A power distribution cabinet dynamic reactive power compensation collaborative control method considering new energy access

Abstract

The present application relates to the field of power distribution cabinet control, especially to a power distribution cabinet dynamic reactive power compensation collaborative control method considering new energy access, the method comprising: obtaining the operating state parameters of the power distribution cabinet; extracting the composite disturbance intensity factor based on the photovoltaic active power sequence, calculating the thermal resistance hysteresis factor of the reactive power compensation device under the current temperature state and determining the thermal-electricity collaborative stress ratio; determining the dynamic reactive power residual rate according to the total capacity of the remaining adjustable reactive power capacity and the reactive power compensation device; calculating the dynamic optimization sensitivity, reconstructing the voltage deviation penalty weight of the model predictive control algorithm based on the dynamic optimization sensitivity; reconstructing the quadratic programming prediction target cost function by using the voltage deviation penalty weight to obtain the improved model predictive control algorithm, and controlling the reactive power compensation device by using the improved model predictive control algorithm. The present application realizes the dynamic balance of efficient power grid compensation and safe operation of the equipment.

Description

Technical Field

[0001] This invention relates to the field of power distribution cabinet control, and in particular to a dynamic reactive power compensation collaborative control method for power distribution cabinets that takes into account the access of new energy sources. Background Technology

[0002] With the profound transformation of the global energy structure, the penetration rate of new energy sources such as photovoltaics and wind power in distribution networks has been increasing year by year. However, new energy power generation has significant intermittency, volatility, and randomness. Taking photovoltaic power generation as an example, cloud cover or dissipation can cause changes in its output active power within a short period of time. When this fluctuating power is injected into the distribution network, it can cause frequent over-limit and flickering of the bus voltage, threatening the power quality and reliability of the local distribution network. To suppress this voltage fluctuation, existing distribution networks typically equip distribution cabinets with reactive power compensation devices such as static var generators (SVG).

[0003] In the control strategy of static var generators (SVRs), model predictive control (MPC) has become the mainstream voltage stability compensation algorithm due to its ability to handle multivariable constraints and its fast dynamic response capability. The core of the MPC algorithm lies in solving a cost function that includes voltage tracking error and control energy consumption through rolling calculations, outputting the optimal switching control sequence. However, the penalty weights of the cost function in traditional MPC algorithms are mostly pre-set static scalars. When facing large fluctuations caused by photovoltaic (PV) grid connection, static weights cannot adaptively adjust the control intensity, easily leading to response hysteresis or overshoot.

[0004] To address this issue, existing technologies typically dynamically adjust the weights of model predictive control. A common approach is to directly sum the variances of photovoltaic (PV) fluctuations or apply a fixed proportionality coefficient to the cost function. However, this superficial mechanism suffers from several problems: First, it ignores the multidimensional characteristics of PV fluctuation trends; simple variance only reflects the amplitude of fluctuations and cannot perceive the physical direction and continuous evolution of rapid power changes. Second, it neglects the underlying physical limitations of reactive power compensation equipment, such as the thermal accumulation state of core power devices and the remaining adjustable reactive power margin. This leads to the control algorithm blindly and aggressively pursuing optimization even when the equipment is already at a high-temperature critical state or its reactive power capacity is about to be exhausted, driven by surface variance data. This can easily cause thermal breakdown and burnout of power devices or lead to high-frequency oscillations under controlled conditions. Summary of the Invention

[0005] To address the technical problem that existing reactive power compensation control strategies are prone to causing high-frequency grid oscillations when dealing with high-frequency power fluctuations generated by new energy grid connection, this invention provides a dynamic reactive power compensation collaborative control method for distribution cabinets that considers new energy access.

