thereby indicating the difference between the model diagnosis result and the true result.

[0031] Further, in step S6, the loss function is:

[0032]

[0033] Among them, N represents the sample size; L represents the number of labels; y i,l Indicates that the true result of the i-th sample is l; It represents the probability that the diagnosis result of the i-th sample is l.

[0034] The present invention also proposes a range extender fault detection and diagnosis method, comprising the following steps:

[0035] Step 1: A range extender fault detection model is constructed by the method described above, and the range extender fault detection model is arranged in the ECU of the vehicle, and sensors at key positions of the range extender are used to collect and obtain multi-channel original signals of the range extender under various working conditions;

[0036] Step 2: Input the multi-channel original signal collected in real time into the ECU, use the range extender fault detection model to perform real-time detection, and output the working condition expression result of the range extender;

[0037] Step 3: If the result of the working condition expression is the label corresponding to the normal working condition of the range extender, it indicates that the range extender is in a normal working condition, and step 4 is performed; if the result of the working condition expression is the label corresponding to the fault working condition of the range extender, it indicates that the range extender is in a fault working condition, and step 5 is performed;

[0038] Step 4: Execute step 2;

[0039] Step 5: The ECU controls the instrument panel to light up the range extender fault indicator light, indicating a range extender fault, and records the range extender operating condition expression results.

[0040] The present invention also proposes a range extender maintenance method, which uses the range extender fault detection and diagnosis method as described above to detect the operating status of the range extender in real time. When it is detected that the range extender is in a faulty operating condition, the operating condition expression result of the range extender is recorded and the fault type is determined to maintain the range extender.

[0041] The beneficial effects of the present invention are:

[0042] The range extender fault detection and diagnosis method of the present invention, by constructing a range extender fault detection model, according to the characteristics of the multi-channel original signal collected during fault detection, multiple nonlinear processing modules are set in series in the range extender fault detection model to deeply explore the potential range extender fault characteristics, so that the model output results are constantly approaching the real results, and finally realize high-speed and high-precision detection and diagnosis of range extender faults; in this way, after the range extender fault detection model is embedded in the ECU of the automobile, during driving, the multi-channel original signal can be collected in real time, and the range extender fault detection model in the ECU is used to perform real-time detection of the current working state of the range extender, and according to the working state expression result of the range extender output by the range extender fault detection model, it can be concluded whether the range extender is in a normal working state or a faulty working state, that is, the range extender fault detection and diagnosis method of the present invention can perform real-time fault detection and diagnosis of the range extender in real time, ensure the safety and stability of the vehicle during driving and reduce subsequent maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to make the purpose, technical solution and beneficial effects of the present invention clearer, the present invention provides the following drawings for illustration:

[0044] Figure 1 A schematic diagram of the structure for arranging sensors at key positions of the range extender;

[0045] Figure 2 This is a principle framework diagram of the range extender fault detection model of this embodiment;

[0046] Figure 3 This is the schematic diagram of the nonlinear processing module;

[0047] Figure 4 It is an information transfer diagram of the serially connected nonlinear processing modules;

[0048] Figure 5 Schematic diagram of automobile range extender fault detection and diagnosis using the range extender fault detection model. DETAILED DESCRIPTION

[0049] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0050] The range extender fault detection and diagnosis method of this embodiment includes the following steps:

[0051] Step 1: Build a range extender fault detection model, place the range extender fault detection model in the vehicle's ECU, and use sensors at key locations on the range extender to collect and obtain multi-channel original signals of the range extender under various working conditions, such as Figure 1 As shown. Figure 2 As shown, it is a principle framework diagram of the range extender fault detection model of this embodiment. In this embodiment, the method for constructing the range extender fault detection model includes the following steps:

[0052] S1: Simulate the operating conditions of the range extender that may occur during driving, including normal operating conditions and various fault conditions. Use sensors arranged at key positions of the range extender to collect and obtain multi-channel original signals of the range extender under various operating conditions, and use the multi-channel original signals to make a range extender operating condition detection data set.

[0053] Specifically, the arrangement positions and types of sensors include but are not limited to: arranging a speed sensor on the crankshaft of the engine; arranging a speed sensor and a torque sensor on the output shaft; arranging a displacement sensor on the main shaft of the generator; arranging a temperature sensor on the inner cavity shell; arranging a voltage sensor and a current sensor at the output end.

