Mobile substation power interface plugging abnormity detection circuit and method

Abstract

The invention relates to the technical field of mobile substation detection, in particular to a circuit and a method for detecting plugging abnormity of a power interface of a mobile substation. Comprising a data monitoring module, a threshold adjustment module, a fault diagnosis module and an alarm prompt module. According to the invention, the data monitoring module detects the voltage, current and temperature of the power interface of the mobile substation, establishes an initial judgment standard for abnormal plugging of the power interface of the mobile substation, and uses the threshold adjustment module to establish a corresponding power-current-voltage model. According to the method, the change ratio obtained after comparison of the changed power and the initial power is taken as a proportionality coefficient, the monitoring range of the current, the voltage and the temperature and the alarm threshold value are adjusted, and the monitoring range of the voltage and the current and the alarm threshold value can be adaptively adjusted when the load power of the mobile substation changes. And the abnormal plugging of the power interface can be detected more accurately.

Description

Technical Field

[0001] The present invention relates to the technical field of mobile substation detection, and in particular to a mobile substation power interface plugging abnormality detection circuit and method. Background Art

[0002] The power interface of a mobile substation is a key component that connects the external power supply and the electrical equipment inside the substation. Its main function is to achieve safe and efficient transmission of electric energy. The power interface of a mobile substation includes a plug-in interface and a fixed interface. The plug-in interface facilitates the quick connection and disconnection of the mobile substation and is widely used in situations where the substation location needs to be frequently moved or where the power supply needs to be temporarily connected. The fixed interface is usually used for mobile substations in a long-term fixed position. The connection is more secure, but the installation and disassembly are relatively complicated. At present, when the power interface of a mobile substation is plugged in and out for abnormal detection, the voltage and current at the interface are detected by a sensor. According to the voltage and current during normal operation of the mobile substation, the abnormal threshold range of voltage and current is set. When the detected voltage and current exceed the threshold range, an alarm signal is issued.

[0003] When the mobile substation is in use, the connected load is changing, and its load power is also changing. If the load power is too high, the voltage and current at the substation interface will also increase. If it exceeds the threshold range, an alarm will be triggered, resulting in a waste of manpower and material resources. If the load power is too low, the voltage and current at the substation interface will also decrease. At this time, an abnormality occurs, resulting in the voltage and current being within the threshold range and unable to trigger an alarm, resulting in the mobile substation being unable to be used normally. Therefore, in order to be able to adaptively adjust the voltage and current monitoring range and the alarm threshold according to the change of load power, so as to more accurately detect the abnormality of the power interface plugging, we propose a mobile substation power interface plugging abnormality detection circuit and method. Summary of the invention

[0004] The purpose of the present invention is to solve the problem that when a mobile substation is in use, its voltage and current monitoring range and alarm threshold are unchanged. When the load power changes, the voltage and current at the power interface of the substation will also change. On the one hand, when the load power increases, the voltage and current at the power interface also increase, causing the voltage and current to exceed the monitoring range and alarm threshold, and issuing an error alarm. On the other hand, when the load power decreases, the voltage and current at the power interface also decrease. If an abnormality occurs at this time, the time for the current and voltage to reach the alarm threshold will increase, and then the alarm cannot be issued in time, causing unnecessary losses. In order to be able to adaptively adjust the voltage and current monitoring range and alarm threshold according to the change of load power, the abnormality of the power interface plugging can be detected more accurately.

[0005] To achieve the above-mentioned purpose, the present invention provides a mobile substation power interface plug-in abnormality detection circuit, including a data monitoring module, a threshold adjustment module, a fault diagnosis module and an alarm prompt module; The data monitoring module detects the voltage, current and temperature at the power interface of the mobile substation through a voltage divider resistor voltage sensor, a Hall current sensor and a temperature sensor, respectively, and converts the analog signals of the voltage, current and temperature into digital signals through an analog-to-digital converter, and calculates the average value, standard deviation and power of the current, voltage and temperature by using a statistical method, and uses this as the initial judgment standard for abnormal plugging of the power interface of the mobile substation, and then transmits the voltage, current and temperature data to the threshold adjustment module; The threshold adjustment module determines the load type according to the voltage and current data transmitted by the data monitoring module, establishes a corresponding power-current-voltage model according to the load type, and when a change in load power is detected, compares the changed power with the initial power to obtain a power change ratio, and uses the proportional coefficient to adjust the monitoring range and alarm threshold of the current, voltage and temperature of the initial judgment standard for abnormal plugging of the power interface of the mobile substation, and establishes a judgment standard for abnormal plugging of the power interface under this load power. If the current, voltage and temperature exceed the threshold range, an alarm signal is transmitted to the alarm prompt module, and at the same time, the change amount of the current, voltage and temperature exceeding the threshold range is transmitted to the fault diagnosis module; The fault diagnosis module uses the changes in current, voltage and temperature that exceed the threshold range as input layer nodes, and the abnormal type of the mobile substation power interface plug-in as the output layer node, establishes a recurrent neural network model, uses the collected historical data to train the model, and determines the abnormal type of the power interface plug-in according to the neural network model, and formulates corresponding solutions, and transmits the abnormal type of the power interface plug-in and the corresponding solutions to the alarm prompt module; When the alarm prompt module receives the alarm signal transmitted by the threshold adjustment module, it immediately triggers the alarm indicator light, and at the same time, voice broadcasts the abnormal type of power interface plug-in and the corresponding solution transmitted by the fault diagnosis module, reminding the staff to repair the power interface of the mobile substation.

