Power distribution network disaster loss emergency resource deployment method and system based on vqnn model and medium

Abstract

The application discloses a power distribution network disaster loss emergency resource deployment method and system based on a VQNN model, relates to the technical field of emergency resource scheduling, and determines the failure probability of each tower of a power distribution network under natural disasters based on basic operation data and disaster environment data; simulates the power outage diffusion path after the tower disconnection failure to obtain the overall failure probability of the power distribution network under natural disasters, and constructs a power distribution network vulnerability curve based on the overall failure probability to evaluate the disaster loss load loss of the power distribution network, and based on the disaster loss load loss and the load gap, the disaster loss load loss and the emergency power supply resource power balance and the emergency power supply resource capacity are taken as constraint conditions, and the minimum scheduling cost is taken as the target to perform dynamic emergency resource deployment; the application realizes all-around accurate prediction from a single device to a system, from static to dynamic, from tradition to AI combined quantum, and provides a reliable data basis for subsequent resource scheduling.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of emergency resource scheduling, and particularly relates to a power distribution network disaster loss emergency resource deployment method and system based on a VQNN model and a medium. BACKGROUND

[0002] With the frequent occurrence of natural disasters such as earthquakes and mudslides, regional power infrastructure is severely damaged, threatening the safe operation of power grids and the security of people's livelihood, and the vulnerability of existing power systems under natural disasters highlights the urgency of improving the emergency power supply capacity of power distribution networks. Therefore, research on the disaster loss emergency recovery method of the power distribution network under natural disasters is an urgent problem for power grid enterprises.

[0003] A power distribution network refers to a network that receives electric energy from a power transmission network or a distributed power source and distributes electric energy to end users through power distribution lines, transformers and other facilities. With the rapid development of the power industry, the penetration rate of distributed resources such as distributed photovoltaic (PV), energy storage devices and diesel generators in power distribution networks is continuously increasing. In extreme situations such as natural disasters, how to effectively integrate these distributed resources to achieve rapid disaster recovery of power distribution networks has become a research hotspot in the current power field.

[0004] Currently, research on the rapid recovery of damaged power distribution networks under natural disaster conditions mainly focuses on two dimensions: disaster prevention and post-disaster repair. In terms of disaster prevention, existing research mainly improves the disaster resistance of the system through reinforcement of power facilities such as towers, transformers and other structural enhancements, and network topology optimization. However, in the face of the strong destructive power of extreme disasters such as earthquakes and mudslides, such traditional transformation schemes have the limitations of high engineering cost and diminishing marginal benefits. In terms of post-disaster emergency, the conventional approach relies on centralized dispatching of emergency resources such as mobile power generation vehicles and diesel generators after disaster assessment. However, practice shows that this method lacks timeliness and accuracy in disaster site assessment, and the geographical distribution of emergency power sources does not match the disaster area well, resulting in difficulty in meeting the demand for key load supply during power supply recovery. In summary, exploring the coordinated dispatching mechanism of distributed power (DG) resources and mobile power generation equipment will become a potential breakthrough direction for improving the emergency power supply recovery capacity under natural disaster conditions. SUMMARY

[0005] The technical problem to be solved by the present application is that the disaster condition assessment is not accurate and the recovery is delayed in the emergency recovery of the distribution network under natural disaster conditions, and the purpose of the present application is to provide a distribution network disaster loss emergency resource deployment method, system and medium based on a VQNN model, determine the failure probability of each tower of the distribution network under natural disasters based on basic operation data and disaster environment data, simulate the power outage diffusion path after the tower line failure to obtain the overall failure probability of the distribution network under natural disasters, and based on the overall failure probability, build a distribution network vulnerability curve to assess the disaster loss load loss of the distribution network, based on the disaster loss load loss and the load gap, with the balance of disaster loss load loss and emergency power supply resource power and the emergency power supply resource capacity as the constraint condition, and the minimum scheduling cost as the target, to carry out dynamic emergency resource deployment, and realize the all-round accurate prediction from a single device to a system, from static to dynamic, from traditional to AI+ quantum, and provide a reliable data basis for subsequent resource scheduling.