[0054] Simulate the normal working condition of the range extender and possible n types of fault conditions during driving. The sensor collects signals from the range extender under various working conditions. The collected signals are converted into digital signals by A / D converter to obtain multi-channel original signals. The multi-channel original signals are uploaded to the industrial computer for storage and marked as The superscript c represents the signal collected by the c-th sensor (c-th channel), and the subscript i represents the i-th sample.

[0055] Assign labels to various working conditions of the range extender. Specifically, in this embodiment, for the signal samples of the simulated normal working condition, the real label is marked as 0; for the signal samples of the simulated fault type 1, the real label is marked as 1; for the signal samples of the simulated fault type 2, the real label is marked as 2... and so on, until all types of simulated signal samples are completely labeled. At this time, there should be n+1 labels, corresponding to 1 signal sample of the normal working condition of the range extender and n signal samples of the fault working condition of the range extender. The real labels of all samples are collectively referred to as y i , subscript i represents the i-th sample.

[0056] S2: Construct a multi-layer neural network as the information encoder of the multi-channel signal, and encode the multi-channel original signal into the latent space feature matrix X i ,and Wherein, Encoder represents an information encoder.

[0057] S3: construct multiple nonlinear processing modules connected in series and obtain a nonlinear processing module group, each of which includes a convolution layer, a pooling layer and a nonlinear activation function layer; input the latent space feature matrix into the nonlinear processing module group, and use the nonlinear processing module group to deeply mine the range extender operating condition characteristics implicit in the multi-channel original signal and integrate them into a high-dimensional semantic feature matrix.

[0058] Specifically, in view of the characteristics of the range extender fault signal, such as high noise, timing, and strong correlation between channels, in the nonlinear processing module group, at least one nonlinear processing module is a denoising module, at least one nonlinear processing module is a time processing module, and at least one nonlinear processing module is a channel processing module. The denoising module adopts a soft threshold denoising strategy, and uses the set multi-layer depooling convolution layer as a filter to generate a reasonable soft threshold to perform denoising operations on the input multi-channel original signal. The time processing module constructs multiple layers of convolution layers with different kernel sizes and a global pooling layer according to the timing characteristics of the signal to perform deep feature mining on long and short range time information. The channel processing module performs a mixed pooling operation after convolution to fully extract the correlation characteristics between multi-channel signals. Figure 3 As shown, although the functions of various nonlinear processing modules may be different, each of the nonlinear processing modules includes a convolution layer, a pooling layer and a nonlinear activation function layer.

[0059] The convolutional layer is used to extract the latent space feature matrix X i Potential information:

[0060]

[0061] in, represents the feature matrix extracted by the convolution layer; k represents the size of the convolution kernel, W represents the convolution kernel weight vector, and p represents the latent space feature matrix Xi The pth element in , s represents the feature matrix obtained by convolution The sth element in ;

[0062] The pooling layer is used to aggregate the high-dimensional features extracted by the convolutional layer:

[0063]

[0064] in, represents the feature matrix extracted by the pooling layer; k′ is the size of the pooling kernel, and pool(·) represents the pooling operation;

[0065] The nonlinear activation function σ(·) is used to change the linear relationship of feature data to make the model learn nonlinearly.

[0066] Latent space feature matrix X i In the nonlinear processing module, the convolution layer, pooling layer and nonlinear activation function layer are sequentially passed to output a high-dimensional feature matrix.

[0067]

[0068] in, A high-dimensional feature matrix representing the output of the nonlinear processing module.

[0069] like Figure 4 As shown, since all nonlinear processing modules are connected in series, the output of the first nonlinear processing module As the input of the second nonlinear processing module, after being processed by the second nonlinear processing module, a higher-dimensional feature matrix is ​​output ... and so on, the n nonlinear processing modules connected in series gradually extract the latent space feature matrix X i The high-dimensional features hidden in the final result are the high-dimensional semantic feature matrix

[0070] S4: Construct a multi-layer neural network as the information decoder of the high-dimensional semantic feature matrix, aggregate and reduce the dimension of the high-dimensional semantic feature matrix layer by layer, and finally output the expression result of the range extender working condition Right now Among them, Decoder represents the information decoder. That is, the range extender operating condition recognition result of the range extender fault detection model for signal sample i.

