Overload flexible prevention and control method and device for building power supply and distribution system

Abstract

The invention relates to a building power supply and distribution system overload flexible prevention and control method and device. According to the method, a prevention and control device is adopted to collect active power of a power supply and distribution incoming line loop of a building, set temperature of an air conditioner and indoor temperature according to a time sequence, a predicted value of active power of an incoming line loop load at a (t + 1) moment is obtained based on a WOA-CNN-BILSTM model algorithm, and if the predicted value is larger than a preset incoming line loop load protection threshold value pmax, the incoming line loop load protection threshold value pmax is determined. If yes, sending a regulation and control command for reducing power to flexible or translational loads such as an air conditioner; the device is composed of a voltage signal conditioning module, a current signal conditioning module, an electric energy calculation module, a single-chip microcomputer module, an alarm module, an expansion module, a reset circuit module and a power line carrier communication module. According to the invention, flexible regulation and control can be carried out on the flexible or translational load in advance according to the prediction state of the building electrical load, overload tripping power outage accidents caused by insufficient capacity of a building power supply and distribution system are avoided, and meanwhile, the power line carrier communication technology is utilized, so that the cost is saved, and construction and maintenance are convenient.

Description

Technical Field

[0001] The present invention belongs to the field of building electrical safety and power distribution system protection, and relates to a method and device for preventing and controlling building electrical overload, and in particular to a method and device for preventing and controlling overload of a building power supply and distribution system. Background Art

[0002] The full implementation of building electrification has resulted in relatively insufficient capacity of the original power supply and distribution systems in buildings, especially in summer and winter when a large number of air-conditioning loads are put into use. The capacity deficit is even greater and the systems often operate in an overloaded state, posing safety risks such as overload burning.

[0003] The overload prevention and control method currently known in the art is the overcurrent tripping protection method. This method detects the current value of the incoming line circuit of the power supply and distribution system. When it exceeds the preset value and the preset time value, the circuit breaker of the incoming line circuit trips and cuts off the power supply to complete the overload protection. This method is simple and easy to implement, but the accidental tripping and power outage due to overload of the incoming line of the power supply and distribution system will seriously affect the normal operation of the entire building, and even cause personal safety accidents and property losses. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the above-mentioned existing methods and to provide a method and device for flexible overload prevention and control of building power supply and distribution systems, which can prevent accidental tripping and power outages while achieving flexible overload prevention and control of the system.

[0005] The present invention provides a flexible overload prevention and control method for a building power supply and distribution system, comprising the following implementation steps.

[0006] Step 1: Through the three-phase voltage signal conditioning module and the three-phase current signal conditioning module of the overload flexible prevention and control device, the instantaneous value i of the three-phase current of the building incoming line circuit is continuously collected according to the time series (such as every 10 minutes). a (t), i b (t), i c (t), instantaneous value of three-phase voltage u a (t),u b (t),u c (t), the electric energy metering unit calculates the actual active power value p of the incoming line circuit load according to the power definition formula s (t);

[0007] The air conditioner built-in controller continuously collects the air conditioner set temperature t in the same time series s (t) and indoor temperature t n (t), the data is transmitted and modulated into the power line through a general communication module (such as HV2002D-HRC, a product of Xi'an Huarui Network Technology Co., Ltd.), and the power line carrier communication unit of the load flexible prevention and control device demodulates the received data from the power line and transmits it to the single-chip microcomputer module;

[0008] Step 2: Use the normalization formula (1) to calculate the normalized actual active power value p′ of the incoming line load. s (t), normalized air conditioning setting temperature value t′ s (t), normalized indoor temperature value t′ n (t), and combined to form a primary eigenvector x(t); where x and x′ represent the values ​​before and after normalization, max(x) and min(x) represent the maximum and minimum values ​​of the variable x, respectively.