[0006] The present application is realized by the following technical solutions:

[0007] The present application provides a distribution network disaster loss emergency resource deployment method based on a VQNN model, which comprises:

[0008] Collecting basic operation data and disaster environment data of the distribution network;

[0009] Determine the failure probability of each tower of the distribution network under natural disasters based on the basic operation data and the disaster environment data, simulate the power outage diffusion path after the tower line failure to obtain the overall failure probability of the distribution network under natural disasters, and based on the overall failure probability, build a distribution network vulnerability curve;

[0010] Assess the disaster loss load loss of the distribution network in combination with the physical propagation process under natural disasters and the distribution network vulnerability curve;

[0011] Predict the load gap of the distribution network based on the VQNN model;

[0012] Based on the disaster loss load loss and the load gap, with the balance of disaster loss load loss and emergency power supply resource power and the emergency power supply resource capacity as the constraint condition, and the minimum scheduling cost as the target, to carry out dynamic emergency resource deployment.

[0013] Further optimization scheme is that the basic operation data includes power grid basic data and operation state data of the distribution network; the power grid basic data includes: distribution network topology, line profile and distribution equipment profile; the operation state data includes: distribution network operation history record, line load data, distribution transformer load data, user historical power load and historical distributed photovoltaic power generation curve; the disaster environment data includes: disaster power outage range, debris flow condition data, earthquake condition data and meteorological data.

[0014] A further optimized solution involves determining the failure probability of each power distribution tower under natural disasters based on basic operational data and disaster environment data; including the following methods:

[0015] Based on the FEA analysis method, the inertial force transformed from the seismic acceleration time history curve is used as the load of the tower caused by the earthquake, and the seismic inertial force F of the tower is calculated. eq The impact force F of the debris flow on the tower was calculated based on hydrodynamics. df ;

[0016] Based on seismic inertial force F eq and impact force F df Analyze the maximum force on the tower, and determine the probability of tower failure under natural disasters based on the maximum force:

[0017] ;

[0018] Where ∫ denotes the integral; S(F eq The symbol F represents the tower under seismic inertial force F. eq The failure probability is as follows: S(F) df The symbol indicates the tower's resistance to the impact force F of a debris flow. df The probability of failure is as follows: S(collapse|F eq ,F df ) represents the probability of conditional failure of the tower obtained through the engineering simulation method FEA.

[0019] A further optimized scheme involves obtaining the overall failure probability of the distribution network under natural disasters by simulating the power outage propagation path after a pole wire breakage fault, and constructing a distribution network vulnerability curve based on the overall failure probability; including the following methods:

[0020] Based on the fault propagation algorithm and distribution network topology, the power outage propagation path after a pole wire breakage fault under natural disasters is simulated.

[0021] Based on the operation logic of the distribution network protection device, the overall failure probability S of the distribution network under natural disasters is obtained. DN :

[0022] ;

[0023] Where, n DN Indicates the number of poles and towers in the power distribution network; S pole.i γ represents the failure probability of the i-th tower in the distribution network under earthquake and debris flow conditions; i This indicates that the fault in the distribution network is caused by the broken wire of the i-th tower in the FEA analysis method.

[0024] Based on the overall failure probability S DN Construct the distribution network vulnerability curve η fail :

[0025] ;

[0026] Where Φ() represents the cumulative function of the standard normal distribution of the probability of natural disaster failures in the distribution network; ln represents the natural logarithm function; and ψ represents the intensity of the natural disaster, which is F in the case of earthquakes. eq In debris flow natural disasters, F df S DN (μ) represents the overall mean of the distribution network failure probability after multiple simulations; S DN (σ) represents the standard deviation of the overall failure probability of the distribution network after multiple simulations.