[0006] Preferably, the data monitoring module includes an analog-to-digital conversion unit and an initial determination unit; The analog-to-digital conversion unit uses an amplifying and filtering circuit to amplify and filter the voltage collected by the sensor, and then the analog-to-digital converter converts the analog signal into a digital signal for subsequent accurate calculation; The initial determination unit calculates the average value, standard deviation and power of current, voltage and temperature by using statistical methods, and formulates the change range and alarm threshold of current, voltage and temperature, which are used as the initial determination standard of abnormal power interface plugging.

[0007] Preferably, the analog-to-digital conversion unit includes an amplifying and filtering circuit, wherein the amplifying and filtering circuit includes a transistor Q1, a transistor Q2 and an operational amplifier A; The base of the transistor Q1 is connected to one end of the resistor R1 and one end of the resistor R2, the other end of the resistor R1 is connected to the positive electrode of the input voltage UI, the emitter of the transistor Q1 is connected to the negative electrode of the input voltage UI, the collector of the transistor Q1 is connected to one end of the resistor R3 and one end of the resistor R4, the other end of the resistor R4 is connected to pin 2 of the operational amplifier A and one end of the resistor R6 and one end of the capacitor C1, the pin 3 of the operational amplifier A is connected to one end of the resistor R5, the other end of the resistor R5 is grounded, the pin 1 of the operational amplifier A is connected to the base of the transistor Q2 and one end of the resistor R6 and the other end of the capacitor C1, the collector of the transistor Q2 is connected to one end of the resistor R7 and the positive electrode of the output voltage UO, the resistor R7 is connected in parallel to the resistor R3 and the resistor R2 to the power supply VCC, and the emitter of the transistor Q2 is connected to the negative electrode of the output voltage UO.

[0008] Preferably, the threshold adjustment module includes a power detection unit and a ratio adjustment unit; The power detection unit draws a current-voltage curve according to the voltage and current data to determine the load characteristics, and uses Ohm's law and power formula to construct a corresponding power-current-voltage model according to the characteristics and connection mode of the load, and substitutes the voltage and current data into the model to obtain the power data at this time; The proportional adjustment unit adopts a linear adjustment strategy, and uses the change ratio of the changed power compared with the initial power as the proportional coefficient to adjust the monitoring range and alarm threshold of the current, voltage and temperature.

[0009] Preferably, the power detection unit draws a current-voltage curve graph using a polynomial fitting algorithm, and determines the load type as a resistive load, an inductive load, or a capacitive load based on the curve graph.

[0010] Preferably, when the proportion adjustment unit adjusts the monitoring range of current, voltage and temperature and the alarm threshold, the proportion coefficient cannot exceed the safety margin ratio of the load, and the calculation formula of the safety margin ratio is: in, is the voltage safety margin ratio, is the maximum allowable voltage, is the rated voltage.

[0011] Preferably, the fault diagnosis module includes a model building unit and a determination processing unit; When constructing the recurrent neural network model, the model construction unit uses ReLU as the activation function of the hidden layer, uses the softmax activation function in the output layer, independently performs probability calculation on each neuron, and uses the Adam optimization algorithm to propagate the gradient of the loss function back to each layer of the neural network through the back propagation algorithm to update the weights and biases, so that the output of the model gradually approaches the real fault type; The determination processing unit determines the type of abnormality of power interface plugging based on the recurrent neural network model, including short circuit fault, overload fault, poor contact fault, poor heat dissipation fault and power fluctuation fault, and formulates corresponding solutions according to the fault type.

[0012] Preferably, when training the recurrent neural network model, the model building unit uses a binary cross entropy loss function to calculate each neuron separately, thereby measuring the difference between the probability distribution of the model output and the actual fault type label.