[0006] This invention provides a dynamic reactive power compensation coordinated control method for distribution cabinets that considers the access of new energy sources, and adopts the following technical solution:

[0007] A method for dynamic reactive power compensation and coordinated control of distribution cabinets considering the access of new energy sources includes the following steps:

[0008] The system acquires the operating status parameters of the distribution cabinet, including the photovoltaic active power sequence, the actual operating temperature of the reactive power compensation equipment, and the remaining adjustable reactive power capacity. Based on the photovoltaic active power sequence, a composite disturbance intensity factor is extracted, and combined with the actual operating temperature, the thermal hysteresis factor of the reactive power compensation equipment under the current temperature condition is calculated. Then, based on the composite disturbance intensity factor and the thermal hysteresis factor, the thermoelectric synergistic stress ratio is determined. The dynamic reactive power surplus rate is determined according to the remaining adjustable reactive power capacity and the total capacity of the reactive power compensation equipment, and the dynamic optimization sensitivity is calculated based on the thermoelectric synergistic stress ratio. The voltage deviation penalty weight of the model predictive control algorithm is reconstructed based on the dynamic optimization sensitivity. The improved model predictive control algorithm is obtained by reconstructing the quadratic programming prediction objective cost function using the voltage deviation penalty weight, and then used to control the reactive power compensation equipment.

[0009] Compared with existing technologies that use static weights or rely solely on shallow variance for dynamic adjustment, which can easily lead to equipment thermal burnout or high-frequency grid oscillations, this new technology, in application scenarios dealing with high-frequency power fluctuations caused by renewable energy grid connection, achieves adaptive and precise adjustment of model predictive control weights by comprehensively considering the multidimensional characteristics of photovoltaic fluctuations and the physical limitations of temperature and remaining capacity at the bottom layer of reactive power compensation equipment. When equipment faces high temperatures or capacity depletion critical points, it can automatically reduce control sensitivity to protect the equipment, effectively avoiding damage to power devices and system instability caused by blindly and aggressively seeking optimization, and improving the overall power quality and power supply reliability of the distribution network.

[0010] Preferably, the steps for obtaining the operating status parameters of the distribution cabinet are as follows: The photovoltaic active power sequence on the bus side of the current distribution cabinet is obtained in real time using a sliding time window of a preset length; isolated noise points in the photovoltaic active power sequence within the current sliding time window are removed using a median filtering algorithm; the actual operating temperature of the reactive power compensation equipment is collected in real time by a temperature sensor deployed on the heat sink substrate of the reactive power compensation equipment, and the maximum allowable insulation temperature of the reactive power compensation equipment is obtained; the remaining adjustable reactive power capacity and the total capacity of the reactive power compensation equipment fed back by the current converter control system are read, and the rated reference power of the distribution network system is read from the system configuration library.

[0011] Compared with traditional solutions that directly use instantaneous sampling data and are easily affected by random noise, in industrial field environmental data acquisition applications, the sliding time window and median filtering algorithm effectively eliminate isolated noise points in the active power sequence. Combined with the underlying hardware sensors, it directly obtains real equipment temperature and capacity data, ensuring high accuracy of input parameters for subsequent control algorithms and the system's anti-interference capability.

[0012] Preferably, the method for calculating the composite disturbance intensity factor is as follows:

[0013] The variance term of the photovoltaic active power sequence within the current sliding time window is divided by the rated reference power of the distribution network system to obtain the first quotient. The absolute value of the average power gradient of the photovoltaic active power sequence within the current sliding time window is multiplied by the sampling time interval of the control system and then divided by the rated reference power of the distribution network system to obtain the second quotient. The square of the first quotient and the square of the second quotient are added together and then the square root is taken to obtain the composite disturbance intensity factor.

[0014] Compared with existing technologies that rely solely on the variance amplitude of time series data, which makes it difficult to distinguish between high-frequency minute fluctuations and continuous unidirectional sudden changes, the introduction of an orthogonal multidimensional analysis method that includes variance and gradient can accurately identify and quantify the true impact intensity of severe oscillations and rapid unidirectional increases or decreases in photovoltaic output power in application scenarios that capture drastic fluctuations in external power grid power, thereby reducing the misjudgment rate of the control system.

[0015] Preferably, the method for calculating the thermal hindrance factor of the reactive power compensation device under the current temperature condition is as follows: divide the actual operating temperature of the reactive power compensation device by the maximum allowable insulation temperature of the reactive power compensation device to obtain the third quotient; calculate the product of the thermistor constant of the heat dissipation substrate material of the reactive power compensation device and the third quotient; perform exponential calculation with the natural constant as the base and the product as the exponent to obtain the thermal hindrance factor.