[0009]

[0010] Step 3: Select x(t) data for 21 consecutive days, classify them into training data set A and test data set B in a ratio of 2:1, train and test the basic CNN-BILSTM model, and the data window length is 8. Specifically, it includes the following contents:

[0011] Substitute the primary feature vector x(t) in the training data set A into the one-dimensional convolutional layer CNN model with 32 and 64 superimposed convolution kernels and 1 feature step size, perform feature extraction and dimensionality reduction calculation, and set the activation layer and the maximum pooling layer to obtain the secondary feature vector x t ; The obtained secondary eigenvector x t Substitute into the BILSTM model, that is, calculate according to formula (2) to obtain the predicted active power value p of the incoming line circuit load at time t y (t),

[0012]

[0013] in, is the hidden layer state of the BILSTM network at time t, are the forward / backward LSTM network hidden layer states at time t, h t-1 is the hidden layer state of the forward LSTM network at time (t-1), h t+1 is the hidden layer state of the backward LSTM network at time (t+1), a t is the hidden layer output weight, b t Optimize the hidden layer bias parameters at time t;

[0014] The forward LSTM network calculation is shown in formula (3):

[0015]

[0016] Among them, f t For the forget gate, i t , is the input gate, ct is the current state of the memory unit, c t-1 is the state of the memory unit at time (t-1), o t 、h t is the output gate, W f , U f , W i , U i , W c , U c , W o , U o are weight matrices, b f , b i , b c , b o is the bias matrix, σ is the sigmoid activation function, and tanh is the hyperbolic tangent activation function.

[0017] Furthermore, the optimized hyperparameter values ​​of the CNN-BILSTM model, such as learning rate, number of hidden layer nodes, maximum training cycle, regularization coefficient, etc., are obtained by optimizing the WOA algorithm.

[0018] The whale population is taken as 10, the number of iterations is taken as 8, and the fitness function f is shown in formula (4), so that the optimized hyperparameter values ​​​​of learning rate, number of hidden layer nodes, maximum training cycle and regularization coefficient can be calculated.

[0019]

[0020] The obtained optimized hyperparameter values, learning rate, number of hidden layer nodes, maximum training cycle, and regularization coefficient are substituted into the basic CNN-BILSTM model to obtain the preliminarily trained WOA-CNN-BILSTM model.

[0021] Substitute the primary feature vector x(t) in the test data set B into the preliminarily trained WOA-CNN-BILSTM model for testing and comparison. If the predicted active power value of the incoming circuit load at time t is p y (t) and the normalized actual active power value of the incoming line load p s When the error between ′(t) and ′(t) is within an acceptable range, the initially trained WOA-CNN-BILSTM model is determined to be the trained WOA-CNN-BILSTM model and saved. Otherwise, the training is continued with an increased amount of data.

[0022] Step 4: Repeat steps 1 and 2, and substitute the obtained primary feature vector x(t) as the real-time vector at time t into the trained WOA-CNN-BILSTM model to obtain the predicted active power value p of the incoming line load at time (t+1) y (t+1).

[0023] Step 5: If p y (t+1) is greater than the load protection threshold p of the incoming line circuit of the distribution system max , the single-chip microcomputer of the overload flexible prevention and control device gives the air conditioner set temperature adjustment instruction to the built-in controller of the air conditioner through the power line carrier communication unit, thereby reducing the air conditioner working active power value p k (t+1) (the air conditioner active power changes by about 7% for every 1°C change in the air conditioner set temperature), or the overload flexible prevention and control device controls some preset shiftable load tripping to ensure that the actual active power value p of the load on the incoming line circuit of the distribution system s (t+1) is less than p max , to prevent the power supply and distribution system from accidentally tripping and power outages due to overload, and at the same time, send an overload flexible prevention and control signal to the control center.

[0024] The flexible overload prevention and control device for the building power supply and distribution system used to implement the above method includes a voltage signal conditioning module, a current signal conditioning module, an electric energy calculation module, a single-chip computer module, an alarm module, an expansion module, a reset circuit module, and a power line carrier communication module. The output end of the voltage signal conditioning module and the output end of the current signal conditioning module are electrically connected to the input end of the electric energy calculation module; the output end of the electric energy calculation module and the reset circuit module are connected to the input end of the single-chip computer module; the output end of the single-chip computer module is respectively connected to the input end of the alarm module and the expansion module; the single-chip computer module is connected to the closing / opening control circuit of the circuit breaker through the expansion module to realize the input / removal control of the movable load; the single-chip computer module is connected to the power line through the power line carrier communication module to complete the data transmission of the built-in controller of the air conditioner and other functions.