[0013] Preferably, the alarm indicator lights of the alarm prompt module are divided into red indicator lights, yellow indicator lights and green indicator lights according to the severity of the fault.

[0014] A second object of the present invention is to provide a method for detecting abnormal plugging of a power interface of a mobile substation, comprising a circuit for detecting abnormal plugging of a power interface of a mobile substation as described in any one of the above, comprising the following steps: S1. Detect the voltage, current and temperature at the power interface of the mobile substation through the data monitoring module, and establish the initial judgment standard for abnormal plugging of the power interface of the mobile substation; S2. Use the threshold adjustment module to establish a corresponding power-current-voltage model, and use the change ratio of the changed power compared with the initial power as the proportional coefficient to adjust the monitoring range and alarm threshold of current, voltage and temperature; S3. Establish a recurrent neural network model through the fault diagnosis module, determine the abnormal type of the power interface plugging according to the neural network model, and formulate corresponding solutions; S4. When the alarm prompt module receives the alarm signal, it immediately triggers the alarm indicator light and simultaneously announces the abnormality type and corresponding solution by voice.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The mobile substation power interface plug-in abnormality detection circuit and method detect the voltage, current and temperature at the mobile substation power interface through a data monitoring module, calculate the average value, standard deviation and power of the current, voltage and temperature by using a statistical method, and use this as the initial judgment standard of the mobile substation power interface plug-in abnormality, use the threshold adjustment module to determine the load type, establish the corresponding power-current-voltage model according to the load type, adopt a linear adjustment strategy, use the change ratio after the change power is compared with the initial power as the proportional coefficient, adjust the monitoring range and alarm threshold of the current, voltage and temperature, and realize that when the load power of the mobile substation changes, the monitoring range and alarm threshold of the voltage and current can be adjusted adaptively, so that the power interface plug-in abnormality can be detected more accurately; A recurrent neural network model is established using the fault diagnosis module. The abnormal type of power interface connection is determined based on the changes in current, voltage and temperature that exceed the threshold range, and corresponding solutions are formulated. The alarm indicator light is triggered by the alarm prompt module, and the abnormal type and corresponding solutions are announced by voice at the same time, reminding the staff to repair the power interface of the mobile substation and reduce unnecessary losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is the overall flow chart of the present invention; Figure 2 It is the overall detailed flow chart of the present invention; Figure 3 It is a flow chart of the fault diagnosis module of the present invention; Figure 4 The amplifying and filtering circuit diagram of the present invention; Figure 5 The figure is a flowchart of the method of the present invention.

[0017] The meaning of each number in the figure is: 100, data monitoring module; 110, analog-to-digital conversion unit; 120, initial determination unit; 200, threshold adjustment module; 210, power detection unit; 220, ratio adjustment unit; 300, fault diagnosis module; 310, model building unit; 320, determination processing unit; 400, alarm prompt module. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] At present, when a mobile substation is in use, the load power changes, causing the voltage and current at the substation interface to change. As a result, when an abnormal plugging occurs at the power interface of the mobile substation, under the influence of the load power, the detected voltage and current cannot exceed the alarm threshold, and thus the alarm cannot be given in time, causing unnecessary losses. In order to be able to adaptively adjust the voltage and current monitoring range and the alarm threshold according to the change of load power, the abnormal plugging of the power interface can be detected more accurately.

[0020] Therefore, the present invention proposes to establish an initial judgment standard for abnormal connection of the power interface of the mobile substation through a data monitoring module, establish a corresponding power-current-voltage model using a threshold adjustment module, use the change ratio after comparing the changed power with the initial power as the proportional coefficient, adjust the monitoring range of current, voltage and temperature and the alarm threshold, establish a recurrent neural network model through a fault diagnosis module, judge the abnormal type of the power interface connection according to the neural network model, and formulate corresponding solutions. When the alarm prompt module receives the alarm signal, it immediately triggers the alarm indicator light and voice broadcasts the abnormal type and the corresponding solution.