[0016] In power distribution scenarios operating under high loads and continuous operation, by introducing an exponential thermal constraint calculation mechanism with nonlinear growth characteristics based on temperature ratios, a sufficiently steep thermal penalty coefficient can be provided when the actual temperature of the equipment approaches the physical limit of insulation, thereby effectively preventing thermal breakdown of power semiconductor devices.

[0017] Preferably, the method for calculating the thermoelectric synergistic stress ratio is as follows: add the thermal stagnation factor of the reactive power compensation device under the current temperature condition to the small constant to prevent calculation overflow, and obtain the first sum; divide the composite disturbance intensity factor within the current sliding time window by the first sum to obtain the thermoelectric synergistic stress ratio.

[0018] Compared with traditional control strategies that view external power grid disturbances and internal equipment status in isolation, this approach can accurately reflect the actual disturbance stress that the equipment can withstand under different health conditions by coupling the actual external impact intensity with the internal thermal damping factor. This allows for the release of control authority when the equipment has sufficient thermal margin and the forced restriction of response impulses when it is in a high-temperature dangerous state.

[0019] Preferably, the method for calculating the dynamic optimization sensitivity is as follows: divide the remaining adjustable reactive power capacity fed back by the current converter control system by the total capacity of the reactive power compensation equipment to obtain the dynamic reactive power surplus rate of the current reactive power compensation equipment.

[0020] Multiply the margin sensing gain of the reactive power compensation device by the current dynamic reactive power surplus rate of the reactive power compensation device, and add the result to a preset constant. Calculate the natural logarithm of the sum to obtain the logarithmic result. Calculate the hyperbolic tangent value of the thermoelectric synergistic stress ratio to obtain the hyperbolic tangent result. Multiply the hyperbolic tangent result by the logarithmic result to obtain the dynamic optimization sensitivity used for parameter tuning of the cost function in the model predictive control algorithm.

[0021] Compared to the solution of forcibly outputting power when the device has no power consumption due to fixed compensation gain, which causes system lockout, in scenarios where the reactive power margin changes dynamically, the hyperbolic tangent function is used to achieve nonlinear soft saturation compression of physical dimensions. Combined with the natural logarithm function to fit the law of diminishing marginal utility, the system provides a smooth gain when the capacity is sufficient, and triggers a sharp sensitivity decay protection when it is about to be exhausted, thus preventing the system from falling into high-frequency oscillations under control constraints.

[0022] Preferably, the method for reconstructing the voltage deviation penalty weight of the model predictive control algorithm based on dynamic optimization sensitivity is as follows: multiply the sensitivity mapping adjustment coefficient by the dynamic optimization sensitivity, and then add it to the basic steady-state weight to obtain the reconstructed voltage deviation penalty weight in the model predictive control algorithm.

[0023] Compared with predictive control models that only have a single regulation mechanism, in application scenarios that deal with complex and variable power grid conditions, by superimposing and reconstructing dynamic optimization sensitivity and basic steady-state weights, it can not only ensure that it can provide strong voltage fluctuation compensation when the system is healthy and subjected to strong external shocks, but also ensure that the control system still has the most basic steady-state voltage tracking capability when there are no external disturbances or when the underlying physical protection state is triggered.

[0024] Preferably, the expression for the quadratic programming prediction objective cost function is:

[0025] ; in, This represents the quadratic programming prediction objective cost function of the model predictive control algorithm within the current rolling optimization time domain; This represents the number of prediction time-domain steps in the model predictive control algorithm; This represents the reconstructed voltage deviation penalty weight in the model predictive control algorithm; This represents the predicted voltage value on the busbar side of the distribution cabinet at the predicted step size k. This indicates the reference voltage on the busbar side of the distribution cabinet; This represents the number of control time-domain steps in the model predictive control algorithm; This represents the energy penalty coefficient for controlling incremental constraints; This represents the control increment sequence of the reactive power compensation device at control step k; This represents the discrete time step index of the model predictive control algorithm in the rolling optimization time domain.