[0027] A further optimized solution involves assessing the disaster-induced load loss of the distribution network by combining the physical propagation process under natural disasters and the distribution network vulnerability curve; including the following methods:

[0028] Earthquakes and debris flows are discretized into spatiotemporal dynamic fields, and the intensity of natural disasters is obtained by combining the vulnerability curve of the power distribution network; the diffusion intensity of natural disasters includes the earthquake diffusion impact force and ψ. eq Debris flow diffusion impact force ψ df ;

[0029] The disaster-induced load loss P of the distribution network is obtained based on the intensity of the natural disaster's spread. loss :

[0030] ;

[0031] ;

[0032] Where k represents the set of load nodes in the distribution network; P DN.j ψ represents the j-th load node in the distribution network; std f represents the impact force that caused the tower to collapse; f() represents the intermediate function.

[0033] A further optimized solution is to predict the load gap of the distribution network based on the VQNN model; including the following methods:

[0034] Construct a VQNN model that includes a quantum feature mapping part, a variable quantum circuit part, and a joint distribution modeling part;

[0035] Based on the VQNN model, the user load P of the distribution network is predicted. ld1 and distributed photovoltaic power output P pv ;

[0036] According to user load P ld1 and distributed photovoltaic power output P pv Obtain the load gap P of the distribution network GAP:

[0037] ;

[0038] wherein, max() represents the maximum function; P es represents the power provided by the energy storage of the power distribution network.

[0039] Further optimization scheme is that the joint distribution modeling part outputs the conditional probability distribution of the user load and the distributed photovoltaic output of the power distribution network, and the user load and the distributed photovoltaic output satisfy a two-dimensional Gaussian distribution, and the probability density function is:

[0040] ;

[0041] wherein, σ ld1 , σ pv respectively represent the standard deviation of the user load and the standard deviation of the PV output; λ ld1 , λ pv respectively represent the mean of the user load and the mean of the PV output; represents the correlation coefficient of the user load and the PV output; p(P ld1 , P pv ) represents the probability density function of the user load and the distributed photovoltaic;

[0042] The joint distribution modeling part also obtains the user load prediction value and the PV output prediction value based on the Pauli measurement operator.

[0043] Further optimization scheme is that the emergency resource deployment is performed based on the disaster load loss and the load gap, the disaster load loss and the power balance of the emergency power supply resource and the emergency power supply resource capacity are taken as constraint conditions, and the minimum scheduling cost is taken as a target; including a method:

[0044] Taking the load gap ≤ the power provided by the emergency power supply resource, the power supply duration ≤ the maximum power supply duration of the emergency power supply resource, and the distance from the emergency power supply resource to the fault point of the power distribution network ≤ the maximum distance that the emergency power supply resource can reach as constraint conditions, and taking the minimum scheduling cost as an objective function, in the offline planning stage, a globally optimal emergency resource deployment scheme is generated based on the NSGA-II algorithm, and in the online adjustment stage, the emergency resource deployment scheme is dynamically adjusted based on the PPO algorithm according to the real-time monitoring of the natural disaster evolution data.

[0045] The scheme also provides a power distribution network disaster emergency resource deployment system based on a VQNN model, which is used to implement the power distribution network disaster emergency resource deployment method based on the VQNN model; the system comprises:

[0046] The acquisition module is used to acquire the basic operation data and disaster environment data of the power distribution network.

[0047] The first calculation module is used to determine the failure probability of each tower in the distribution network under natural disasters based on basic operating data and disaster environment data; to simulate the power outage propagation path after a tower wire breakage fault to obtain the overall failure probability of the distribution network under natural disasters; and to construct the distribution network vulnerability curve based on the overall failure probability.

[0048] The assessment module is used to assess the disaster load loss of the distribution network by combining the physical propagation process under natural disasters and the distribution network vulnerability curve;

[0049] The second calculation module is used to predict the load gap of the distribution network based on the VQNN model;

[0050] The deployment module is used to dynamically deploy emergency resources based on disaster load loss and load gap, with the constraints of disaster load loss and emergency power supply resource power balance and emergency power supply resource capacity, and with the goal of minimizing scheduling costs.

[0051] This solution also provides a computer-readable medium storing a computer program, which, when executed by a processor, can implement the above-described method for deploying emergency resources for power distribution network disasters based on the VQNN model.