[0021] The details are as follows: like Figure 1 As shown, one of the purposes of the present invention is to provide a mobile substation power interface plug-in abnormality detection circuit, including a data monitoring module 100, a threshold adjustment module 200, a fault diagnosis module 300 and an alarm prompt module 400; The data monitoring module 100 detects the voltage, current and temperature at the power interface of the mobile substation through a voltage divider resistor voltage sensor, a Hall current sensor and a temperature sensor, respectively, and converts the analog signals of the voltage, current and temperature into digital signals through an analog-to-digital converter, and calculates the average value, standard deviation and power of the current, voltage and temperature by using a statistical method, and uses this as the initial judgment standard for abnormal plugging of the power interface of the mobile substation, and then transmits the voltage, current and temperature data to the threshold adjustment module 200; The threshold adjustment module 200 determines the load type according to the voltage and current data transmitted by the data monitoring module 100, and establishes a corresponding power-current-voltage model according to the load type. When a change in load power is detected, the changed power is compared with the initial power to obtain a power change ratio, and the current, voltage and temperature monitoring range and alarm threshold of the initial judgment standard for abnormal plugging of the power interface of the mobile substation are adjusted with the proportional coefficient to establish a judgment standard for abnormal plugging of the power interface under this load power. If the current, voltage and temperature exceed the threshold range, an alarm signal is transmitted to the alarm prompt module 400, and at the same time, the change amount of the current, voltage and temperature exceeding the threshold range is transmitted to the fault diagnosis module 300; The fault diagnosis module 300 uses the variation of current, voltage and temperature exceeding the threshold range as the input layer node and the abnormal type of the mobile substation power interface plug-in as the output layer node, establishes a recurrent neural network model, trains the model using the collected historical data, determines the abnormal type of the power interface plug-in according to the neural network model, formulates corresponding solutions, and transmits the abnormal type of the power interface plug-in and the corresponding solutions to the alarm prompt module 400; When the alarm prompt module 400 receives the alarm signal transmitted by the threshold adjustment module 200, it immediately triggers the alarm indicator light, and at the same time voice broadcasts the abnormal type of power interface plug-in and the corresponding solution transmitted by the fault diagnosis module 300, reminding the staff to repair the power interface of the mobile substation.

[0022] like Figure 2 As shown, the data monitoring module 100 includes an analog-to-digital conversion unit 110 and an initial determination unit 120; The analog-to-digital conversion unit 110 uses an amplifying and filtering circuit to amplify and filter the voltage collected by the sensor, and then the analog-to-digital converter converts the analog signal into a digital signal for subsequent accurate calculation; The initial determination unit 120 uses a statistical method to calculate the average value, standard deviation and power of the current, voltage and temperature, and formulates the change range and alarm threshold of the current, voltage and temperature, so as to use it as the initial determination standard of abnormal plugging of the power interface; Taking the current as an example, the average value is calculated as follows: in, is the average current value, For the number, is the corresponding current value.

[0023] The current variance formula is: in, is the variance of the current, is the average current value, For the number, is the corresponding current value.

[0024] The formula for the current error is: in, is the standard deviation of the current, is the variance of the current.

[0025] For example, the following current, voltage, and temperature data are collected: 1 2 3 4 5 6 7 8 9 10 Current 2 3 4 3 2.5 3.5 2 3 4 3 Voltage 220 218 222 219 221 220 221 219 221 218 temperature 30 32 31 29 33 30 31 32 30 31 The average value of the current is: The variance of the current is: The standard deviation of the current is: For each data point , the anomaly coefficient is: For example, for , .

[0026] The corresponding current variation range is: Lower limit: , upper limit: .

[0027] The alarm thresholds are: the lower alarm threshold is 2.1A, and an alarm is triggered when the current is less than 2.1A; the upper alarm threshold is 3.9A, and an alarm is triggered when the current is greater than 3.9A.

[0028] Similarly, the voltage and temperature variation ranges and alarm thresholds are obtained to establish an initial judgment standard for abnormal power interface plugging.