[0026] In the real-time rolling optimization distribution network control scenario, a quadratic programming prediction objective cost function is constructed that integrates the voltage deviation penalty weight and energy penalty coefficient after reconfiguration. This enables the solver to achieve the best dynamic balance between rapid voltage tracking compensation and control energy consumption, thereby outputting a smoother and more reasonable control increment sequence.

[0027] Preferably, the method for controlling reactive power compensation equipment using the improved model predictive control algorithm is as follows:

[0028] The actual initial voltage of the distribution cabinet busbar collected at the current moment, the control output state at the previous moment, the reference voltage, and the reconstructed voltage deviation penalty weight are input into the improved model predictive control algorithm to obtain the optimal control increment sequence. The first time step element in the optimal control increment sequence is extracted, converted into the corresponding pulse width modulation switch control sequence, and then sent to the underlying hardware driver unit of the static var generator to trigger the switching transistors on the converter bridge arm to turn on and off.

[0029] The present invention has the following technical effects:

[0030] This invention integrates the multidimensional characteristics of external photovoltaic fluctuations with the underlying physical limitations of reactive power compensation equipment. By calculating the composite disturbance intensity, thermal hindrance factor, and dynamic reactive power surplus rate, it adaptively reconfigures the cost function penalty weights of the control algorithm. This scheme can quickly smooth voltage when the grid experiences severe fluctuations and actively reduce its response to protect the hardware when the equipment faces high temperatures or capacity depletion, achieving a dynamic balance between efficient grid compensation and safe equipment operation. Attached Figure Description

[0031] Figure 1 This is a flowchart of a dynamic reactive power compensation collaborative control method for power distribution cabinets that takes into account the access of new energy sources, according to the present invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] This invention discloses a dynamic reactive power compensation coordinated control method for distribution cabinets that considers the access of new energy sources, referring to... Figure 1 This includes the following steps:

[0034] Step S1: Obtain the operating status parameters of the power distribution cabinet.

[0035] Data acquisition is performed using an industrial control computer deployed next to the power distribution cabinet to obtain the cabinet's operating status parameters. Specifically, the industrial control computer utilizes a length of... The sliding time window is used to obtain the photovoltaic active power sequence on the current distribution cabinet bus side in real time. The median filtering algorithm is used to remove isolated noise points from the photovoltaic active power sequence within the current sliding time window. In this embodiment, the duration of the sliding time window is set to [value missing]. One sampling point.

[0036] Simultaneously, the rated reference power of the distribution network system is read from the system configuration library. The actual operating temperature of the reactive power compensation equipment is collected in real time by using a fiber Bragg grating temperature sensor deployed on the heat sink substrate of the reactive power compensation equipment. The unit is: And obtain the maximum allowable insulation temperature of the reactive power compensation equipment. The unit is: Simultaneously, it reads the remaining adjustable reactive power capacity fed back from the current converter control system. Total capacity of reactive power compensation equipment .

[0037] Step S2: Perturbation feature extraction based on multidimensional analysis.

[0038] Traditional fluctuation assessment methods typically rely solely on the variance amplitude of the time series, making it difficult to distinguish the physical differences between high-frequency minute fluctuations and sustained unidirectional sudden changes, which can easily lead to misjudgments in the control system. Therefore, this step utilizes the geometric properties of signal statistics and calculus to perform orthogonal multidimensional analysis of the variance and gradient of the photovoltaic active power series, extracting a composite disturbance intensity factor to reflect the true impact intensity of photovoltaic injection on the distribution network. The expression for the composite disturbance intensity factor is as follows:

[0039]

[0040] in, This represents the composite disturbance intensity factor within the current sliding time window; This represents the variance term of the photovoltaic active power sequence within the current sliding time window; Indicates the rated reference power of the power distribution network system; This represents the average power gradient of the photovoltaic active power sequence within the current sliding time window; This indicates the sampling time interval of the control system.

[0041] Variance term of the photovoltaic active power sequence within the current sliding time window When this value increases, it reflects the phenomenon of violent oscillations in photovoltaic output power, leading to an increase in the composite perturbation intensity factor within the current sliding time window. The average power gradient of the photovoltaic active power sequence within the current sliding time window tends to increase. When the absolute value of increases, it reflects a rapid unidirectional increase or decrease in photovoltaic power, leading to an increase in the composite disturbance intensity factor within the current sliding time window. The tendency is to increase, thereby achieving effective capture of sustained power surges.