[0071] S5: Compare the expression result of the range extender working condition with the actual result recorded in the range extender working condition detection data set and calculate the cross entropy. The calculation method of the cross entropy is:

[0072]

[0073] Where m represents the sample size; u(·) represents the true distribution of the sample; v(·) represents the expression distribution of the model; and the cross entropy H(u,v) is used to measure the difference between u and v, thereby indicating the difference between the model diagnosis result and the true result.

[0074] S6: Use one-hot encoding to achieve multi-classification expression, use cross entropy as the loss function, use the gradient descent optimizer to iteratively optimize the loss function, and update the parameters of each nonlinear processing module. Specifically, the loss function is:

[0075]

[0076] Among them, N represents the sample size; L represents the number of labels; y i,l Indicates that the true result of the i-th sample is l; It represents the probability that the diagnosis result of the i-th sample is l.

[0077] The Adam optimizer is used to iteratively optimize the loss function Loss, continuously fitting the expression distribution and true distribution of the range extender fault detection model, so that the diagnosis results of the model continue to approach the true results.

[0078] S7: Loop through steps S2 to S6, use the range extender operating condition detection data set to train the model, iteratively perform forward operations and optimize model parameters, so that the model gradually learns the potential correspondence between multi-channel original signals and various range extender operating conditions, and obtains the range extender fault detection model after training is completed.

[0079] Step 2: Input the multi-channel original signal collected in real time into the ECU, use the range extender fault detection model to perform real-time detection, and output the range extender working condition expression result, such as Figure 5 shown.

[0080] Step 3: If the working condition expression result is the label corresponding to the normal working condition of the range extender, that is, the diagnosis result of the range extender fault detection model It indicates that the range extender is in normal working condition, and step 4 is executed; if the working condition expression result is the label corresponding to the range extender fault working condition, that is, the diagnosis result of the range extender fault detection model This indicates that the range extender is in a faulty condition. Go to step 5.

[0081] Step 4: Execute step 2 and continue to monitor the operating status of the range extender in real time.

[0082] Step 5: The ECU controls the instrument panel to light up the range extender fault indicator light, indicating a range extender fault, and records the range extender operating condition expression results.

[0083] This embodiment also proposes a range extender maintenance method, which uses the range extender fault detection and diagnosis method described in this embodiment to detect the range extender's operating state in real time. When the range extender is detected to be in a faulty operating state, the range extender's operating state expression result is recorded and the fault type is determined to maintain the range extender. Specifically, if Indicates that the range extender has a type 1 fault condition; if Indicates that the range extender has a type 2 fault condition... and so on. Indicates that the range extender has a fault condition of type n. Perform repair and maintenance on the range extender according to the fault type.

[0084] The above-described embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or changes made by those skilled in the art based on the present invention are within the protection scope of the present invention. The protection scope of the present invention shall be subject to the claims.

Claims

1. A method for constructing a range extender fault detection model, characterized in that: The steps include: S1: Simulate the working conditions of the range extender, including normal working conditions and various fault working conditions, which may occur during driving. Use sensors arranged at key positions of the range extender to collect and obtain multi-channel original signals of the range extender under various working conditions, and use the multi-channel original signals to make a range extender working condition detection data set; S2: Construct a multi-layer neural network as the information encoder of the multi-channel signal to encode the multi-channel original signal into a latent space feature matrix; S3: construct multiple nonlinear processing modules connected in series and obtain a nonlinear processing module group, each of which includes a convolution layer, a pooling layer and a nonlinear activation function layer; input the latent space feature matrix into the nonlinear processing module group, and use the nonlinear processing module group to deeply mine the range extender operating condition features hidden in the multi-channel original signal and integrate them into a high-dimensional semantic feature matrix; In the nonlinear processing module group, at least one nonlinear processing module is a denoising module, at least one nonlinear processing module is a time processing module, and at least one nonlinear processing module is a channel processing module; The denoising module adopts a soft threshold denoising strategy, uses the multi-layer depooling convolution layer as a filter, generates a reasonable soft threshold to perform denoising operations on the input multi-channel original signal; the time processing module constructs a multi-layer convolution layer with different kernel sizes and a global pooling layer according to the timing characteristics of the signal to perform deep feature mining on the long and short range time information; the channel processing module performs a mixed pooling operation after convolution to fully extract the correlation characteristics between multi-channel signals; S4: Construct a multi-layer neural network as an information decoder of the high-dimensional semantic feature matrix, aggregate and reduce the dimension of the high-dimensional semantic feature matrix layer by layer, and finally output the expression result of the range extender working condition; S5: comparing the expression result of the range extender operating condition with the actual result recorded in the range extender operating condition detection data set and calculating the cross entropy; S6: Use the cross entropy as the loss function, use the gradient descent optimizer to iteratively optimize the loss function, and update the parameters of each nonlinear processing module; S7: Loop through steps S2 to S6, use the range extender operating condition detection data set to train the model, iteratively perform forward operations and optimize model parameters, so that the model gradually learns the potential correspondence between multi-channel original signals and various range extender operating conditions, and obtains the range extender fault detection model after training is completed.