[0029] like Figure 4 As shown, since the signals output by the current, voltage and temperature sensors are usually weak analog signals, in order to more accurately reflect the current, voltage and temperature data, the analog-to-digital conversion unit 110 includes an amplifying and filtering circuit, wherein the amplifying and filtering circuit includes a transistor Q1, a transistor Q2 and an operational amplifier A; The base of transistor Q1 is connected to one end of resistor R1 and one end of resistor R2 in parallel, the other end of resistor R1 is connected to the positive electrode of input voltage UI, the emitter of transistor Q1 is connected to the negative electrode of input voltage UI, the collector of transistor Q1 is connected to one end of resistor R3 and one end of resistor R4 in parallel, the other end of resistor R4 is connected to pin 2 of operational amplifier A and one end of resistor R6 and one end of capacitor C1 in parallel, pin 3 of operational amplifier A is connected to one end of resistor R5, the other end of resistor R5 is grounded, pin 1 of operational amplifier A is connected to the base of transistor Q2 and one end of resistor R6 and the other end of capacitor C1, the collector of transistor Q2 is connected to one end of resistor R7 and the positive electrode of output voltage UO, resistor R7 is connected in parallel to resistor R3 and resistor R2 to power supply VCC, and the emitter of transistor Q2 is connected to the negative electrode of output voltage UO; In this circuit, a first-stage amplifier circuit is formed by transistor Q1 and resistors R1, R2, R3 and R4. When the input voltage UI changes slightly, the base current of transistor Q1 will change. According to the current amplification effect of transistor Q1, the collector current of transistor Q1 will be amplified. As the collector current of transistor Q1 increases, the voltage on resistor R4 will increase, thereby realizing the first-stage amplification of input voltage UI. The voltage amplified by the first-stage amplifier circuit is connected to a filter circuit composed of an operational amplifier A, resistors R5, R6, and capacitor C1. According to the virtual short and virtual open characteristics of the operational amplifier A, the voltages at the non-inverting input terminal 2 and the inverting input terminal 1 are approximately equal (virtual short), and the current flowing into the inverting input terminal 1 is approximately zero (virtual open). The input voltage passes through a low-pass filter network composed of resistor R6 and capacitor C1 to be low-pass filtered. The voltage after low-pass filtering by the filter circuit is secondary amplified by a diode amplifier circuit composed of a transistor Q2 and a resistor R7. The principle is the same as that of the amplifier circuit to obtain an output voltage UO.

[0030] The input voltage UI enters the first-stage amplifier circuit for amplification to obtain an amplified voltage signal, which then enters the low-pass filter circuit. Since the capacitive reactance of the capacitor decreases with increasing frequency, high-frequency noise signals are more easily bypassed to the ground through the capacitor, thereby filtering out high-frequency noise and allowing only low-frequency useful signals to pass. The filtered signal then enters the second-stage amplifier circuit for further amplification to a suitable amplitude for subsequent processing. Filtering first and then amplifying the next stage can effectively suppress the influence of high-frequency noise, improve the quality of the output signal of the entire circuit, and ensure the accuracy and reliability of the current signal detection of the mobile substation power interface.

[0031] The threshold adjustment module 200 includes a power detection unit 210 and a ratio adjustment unit 220; The power detection unit 210 draws a current-voltage curve according to the voltage and current data to determine the load characteristics, and uses Ohm's law and the power formula to construct a corresponding power-current-voltage model according to the characteristics and connection mode of the load, and substitutes the voltage and current data into the model to obtain the power data at this time; The proportional adjustment unit 220 adopts a linear adjustment strategy, and uses the change ratio of the changed power compared with the initial power as the proportional coefficient to adjust the monitoring range and alarm threshold of the current, voltage and temperature; In order to more accurately determine the load type, the power detection unit 210 uses a polynomial fitting algorithm to draw a current-voltage curve graph, and determines the load type as a resistive load, an inductive load, and a capacitive load according to the curve graph; Among them, polynomial fitting is a method of approximating given data points with polynomial functions. Data points ,use Polynomial , use the least squares method to determine the coefficients of the polynomial , the goal of the least squares method is to make the sum of squared errors The minimum, among which, is a polynomial in The value of ; If the fitted curve is a straight line and passes through the origin (or is very close to the origin), then it is a resistive load; For inductive loads, the current lags behind the voltage. On the current-voltage graph, the curve may not be a straight line. Especially in AC circuits, the change in current lags behind the change in voltage. If the peak value of the current lags behind the peak value of the voltage, or the fitted polynomial is not a simple linear function, and the rate of change of the current is not constant as the voltage changes (the curve is curved), it is an inductive load; The current of a capacitive load is ahead of the voltage. On the graph, similar to an inductive load, the curve is usually not a straight line. When the voltage starts to change, the capacitor will charge or discharge quickly, causing the current to change before the voltage changes. If the peak of the current is found to be ahead of the peak of the voltage, or the fitting polynomial shows that the relationship between current and voltage does not conform to the linear law of a resistive load, and the current changes greatly at the beginning of the voltage change and then gradually stabilizes, this is a capacitive load.

[0032] In order to better limit the monitoring range and alarm threshold of current, voltage and temperature, when the proportion adjustment unit (220) adjusts the monitoring range and alarm threshold of current, voltage and temperature, the proportion coefficient cannot exceed the safety margin ratio of the load. The calculation formula of the safety margin ratio is: in, is the voltage safety margin ratio, is the maximum allowable voltage, is the rated voltage.