[0042] In summary, the output variables Essentially, it reflects the overall disturbance intensity caused by photovoltaic (PV) injection to the distribution network. When When the value is large, it is determined that the power grid is under a strong external power surge; when When the value is relatively small, it is determined that the power grid is in a relatively stable operating state.

[0043] For example, suppose that within a certain sampling period, the rated reference power of the distribution network system is... The variance term of the extracted photovoltaic active power sequence within the current sliding time window. The average power gradient of the photovoltaic active power sequence within the current sliding time window. The sampling time interval of the control system Substituting into the formula, we get: At this point, the underlying perturbation basis data was extracted.

[0044] Step S3: Calculate the thermoelectric synergistic stress ratio based on the equipment thermal hysteresis constraint.

[0045] After obtaining the characteristics of external disturbances, directly using them to amplify the control gain can easily lead to thermal breakdown of power semiconductor devices when the reactive power compensation equipment operates at high loads for extended periods, resulting in severe internal heat accumulation. To assess the true cost of external disturbances to the equipment under current thermal conditions, this step uses the physical temperature boundary as a constraint and introduces a thermal hysteresis factor, expressed as:

[0046]

[0047] in, This indicates the thermal hysteresis factor of the reactive power compensation device under the current temperature conditions. This indicates the thermistor constant of the heat dissipation substrate material of the reactive power compensation equipment; This indicates the actual operating temperature of the reactive power compensation equipment; This indicates the maximum permissible insulation temperature of the reactive power compensation equipment. (Coefficient) The value range is [3.0, 5.0]. In this embodiment, the value is 4.0. The value of 4.0 is to provide a sufficiently steep exponential penalty when the equipment temperature exceeds the rated value by 80%. It represents an exponential function with a base of constant e, which can provide a thermal penalty coefficient that grows non-linearly as the equipment temperature approaches its physical limit.

[0048] Then, the thermoelectric synergistic stress ratio is calculated, and the expression is:

[0049]

[0050] in, This represents the thermoelectric synergistic stress ratio that couples external power disturbance with internal thermal state. This represents the composite disturbance intensity factor within the current sliding time window; This indicates the thermal hysteresis factor of the reactive power compensation device under the current temperature conditions. This represents a tiny constant used to prevent computational overflow; in this embodiment, it takes the value of... .

[0051] When the actual operating temperature of the reactive power compensation equipment The temperature rises and approaches the maximum allowable insulation temperature of the reactive power compensation equipment. This indicates that the equipment is overheating, causing a sharp increase in the thermal hindrance factor F of the reactive power compensation equipment at the current temperature, which in turn affects the final output thermoelectric synergistic stress ratio. Formation of inhibition.

[0052] In summary, the thermoelectric synergistic stress ratio Essentially, it reflects the actual disturbance stress that the equipment is currently experiencing. When the value is large, it is determined that the equipment has sufficient thermal margin and the external disturbance is strong, and control authority can be released; when When the value is small, the equipment is judged to be in a high-heat dangerous state, and the disturbance stress can be forcibly compressed.

[0053] For example, assume the actual operating temperature of the current reactive power compensation equipment. Maximum permissible temperature of reactive power compensation equipment insulation Thermosensitive constant of the heat dissipation substrate material of reactive power compensation equipment Calculate the thermal hysteresis factor of the reactive power compensation equipment under the current temperature conditions. Further calculation of the thermoelectric synergistic stress ratio. This indicates that the high temperature significantly weakens the system's response impulse to external disturbances.

[0054] Step S4: Reconstruct voltage deviation penalty weights.

[0055] Even if the equipment temperature is safe, excessive compensation commands can still lead to equipment shutdown or high-frequency grid oscillations if the remaining capacity of the reactive power compensation equipment is depleted. Therefore, this step introduces the reactive power surplus rate to perform a final smooth reconstruction of the thermoelectric synergistic stress ratio. First, the dimensionless dynamic reactive power surplus rate is calculated, expressed as:

[0056]

[0057] in, This indicates the current dynamic reactive power surplus rate of the reactive power compensation equipment; This indicates the remaining adjustable reactive power capacity fed back by the current converter control system; This indicates the total capacity of the reactive power compensation equipment.