[0025] The voltage signal conditioning module is composed of an electrically connected voltage transformer, an I / V conversion circuit, an RC filtering circuit, an amplifying circuit and a limiting circuit.

[0026] The current signal conditioning module is composed of an electrically connected current transformer, an I / V conversion circuit, an RC filtering circuit, an amplifying circuit and a limiting circuit.

[0027] The power line carrier communication module is composed of a universal power line carrier chip and a peripheral universal circuit.

[0028] The present invention has the following beneficial effects:

[0029] 1. The method and device of the present invention can flexibly regulate flexible loads such as air conditioners or shiftable loads in advance according to the predicted state of the building's power load, thereby avoiding overload tripping and power outage accidents caused by insufficient capacity of the power supply and distribution system, which affects the normal operation of the building;

[0030] 2. The method and device described in the present invention utilize power carrier communication. The power line is both a power line and a signal line. There is no need to deploy dedicated communication cables, which saves costs, is simple to construct, and is easy to maintain. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a structural schematic diagram of the flexible prevention and control device of the building power supply and distribution system according to the present invention. DETAILED DESCRIPTION

[0032] The following is combined with Figure 1 The present invention is described in further detail.

[0033] The basic content of the flexible overload prevention and control method for building power supply and distribution system of the present invention is to continuously collect the instantaneous value i of the three-phase current in the building incoming line circuit according to the time series through the voltage signal conditioning module and the current signal conditioning module of the overload flexible prevention and control device. a (t), i b (t), i c (t), instantaneous value of three-phase voltage u a (t),u b (t),u c (t), after amplification, limiting, filtering, and I / V conversion, it is sent to the electric energy calculation module with a built-in A / D converter, and the electric energy metering chip calculates the actual active power value p of the incoming line circuit load s (t) and stored in the register, which is read and processed by the single-chip microcomputer module; the single-chip microcomputer module performs two-way data transmission with the built-in controller of the air conditioner through the power line carrier communication module, the power line, and HV2002D-HRC, on the one hand, it collects the air conditioner setting temperature t s (t), indoor temperature t n (t), on the one hand, changing the set temperature value of the air conditioner to adjust the operating power of the air conditioner;

[0034] The single-chip microcomputer module performs linear normalization processing on the sampled data according to formula (1) and combines them into x(t);

[0035] Offline, we selected 21 consecutive days of x(t) data and classified them into training data set A and test data set B in a ratio of 2:1. The time sliding window was set to 8. We substituted x(t) in A into a one-dimensional convolutional layer CNN model with 32 and 64 convolution kernels for feature extraction and dimensionality reduction, and obtained x(t) t ; Continue to add x t Substitute into the BILSTM model for calculation, that is, calculate according to formula (2), where the LSTM calculation process is as shown in formula (3), and obtain the predicted active power value p of the incoming line circuit load at time t y(t), the hyperparameters in the model, such as learning rate, hidden layer nodes, maximum training cycle, and regularization coefficient, are optimized by the WOA algorithm and substituted into the basic CNN-BILSTM model to obtain the preliminarily trained WOA-CNN-BILSTM model;

[0036] Substitute x(t) in B into the preliminarily trained WOA-CNN-BILSTM model for testing and comparison. The predicted active power value p of the incoming line load at time t is y (t) and the normalized actual active power value of the incoming line load p s When the error between ′(t) and ′(t) is within an acceptable range, the initially trained WOA-CNN-BILSTM model is determined to be the trained WOA-CNN-BILSTM model and saved, otherwise the training is continued with an increased amount of data;