[0033] The safety margin ratio refers to a safety margin relative to the rated parameters reserved when designing the load operating parameters in order to ensure the safe and reliable operation of the load. For example, for a load with a rated current of 10A, considering possible voltage fluctuations, short-term overloads, etc., the maximum safe current it can actually withstand may be set to 12A, then the safety margin ratio is ; When adjusting the current monitoring range, the maximum current allowed by the load safety margin cannot be exceeded. For example, the load rated current is , the safety margin ratio is , then the maximum safe current is When adjusting the current monitoring range according to factors such as load power, the upper limit should be less than or equal to If this value is exceeded, accurate monitoring may fail when the load approaches or exceeds the safety limit, increasing the risk of load damage. The same is true for voltage and temperature. If the proportional coefficient of the monitoring range and alarm threshold is adjusted to exceed the load safety margin ratio, the load overload may not be detected in time. For example, for a motor load, the rated current is 10A, and the safety margin ratio is set to 20%, that is, the maximum safe current is 12A. If the current monitoring range and alarm threshold are adjusted so that the alarm is not sounded until the current reaches 13A, the motor may continue to run when the safety current exceeds the safety current, causing winding overheating, insulation damage, and even motor burnout. Controlling the proportional coefficient within the safety margin ratio range can ensure that the alarm is sounded before the load approaches the safety limit, and timely measures can be taken, such as reducing the load power or cutting off the power supply, thereby effectively preventing the load from being damaged due to overload.

[0034] like Figure 3 As shown, the fault diagnosis module 300 includes a model building unit 310 and a determination processing unit 320; When constructing the recurrent neural network model, the model construction unit 310 uses ReLU as the activation function of the hidden layer, uses the softmax activation function in the output layer, independently performs probability calculation on each neuron, and uses the Adam optimization algorithm to propagate the gradient of the loss function back to each layer of the neural network through the back propagation algorithm to update the weights and biases, so that the output of the model gradually approaches the real fault type; The determination processing unit 320 determines the type of abnormal connection of the power interface according to the recurrent neural network model, including short circuit fault, overload fault, poor contact fault, poor heat dissipation fault and power fluctuation fault, and formulates corresponding solutions according to the fault type; In order to better measure the difference between the probability distribution output by the model and the real fault type label, the model building unit (310) uses a binary cross entropy loss function to calculate each neuron separately when training the recurrent neural network model, thereby measuring the difference between the probability distribution output by the model and the real fault type label; The principle of the binary cross entropy loss function, for the binary classification problem, assuming that the true label is , the model predicts the output to be , then the binary cross entropy loss function is , the smaller the value of this function, the better the predicted probability distribution of the model With the true label The closer; Assume that the output layer of the recurrent neural network has Neurons, each neuron corresponds to a fault type (here simplified to two categories, such as normal and abnormal), for each neuron , the output of the neuron is converted into a probability distribution through the softmax function, assuming that the neuron The output is , the probability obtained after the softmax function , this probability Indicates that the model predicts that the sample belongs to The probability of a failure type; When there is a real fault type label (Assumption is a one-hot encoding vector. For example, if the true fault type is If a fault occurs, No. elements are 1 and the rest are 0), for each neuron , the loss is calculated according to the binary cross entropy loss function. If the first The prediction of each neuron is considered as a binary classification problem (i.e. type of failure or not), then the loss of this neuron , here Is the real label The vector elements; For example, if the model outputs a probability distribution , the true label , then for the first neuron: ; For the second neuron: ; By analogy, the loss function value of each neuron can be calculated, which can more accurately measure the difference between the probability distribution output by the model and the actual fault type label.

[0035] Among them, the alarm indicator lights of the alarm prompt module 400 are divided into red indicator lights, yellow indicator lights and green indicator lights according to the severity of the fault. By assigning alarm indicator lights of specific colors to faults of different severity, the staff can quickly and intuitively judge the approximate urgency of the fault at a long distance or only by peripheral vision, without having to carefully check the specific fault prompt information or analyze related data, so that corresponding response actions can be made quickly, precious fault handling time can be gained, and the impact of the fault on the operation of the mobile substation can be minimized.

[0036] The second object of the present invention is to provide a method for detecting abnormal plugging of a power interface of a mobile substation, comprising any one of the above-mentioned abnormal plugging detection circuits for a power interface of a mobile substation, such as Figure 5 As shown, the following steps are included: S1. Detect the voltage, current and temperature at the power interface of the mobile substation through the data monitoring module 100, and establish an initial judgment standard for abnormal plugging of the power interface of the mobile substation; S2, using the threshold adjustment module 200 to establish a corresponding power-current-voltage model, using the change ratio of the changed power compared with the initial power as the proportional coefficient, and adjusting the monitoring range and alarm threshold of the current, voltage and temperature; S3, establishing a recurrent neural network model through the fault diagnosis module 300, determining the abnormal type of the power interface plugging according to the neural network model, and formulating corresponding solutions; S4. When the alarm prompt module 400 receives the alarm signal, it immediately triggers the alarm indicator light and simultaneously announces the abnormality type and corresponding solution by voice.