[0058] Then, the dynamic optimization sensitivity is calculated, expressed as:

[0059]

[0060] in, This represents the dynamic optimization sensitivity used for tuning the cost function in the model predictive control algorithm. This represents the thermoelectric synergistic stress ratio that couples external power disturbance with internal thermal state. This indicates the current dynamic reactive power surplus rate of the reactive power compensation equipment; This coefficient represents the margin-sensing gain of the reactive power compensation equipment. The value range is typically [5.0, 15.0], and in this selected embodiment, it is set to 10.0. The value of 10.0 is chosen to trigger sufficiently sensitive logarithmic smoothing decay when the remaining capacity is below 20%. The hyperbolic tangent function represents the physical meaning of using it, which lies in nonlinearly and softly saturating the infinitely extending thermoelectric stress ratio to a value that is reduced to a specific value. Within the safe range; using the natural logarithm function ( The physical significance of this is that it conforms to the law of diminishing marginal utility of reactive power margin, providing a gradual gain when capacity is sufficient, and generating a sharp sensitivity decay protection when capacity is about to be exhausted.

[0061] The reconstructed voltage deviation penalty weight is expressed as follows:

[0062]

[0063] in, This represents the reconstructed voltage deviation penalty weight in the model predictive control algorithm; Represents the basic steady-state weights; This represents the sensitivity mapping adjustment coefficient; Represents the dynamic optimization sensitivity used for parameter tuning of the cost function in the model predictive control algorithm, and the basic steady-state weights. The value range of is typically [1.0, 5.0]. In this preferred embodiment, the value is 2.0. The value of 2.0 is chosen to ensure that the MPC algorithm still possesses the most basic steady-state voltage tracking capability when there are no external disturbances or the equipment is in a limit protection state. This coefficient The value range is usually [0.5, 2.0]. In this embodiment, the value is 1.0 to ensure a direct and smooth mapping between the units and the numerical scale.

[0064] Reconstructed voltage deviation penalty weight This reflects the final optimization effort after comprehensively considering external disturbances, thermal boundaries, and capacity boundaries. When the value is large, the system is considered healthy and requires strong compensation; when When the value is small, it is determined that the system has triggered the underlying physical limit, and conservative compensation is performed.

[0065] For example, the thermoelectric synergistic stress ratio Assuming the current converter control system has feedback on the remaining adjustable capacity Total capacity of reactive power compensation equipment The dynamic reactive power surplus rate of the current reactive power compensation equipment is then... Calculate the dynamic optimization sensitivity for parameter tuning of the cost function in the model predictive control algorithm. Final voltage deviation penalty weight The numerical transition was smooth, without any radical weight jumps caused by external disturbances.

[0066] Step S5: Improve the model predictive control algorithm to achieve control of reactive power compensation equipment.

[0067] After the weights are dynamically established, the quadratic programming prediction objective cost function in the model predictive control algorithm is reconstructed to obtain the improved model predictive control algorithm. The expression for the quadratic programming prediction objective cost function is as follows:

[0068]

[0069] in, This represents the quadratic programming prediction objective cost function of the model predictive control algorithm within the current rolling optimization time domain; This represents the number of prediction time-domain steps in the model predictive control algorithm; This represents the reconstructed voltage deviation penalty weight in the model predictive control algorithm; This represents the predicted voltage value on the busbar side of the distribution cabinet at the predicted step size k. This indicates the reference voltage on the busbar side of the distribution cabinet; This represents the number of control time-domain steps in the model predictive control algorithm; This represents the energy penalty coefficient for controlling incremental constraints; This represents the control increment sequence of the reactive power compensation device at control step k; where k represents the discrete time step index of the model predictive control algorithm in the rolling optimization time domain.

[0070] Predict time-domain steps In this embodiment, the value is 10, parameter The value of is 3 in this embodiment, which is the energy penalty coefficient. The value is 0.1. The above values ​​are all classic settings in conventional model predictive control algorithms to balance computational burden and predictability.