[0037] The trained WOA-CNN-BILSTM model is saved in the MCU module, and the MCU module continues to collect and process p in the same way as above. s (t), t s (t), t n (t), and substitute the obtained x(t) as the real-time application vector at time t into the stored trained WOA-CNN-BILSTM model to calculate the predicted active power value p of the incoming line load at time (t+1) y (t+1);

[0038] If p y (t+1) is greater than the load protection threshold p of the incoming line circuit of the distribution system max , the single-chip microcomputer module gives the air conditioner set temperature adjustment instruction to the air conditioner built-in controller through the power line carrier communication unit, power line, and HV2002D-HRC, thereby reducing the air conditioner working active power value p at the subsequent moment k (t+1), or the single-chip microcomputer module controls some preset shiftable load tripping through the expansion module to ensure the active power value p of the load in the incoming line circuit of the distribution system s (t+1) is less than p max , to prevent the power supply and distribution system from accidentally tripping and power outages due to overload. At the same time, the microcontroller module sends an overload flexible prevention and control signal to the control center through the alarm module.

[0039] The structure of the embodiment of the flexible overload prevention and control device for building power supply and distribution system of the present invention is as follows: Figure 1 As shown, it includes a voltage signal conditioning module, a current signal conditioning module, an electric energy calculation module, a single-chip computer module, an alarm module, an expansion module, a reset circuit module, a power line carrier communication module, etc., wherein:

[0040] The output end of the voltage signal conditioning module and the output end of the current signal conditioning module are electrically connected to the input end of the electric energy calculation module. The output ends of the electric energy calculation module and the reset circuit module are connected to the input end of the single-chip computer module. The output end of the single-chip computer module is respectively connected to the input ends of the alarm module and the expansion module. The single-chip computer module is connected to the closing / opening control circuit of the circuit breaker K3 through the expansion module to realize the input / removal control of the movable load. The single-chip computer module is connected to the power line through the power line carrier communication module to complete the data transmission of the built-in controller of the air conditioner and other functions;

[0041] The voltage signal conditioning module is composed of an electrically connected voltage transformer (such as GPT-262B), a universal I / V conversion circuit, an RC filtering circuit, an amplification circuit and a limiting circuit;

[0042] The current signal conditioning module is composed of an electrically connected current transformer (such as GCT-210EB), an I / V conversion circuit, an RC filter circuit, an amplifier circuit and a limiter circuit;

[0043] The electric energy calculation module is composed of a universal electric power metering chip (such as ATT7022E) and a peripheral universal resistance and capacitance circuit;

[0044] The single-chip microcomputer module is composed of an electrically connected single-chip microcomputer (such as STM32F103RCT6), a general crystal oscillator circuit, etc.;

[0045] The alarm module and the expansion module are composed of a general photoelectric coupler (such as TLP187) and a general relay;

[0046] The power line carrier communication module consists of a general power line carrier chip (such as SSC1667) and peripheral general circuits.

[0047] The preliminary training and testing process of the device of the present invention is as follows:

[0048] Step 1: Power on the system, self-check, and initialize. If the self-check fails, jump to step 6.

[0049] Step 2: Continuously collect the instantaneous value u of the phase voltage in the incoming line circuit a (t),u b (t),u c (t), instantaneous value of phase current i a (t), i b (t), i c (t) and calculate the actual active power value p of the incoming line circuit load in real time s (t), continuously collect the air conditioning setting temperature t s (t), indoor temperature t n (t);

[0050] Step 3: According to formula (1), the sampled data is linearly normalized and then combined into x(t);

[0051] Step 4: Select x(t) data for 21 consecutive days and classify them into training data set A and test data set B in a ratio of 2:1;

[0052] Step 5: Use x(t) in A to train and optimize the basic CNN-BILSTM model to obtain the preliminarily trained WOA-CNN-BILSTM, and use x(t) in B to test and compare the preliminarily trained WOA-CNN-BILSTM model to obtain the trained WOA-CNN-BILSTM model and save it in the single-chip microcomputer module of the device;

[0053] Step 6: End.