[0037] In summary, the working principle of this solution is as follows: The mobile substation power interface plug-in abnormality detection circuit and method detect the voltage, current and temperature at the mobile substation power interface through the data monitoring module 100, calculate the average value, standard deviation and power of the current, voltage and temperature by using a statistical method, and use this as the initial judgment standard of the mobile substation power interface plug-in abnormality, use the threshold adjustment module 200 to determine the load type, establish the corresponding power-current-voltage model according to the load type, adopt a linear adjustment strategy, use the change ratio after the change power is compared with the initial power as the proportional coefficient, adjust the monitoring range of current, voltage and temperature and the alarm threshold, so that when the load power of the mobile substation changes, the monitoring range of voltage and current and the alarm threshold can be adjusted adaptively, so that the power interface plug-in abnormality can be detected more accurately, and the fault diagnosis module 300 is used to establish a recurrent neural network model, and the abnormal type of the power interface plug-in is judged according to the change amount of the current, voltage and temperature exceeding the threshold range, and the corresponding solution is formulated, and the alarm indicator light is triggered by the alarm prompt module 400, and the abnormal type and the corresponding solution are voice broadcast to remind the staff to repair the mobile substation power interface to reduce unnecessary losses.

[0038] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A mobile substation power interface plug-in abnormality detection circuit, characterized in that: It comprises a data monitoring module (100), a threshold adjustment module (200), a fault diagnosis module (300) and an alarm prompt module (400); The data monitoring module (100) detects the voltage, current and temperature at the mobile substation power interface respectively through a voltage divider resistor voltage sensor, a Hall current sensor and a temperature sensor, converts the analog signals of the voltage, current and temperature into digital signals through an analog-to-digital converter, calculates the average value, standard deviation and power of the current, voltage and temperature using a statistical method, and uses this as an initial judgment standard for abnormal plugging of the mobile substation power interface, and then transmits the voltage, current and temperature data to the threshold adjustment module (200); The threshold adjustment module (200) determines the load type based on the voltage and current data transmitted by the data monitoring module (100), establishes a corresponding power-current-voltage model based on the load type, and when a change in load power is detected, compares the changed power with the initial power to obtain a power change ratio, and uses the ratio coefficient to adjust the current, voltage and temperature monitoring range and alarm threshold of the initial judgment standard for abnormal plugging of the power interface of the mobile substation, and establishes a judgment standard for abnormal plugging of the power interface under this load power. If the current, voltage and temperature exceed the threshold range, an alarm signal is transmitted to the alarm prompt module (400), and at the same time, the change amount of the current, voltage and temperature exceeding the threshold range is transmitted to the fault diagnosis module (300); The fault diagnosis module (300) uses the variation of current, voltage and temperature exceeding the threshold range as input layer nodes and the abnormal type of the mobile substation power interface plugging as output layer nodes, establishes a recurrent neural network model, trains the model using the collected historical data, determines the abnormal type of the power interface plugging according to the neural network model, formulates corresponding solutions, and transmits the abnormal type of the power interface plugging and the corresponding solutions to the alarm prompt module (400); When the alarm prompt module (400) receives the alarm signal transmitted by the threshold adjustment module (200), it immediately triggers the alarm indicator light and simultaneously voice broadcasts the abnormal type of power interface connection transmitted by the fault diagnosis module (300) and the corresponding solution, thereby reminding the staff to repair the mobile substation power interface.

2. The mobile substation power interface plug-in abnormality detection circuit according to claim 1, characterized in that: The data monitoring module (100) comprises an analog-to-digital conversion unit (110) and an initial determination unit (120); The analog-to-digital conversion unit (110) uses an amplifying and filtering circuit to amplify and filter the voltage collected by the sensor, and then uses an analog-to-digital converter to convert the analog signal into a digital signal for subsequent accurate calculation; The initial determination unit (120) uses a statistical method to calculate the average value, standard deviation and power of current, voltage and temperature, and formulates the change range and alarm threshold of current, voltage and temperature, so as to use them as the initial determination standard of abnormal power interface plugging.