[0071] The industrial control computer will collect the actual initial voltage of the distribution cabinet busbar in real time. The control output state at the previous moment, and the reference voltage issued by the upper-level power grid dispatcher. and voltage deviation penalty weight The input is fed into the improved model predictive control algorithm, which outputs the optimal control increment sequence. Based on the rolling optimization principle of model predictive control, the industrial control computer extracts the first time step element from this sequence. After converting it into the corresponding pulse width modulation (PWM) switching control sequence, it is directly sent to the underlying hardware driver unit of the static var generator to precisely trigger the on and off of the IGBT switching transistors on the strain gauge bridge arm, thereby completing the physical closed loop of a single dynamic reactive power compensation. Then the time window slides forward to enter the next round of real-time sampling and optimization.

[0072] The improved model predictive control algorithm, during optimization calculations and dynamic simulations, when the dynamic input quantity... When the value increases, it physically reflects that the current power grid is facing a strong photovoltaic power surge, and that the internal heat capacity and reactive power margin of the distribution cabinet are extremely sufficient, leading to a higher cost function. The sensitivity to the first voltage deviation increases dramatically. The solver, in order to reduce the global cost... Minimization tends to output a more aggressive control increment sequence with larger amplitude. This quickly suppresses drastic fluctuations in bus voltage; conversely, when When the value decreases, it physically reflects that the equipment temperature is approaching the insulation limit or the reactive power margin is severely depleted, and the cost function... The penalty for tracking voltage deviations is forcibly weakened. At this point, the solver is subject to a second energy penalty. The rigid constraints will result in a conservative and gradual increase in control output.

[0073] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for dynamic reactive power compensation and coordinated control of a distribution cabinet considering the access of new energy sources, characterized in that, Including the following steps: Obtain the operating status parameters of the distribution cabinet, including the photovoltaic active power sequence, the actual operating temperature of the reactive power compensation equipment, and the remaining adjustable reactive power capacity; The composite disturbance intensity factor is extracted based on the photovoltaic active power sequence, and the thermal stagnation factor of the reactive power compensation equipment under the current temperature condition is calculated in combination with the actual operating temperature. Then, the thermoelectric synergistic stress ratio is determined based on the composite disturbance intensity factor and the thermal stagnation factor. The dynamic reactive power surplus rate is determined based on the remaining adjustable reactive power capacity and the total capacity of the reactive power compensation equipment. The dynamic optimization sensitivity is calculated by combining the thermoelectric synergistic stress ratio. The voltage deviation penalty weight of the model predictive control algorithm is reconstructed based on the dynamic optimization sensitivity. The improved model predictive control algorithm is obtained by reconstructing the quadratic programming prediction objective cost function using the voltage deviation penalty weight. The improved model predictive control algorithm is then used to control the reactive power compensation equipment.

2. The method for dynamic reactive power compensation and coordinated control of a distribution cabinet considering the access of new energy sources, as described in claim 1, is characterized in that... The steps for obtaining the operating status parameters of the distribution cabinet are as follows: First, the photovoltaic active power sequence on the bus side of the current distribution cabinet is obtained in real time using a sliding time window of a preset length. Second, isolated noise points in the photovoltaic active power sequence within the current sliding time window are removed using a median filtering algorithm. Third, the actual operating temperature of the reactive power compensation equipment is collected in real time using a temperature sensor deployed on the heat sink substrate of the reactive power compensation equipment, and the maximum allowable insulation temperature of the reactive power compensation equipment is obtained. Fourth, the remaining adjustable reactive power capacity and the total capacity of the reactive power compensation equipment are read from the current converter control system, and the rated reference power of the distribution network system is read from the system configuration library.

3. The method for dynamic reactive power compensation and coordinated control of a distribution cabinet considering the access of new energy sources, as described in claim 2, is characterized in that... The method for calculating the composite disturbance intensity factor is as follows: The variance term of the photovoltaic active power sequence within the current sliding time window is divided by the rated reference power of the distribution network system to obtain the first quotient. The absolute value of the average power gradient of the photovoltaic active power sequence within the current sliding time window is multiplied by the sampling time interval of the control system and then divided by the rated reference power of the distribution network system to obtain the second quotient. The square of the first quotient and the square of the second quotient are added together and then the square root is taken to obtain the composite disturbance intensity factor.