[0054] The working process of the device of the present invention in practical application is as follows:

[0055] Step 1: Power on the system, self-check, and initialize. If the self-check fails, jump to step 8.

[0056] Step 2: Set the load protection threshold p of the incoming line circuit of the power supply and distribution system max ;

[0057] Step 3: Continuously collect the instantaneous value u of the phase voltage in the incoming line circuit a (t),u b (t),u c (t), instantaneous value of phase current i a (t), i b (t), i c (t) and calculate the actual active power value p of the incoming line circuit load in real time s (t), continuously collect the air conditioning setting temperature t s (t), indoor temperature t n (t);

[0058] Step 4: According to formula (1), the sampled data is linearly normalized and then combined into a primary feature vector x(t);

[0059] Step 5: Substitute the obtained x(t) as the application vector at time t into the trained WOA-CNN-BILSTM model stored in the single-chip module of this device to calculate the predicted active power value p of the incoming line load at time (t+1) y (t+1);

[0060] Step 6: If p y (t+1) is greater than the load protection threshold p of the incoming line circuit of the distribution system max, the device gives the air conditioner built-in controller an air conditioner set temperature adjustment instruction, thereby reducing the air conditioner working active power value p k (t+1), or issue a trip command to the preset translatable load, and at the same time, send an overload flexible prevention and control signal to the control center through the alarm module of this device; if p y (t+1) is less than p max , jump to step eight;

[0061] Step 7: Repeat step 3;

[0062] Step 8: End.

[0063] The above description and specific embodiments describe the basic principles, main methods and advantages of the present invention. The present invention can achieve flexible prevention and control of overload of the building power supply and distribution system, and avoid overload tripping and power outage accidents caused by insufficient capacity of the power supply and distribution system. The actual application mode and application occasions of the present invention are not limited by the above embodiments. Without departing from the scope of the main subject of the present invention, the present invention will also have various corresponding changes (for example, the application occasions are changed to other industrial control systems) and improvements. These changes and improvements should fall within the scope of protection required by the present invention, and the scope of protection can be defined by the claims attached to the application text of the present invention and their equivalents.

Claims

1. A flexible overload prevention and control method for a building power supply and distribution system, characterized in that: The implementation steps include the following: Step 1: Through the three-phase voltage signal conditioning module and the three-phase current signal conditioning module of the overload flexible prevention and control device, the instantaneous value i of the three-phase current of the building incoming line circuit is continuously collected according to the time series. a (t), i b (t), i c (t), instantaneous value of three-phase voltage u a (t),u b (t),u c (t), the electric energy metering unit calculates the actual active power value p of the incoming line circuit load according to the power definition formula s (t); The air conditioner built-in controller continuously collects the air conditioner set temperature t in the same time series s (t) and indoor temperature t n (t), the data is transmitted and modulated to the power line through the universal communication module, and the power line carrier communication unit of the load flexible prevention and control device demodulates the received data from the power line and transmits it to the single chip microcomputer module; Step 2: Use the normalization formula (1) to calculate the normalized actual active power value p′ of the incoming line load. s (t), normalized air conditioning setting temperature value t′ s (t), normalized indoor temperature value t′ n (t), and combined to form a primary eigenvector x(t); where x and x′ represent the values ​​before and after normalization, max(x) and min(x) represent the maximum and minimum values ​​of the variable x, respectively. Step 3: Select x(t) data for 21 consecutive days and classify them into training data set A and test data set B in a ratio of 2:

1. Train and test the basic CNN-BILSTM model with a data window length of 8. Specifically, it includes the following contents: Substitute the primary feature vector x(t) in the training data set A into the one-dimensional convolutional layer CNN model with 32 and 64 superimposed convolution kernels and 1 feature step size, perform feature extraction and dimensionality reduction calculation, and set the activation layer and the maximum pooling layer to obtain the secondary feature vector x t ; The obtained secondary eigenvector x t Substitute into the BILSTM model, that is, calculate according to formula (2) to obtain the predicted active power value p of the incoming line circuit load at time t y (t), in, is the hidden layer state of the BILSTM network at time t, are the forward / backward LSTM network hidden layer states at time t, h t-1 is the hidden layer state of the forward LSTM network at time (t-1), h t+1 is the hidden layer state of the backward LSTM network at time (t+1), a t is the hidden layer output weight, b t Optimize the hidden layer bias parameters at time t; The forward LSTM network calculation is shown in formula (3): Among them, f t For the forget gate, i t , is the input gate, c t is the current state of the memory unit, c t-1 is the state of the memory unit at time (t-1), o t 、h t is the output gate, W f , U f , W i , U i , W c , U c , W o , U o are weight matrices, b f 、b i 、b c , b is the bias matrix, σ is the sigmoid activation function, and tanh is the hyperbolic tangent activation function; The hyperparameters of the CNN-BILSTM model are optimized by the WOA algorithm. The specific method is as follows: Taking the whale population as 10 and the number of iterations as 8, the fitness function f is shown in formula (4), so that the optimized hyperparameter values ​​of learning rate, number of hidden layer nodes, maximum training cycle and regularization coefficient can be calculated; Substitute the obtained optimized hyperparameter values, learning rate, number of hidden layer nodes, maximum training cycle, and regularization coefficient, into the basic CNN-BILSTM model to obtain the preliminarily trained WOA-CNN-BILSTM model; Substitute the primary feature vector x(t) in the test data set B into the preliminarily trained WOA-CNN-BILSTM model for testing and comparison. If the predicted active power value of the incoming circuit load at time t is p y (t) and the normalized actual active power value of the incoming line load p s When the error between ′(t) and ′(t) is within an acceptable range, the initially trained WOA-CNN-BILSTM model is determined to be the trained WOA-CNN-BILSTM model and saved, otherwise the training is continued by increasing the amount of data; Step 4: Repeat steps 1 and 2, and substitute the obtained primary feature vector x(t) as the real-time vector at time t into the trained WOA-CNN-BILSTM model to obtain the predicted active power value p of the incoming line load at time (t+1) y (t+1); Step 5: If p y (t+1) is greater than the load protection threshold p of the incoming line circuit of the distribution system max , the single-chip microcomputer of the overload flexible prevention and control device gives the air conditioner set temperature adjustment instruction to the built-in controller of the air conditioner through the power line carrier communication unit, thereby reducing the active power value p of the air conditioner k (t+1), or the overload flexible prevention and control device controls some preset shiftable load tripping to ensure the actual active power value p of the load on the incoming line circuit of the distribution system s (t+1) is less than p max , to prevent the power supply and distribution system from accidentally tripping and power outages due to overload, and at the same time, send an overload flexible prevention and control signal to the control center.

2. A flexible overload prevention and control device for a building power supply and distribution system for implementing the method of claim 1, characterized in that: It includes a voltage signal conditioning module, a current signal conditioning module, an electric energy calculation module, a single chip computer module, an alarm module, an expansion module, a reset circuit module, and a power line carrier communication module; The output end of the voltage signal conditioning module and the output end of the current signal conditioning module are electrically connected to the input end of the electric energy calculation module; the output end of the electric energy calculation module and the reset circuit module are connected to the input end of the single-chip computer module; the output end of the single-chip computer module is respectively connected to the input end of the alarm module and the expansion module; The single-chip microcomputer module is connected to the closing / opening control circuit of the circuit breaker through the expansion module to realize the input / removal control of the movable load; the single-chip microcomputer module is connected to the power line through the power line carrier communication module to complete the data transmission function of the built-in controller of the air conditioner; The voltage signal conditioning module is composed of an electrically connected voltage transformer, an I / V conversion circuit, an RC filter circuit, an amplifier circuit and a limiter circuit; The current signal conditioning module is composed of an electrically connected current transformer, an I / V conversion circuit, an RC filter circuit, an amplifier circuit and a limiter circuit; The power line carrier communication module is composed of a universal power line carrier chip and a peripheral universal circuit.

Images