3. The mobile substation power interface plug-in abnormality detection circuit according to claim 2, characterized in that: The analog-to-digital conversion unit (110) comprises an amplifying and filtering circuit, wherein the amplifying and filtering circuit comprises a transistor Q1, a transistor Q2 and an operational amplifier A; The base of the transistor Q1 is connected to one end of the resistor R1 and one end of the resistor R2, the other end of the resistor R1 is connected to the positive electrode of the input voltage UI, the emitter of the transistor Q1 is connected to the negative electrode of the input voltage UI, the collector of the transistor Q1 is connected to one end of the resistor R3 and one end of the resistor R4, the other end of the resistor R4 is connected to pin 2 of the operational amplifier A and one end of the resistor R6 and one end of the capacitor C1, the pin 3 of the operational amplifier A is connected to one end of the resistor R5, the other end of the resistor R5 is grounded, the pin 1 of the operational amplifier A is connected to the base of the transistor Q2 and one end of the resistor R6 and the other end of the capacitor C1, the collector of the transistor Q2 is connected to one end of the resistor R7 and the positive electrode of the output voltage UO, the resistor R7 is connected in parallel to the resistor R3 and the resistor R2 to the power supply VCC, and the emitter of the transistor Q2 is connected to the negative electrode of the output voltage UO.

4. The mobile substation power interface plug-in abnormality detection circuit according to claim 1, characterized in that: The threshold adjustment module (200) comprises a power detection unit (210) and a ratio adjustment unit (220); The power detection unit (210) draws a current-voltage curve according to the voltage and current data, determines the load characteristics, and uses Ohm's law and a power formula to construct a corresponding power-current-voltage model according to the characteristics and connection mode of the load, and substitutes the voltage and current data into the model to obtain the power data at this time; The proportion adjustment unit (220) adopts a linear adjustment strategy, using the change ratio after the changed power is compared with the initial power as a proportion coefficient, to adjust the monitoring range and alarm threshold of the current, voltage and temperature.

5. The mobile substation power interface plug-in abnormality detection circuit according to claim 4, characterized in that: The power detection unit (210) draws a current-voltage curve graph using a polynomial fitting algorithm, and determines the load type as a resistive load, an inductive load, or a capacitive load based on the curve graph.

6. The mobile substation power interface plug-in abnormality detection circuit according to claim 4, characterized in that: When the proportion adjustment unit (220) adjusts the monitoring ranges of current, voltage and temperature and the alarm threshold, the proportion coefficient cannot exceed the safety margin ratio of the load. The calculation formula of the safety margin ratio is: in, is the voltage safety margin ratio, is the maximum allowable voltage, is the rated voltage.

7. The mobile substation power interface plug-in abnormality detection circuit according to claim 1, characterized in that: The fault diagnosis module (300) comprises a model building unit (310) and a determination processing unit (320); When constructing a recurrent neural network model, the model construction unit (310) uses ReLU as an activation function of a hidden layer, uses a softmax activation function in an output layer, independently performs probability calculation on each neuron, and uses an Adam optimization algorithm to propagate the gradient of the loss function back to each layer of the neural network through a back propagation algorithm to update weights and biases, so that the output of the model gradually approaches the actual fault type; The determination processing unit (320) determines the type of abnormal connection of the power interface based on the recurrent neural network model, including short circuit fault, overload fault, poor contact fault, poor heat dissipation fault and power fluctuation fault, and formulates corresponding solutions according to the fault type.

8. The mobile substation power interface plug-in abnormality detection circuit according to claim 7, characterized in that: When training the recurrent neural network model, the model building unit (310) uses a binary cross entropy loss function to calculate each neuron separately, thereby measuring the difference between the probability distribution output by the model and the actual fault type label.

9. The mobile substation power interface plug-in abnormality detection circuit according to claim 1, characterized in that: The alarm indicator lights of the alarm prompt module (400) are divided into a red indicator light, a yellow indicator light and a green indicator light according to the severity of the fault.

10. A method for detecting abnormal plugging of a power interface of a mobile substation, comprising a circuit for detecting abnormal plugging of a power interface of a mobile substation as claimed in any one of claims 1 to 9, characterized in that: The following steps are involved: S1. Detecting the voltage, current and temperature at the power interface of the mobile substation through the data monitoring module (100), and establishing an initial judgment standard for abnormal plugging of the power interface of the mobile substation; S2, using the threshold adjustment module (200) to establish a corresponding power-current-voltage model, using the change ratio of the changed power compared with the initial power as a proportional coefficient, and adjusting the monitoring range and alarm threshold of the current, voltage and temperature; S3, establishing a recurrent neural network model through the fault diagnosis module (300), determining the abnormal type of the power interface plugging according to the neural network model, and formulating corresponding solutions; S4. When the alarm prompt module (400) receives the alarm signal, it immediately triggers the alarm indicator light and simultaneously announces the abnormality type and corresponding solution by voice.

Images