2. The range extender fault detection model construction method according to claim 1, characterized in that: In step S3, the convolutional layer is used to extract the latent space feature matrix X i Potential information: in, represents the feature matrix extracted by the convolution layer; k represents the size of the convolution kernel, W represents the convolution kernel weight vector, and p represents the latent space feature matrix X i The pth element in , s represents the feature matrix obtained by convolution The sth element in ; The pooling layer is used to aggregate the high-dimensional features extracted by the convolutional layer: in, represents the feature matrix extracted by the pooling layer; k′ is the size of the pooling kernel, and pool(·) represents the pooling operation; The nonlinear activation function σ(·) is used to change the linear relationship of feature data to make the model learn nonlinearly.

3. The range extender fault detection model construction method according to claim 2, characterized in that: Latent space feature matrix X i In the nonlinear processing module, the convolution layer, pooling layer and nonlinear activation function layer are sequentially passed to output a high-dimensional feature matrix. in, A high-dimensional feature matrix representing the output of the nonlinear processing module.

4. The range extender fault detection model construction method according to claim 3, characterized in that: All the nonlinear processing modules are connected in series, and the output of the first nonlinear processing module As the input of the second nonlinear processing module, after being processed by the second nonlinear processing module, a higher-dimensional feature matrix is ​​output By analogy, the n nonlinear processing modules connected in series gradually extract the latent space feature matrix X i The high-dimensional features hidden in the final result are the high-dimensional semantic feature matrix 5. The range extender fault detection model construction method according to claim 1, characterized in that: In step S5, the cross entropy is calculated as follows: Where m represents the sample size; u(·) represents the true distribution of the sample; v(·) represents the expression distribution of the model; and the cross entropy H(u,v) is used to measure the difference between u and v, thereby indicating the difference between the model diagnosis result and the true result.

6. The range extender fault detection model construction method according to claim 1, characterized in that: In step S6, the loss function is: Among them, N represents the sample size; L represents the number of labels; y i,l Indicates that the true result of the i-th sample is l; It represents the probability that the diagnosis result of the i-th sample is l.

7. A range extender fault detection and diagnosis method, characterized in that: The steps include: Step 1: construct a range extender fault detection model using the method described in any one of claims 1 to 6, place the range extender fault detection model in the ECU of the vehicle, and use sensors at key positions of the range extender to collect and obtain multi-channel original signals of the range extender under various working conditions; Step 2: Input the multi-channel original signal collected in real time into the ECU, use the range extender fault detection model to perform real-time detection, and output the working condition expression result of the range extender; Step 3: If the result of the working condition expression is the label corresponding to the normal working condition of the range extender, it indicates that the range extender is in a normal working condition, and step 4 is performed; if the result of the working condition expression is the label corresponding to the fault working condition of the range extender, it indicates that the range extender is in a fault working condition, and step 5 is performed; Step 4: Execute step 2; Step 5: The ECU controls the instrument panel to light up the range extender fault indicator light, indicating a range extender fault, and records the range extender operating condition expression results.

8. A range extender maintenance method, characterized in that: The range extender fault detection and diagnosis method as claimed in claim 7 is used to detect the operating status of the range extender in real time. When it is detected that the range extender is in a faulty operating condition, the operating condition expression result of the range extender is recorded and the fault type is determined to maintain the range extender.

Images