4. The method for dynamic reactive power compensation and coordinated control of a distribution cabinet considering the access of new energy sources, as described in claim 1, is characterized in that... The method for calculating the thermal hindrance factor of reactive power compensation equipment under the current temperature condition is as follows: divide the actual operating temperature of the reactive power compensation equipment by the maximum allowable insulation temperature of the reactive power compensation equipment to obtain the third quotient; calculate the product of the thermistor constant of the heat dissipation substrate material of the reactive power compensation equipment and the third quotient; perform exponential calculation with the natural constant as the base and the product as the exponent to obtain the thermal hindrance factor.

5. The method for dynamic reactive power compensation and coordinated control of a distribution cabinet considering the access of new energy sources, as described in claim 4, is characterized in that... The method for calculating the thermoelectric synergistic stress ratio is as follows: add the thermal stagnation factor of the reactive power compensation device under the current temperature condition to the small constant to prevent calculation overflow, and obtain the first sum; divide the composite disturbance intensity factor within the current sliding time window by the first sum to obtain the thermoelectric synergistic stress ratio.

6. The method for dynamic reactive power compensation and coordinated control of a distribution cabinet considering the access of new energy sources according to claim 1, characterized in that, The method for calculating the dynamic optimization sensitivity is as follows: divide the remaining adjustable reactive power capacity fed back by the current converter control system by the total capacity of the reactive power compensation equipment to obtain the dynamic reactive power surplus rate of the current reactive power compensation equipment. Multiply the margin sensing gain of the reactive power compensation device by the current dynamic reactive power surplus rate of the reactive power compensation device, and add the result to a preset constant. Calculate the natural logarithm of the sum to obtain the logarithmic result. Calculate the hyperbolic tangent value of the thermoelectric synergistic stress ratio to obtain the hyperbolic tangent result. Multiply the hyperbolic tangent result by the logarithmic result to obtain the dynamic optimization sensitivity used for parameter tuning of the cost function in the model predictive control algorithm.

7. The method for dynamic reactive power compensation and coordinated control of a distribution cabinet considering the access of new energy sources according to claim 1, characterized in that, The method for reconstructing the voltage deviation penalty weight based on the dynamic optimization sensitivity is as follows: multiply the sensitivity mapping adjustment coefficient by the dynamic optimization sensitivity, and then add it to the basic steady-state weight to obtain the reconstructed voltage deviation penalty weight in the model predictive control algorithm.

8. The method for dynamic reactive power compensation and coordinated control of a distribution cabinet considering the access of new energy sources according to claim 1, characterized in that, The expression for the objective cost function predicted by quadratic programming is: ； in, This represents the quadratic programming prediction objective cost function of the model predictive control algorithm within the current rolling optimization time domain; This represents the number of prediction time-domain steps in the model predictive control algorithm; This represents the reconstructed voltage deviation penalty weight in the model predictive control algorithm; This represents the predicted voltage value on the busbar side of the distribution cabinet at the predicted step size k. This indicates the reference voltage on the busbar side of the distribution cabinet; This represents the number of control time-domain steps in the model predictive control algorithm; This represents the energy penalty coefficient for controlling incremental constraints; This represents the control increment sequence of the reactive power compensation device at control step k; This represents the discrete time step index of the model predictive control algorithm in the rolling optimization time domain.

9. A method for dynamic reactive power compensation and coordinated control of a distribution cabinet considering the access of new energy sources, as described in claim 8, is characterized in that... The method for controlling reactive power compensation equipment using the improved model predictive control algorithm is as follows: The actual initial voltage of the distribution cabinet busbar collected at the current moment, the control output state at the previous moment, the reference voltage, and the reconstructed voltage deviation penalty weight are input into the improved model predictive control algorithm to obtain the optimal control increment sequence. The first time step element in the optimal control increment sequence is extracted, converted into the corresponding pulse width modulation switch control sequence, and then sent to the underlying hardware driver unit of the static var generator to trigger the switching transistors on the converter bridge arm to turn on and off.

Images