[0052] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0053] 1. The present invention provides a method, system and medium for deploying emergency resources for power distribution network disasters based on the VQNN model; the solution introduces quantum computing and AI to optimize prediction accuracy and speed, and elevates emergency resource deployment from a simple "post-disaster response" action to a process that runs through the entire process of power grid planning, early warning, response and recovery.

[0054] 2. The distribution network disaster emergency resource deployment method, system, and medium provided by this invention, based on the VQNN model, not only considers disaster environmental data but also combines basic power grid operation data, integrating physical attributes with real-time status, making the estimation of distribution network fault probability far more accurate than a single method; by simulating the power outage propagation path, the chain reaction of the fault can be dynamically evaluated, thereby obtaining the overall fault probability and plotting vulnerability curves. This avoids treating the system as a collection of isolated components, can capture the topological vulnerability of the power grid, and the assessed disaster load loss is more in line with reality.

[0055] 3. The power distribution network disaster loss emergency resource deployment method and system based on the VQNN model and the medium are provided, the load gap of the power distribution network is predicted based on the VQNN model, the variable quantum neural network combines the parallel processing capability of quantum computing and the learning capability of a neural network, can find a global optimal solution or an approximate optimal solution in a huge solution space faster, and has the explicit target of minimum cost, comprehensively considers the transportation cost, the start-up cost and the operation cost of emergency rescue resources under the premise of meeting the demand, and maximizes economic benefits. BRIEF DESCRIPTION OF DRAWINGS

[0056] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows, and it should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope, and for those skilled in the art, other related drawings can also be obtained without creative labor. In the drawings:

[0057] Figure 1 The figure is a flowchart of the power distribution network disaster loss emergency resource deployment method based on the VQNN model.

[0058] Figure 2 The figure is a structural diagram of the power distribution network disaster loss emergency resource deployment system based on the VQNN model. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solutions and advantages of the present application more clear and obvious, the present application will be further described in detail in combination with the embodiments and the drawings, the exemplary embodiments of the present application and the description thereof are only used to explain the present application, and do not limit the present application.

[0060] In view of this, the present application provides the following embodiments to solve the above technical problems:

[0061] Embodiment 1: The embodiment provides a power distribution network disaster loss emergency resource deployment method based on a VQNN model, as shown in the figure, which comprises: Figure 1

[0062] Step 1: Collecting basic operation data and disaster environment data of the power distribution network;

[0063] The basic operation data includes power grid basic data and operation state data of the power distribution network;

[0064] ​The power grid basic data includes: power distribution network topology (including the feeder branch relationship of the power distribution network, the node capacity of the power distribution network, the line switch position of the power distribution network, and the user access point distribution, etc.), line archives (including the voltage level of the power distribution network, the conductor model of the power distribution network, the tower model of the power distribution network, the windproof / waterproof level, and the anti-seismic capacity, etc.), and power distribution equipment archives (power distribution transformer capacity, distributed photovoltaic installed capacity, energy storage capacity and power, windproof or waterproof level, anti-seismic capacity, etc.);

[0065] The operation state data includes: power distribution network operation history records (power distribution network equipment operation time, fault records, power distribution network structure reinforcement records), line load data (including line 96-point or daily load curves), power distribution transformer load data (including line 96-point or daily load curves), user historical power consumption load (including user 96-point or daily power consumption load curves), and historical distributed photovoltaic power generation curves (including distributed photovoltaic 96-point or daily power generation curves);

[0066] The disaster environment data includes: disaster power failure range (including the number of power failure feeders, the number of power failure subareas, and the number of power failure users), debris flow conditions (occurrence location, impact range, flow rate (m / s), and flow volume (m³ / s)), earthquake conditions (including magnitude, epicenter location, focal depth, and intensity distribution), and meteorological data (including temperature, humidity, rainfall, solar radiation, wind speed, and wind direction); the power distribution automation system can be responsible for collecting power grid basic data such as power distribution network topology, line archives, and power distribution equipment archives, as well as power distribution transformer load and line load data; the electricity information system can collect user historical power consumption load and distributed photovoltaic power generation curves; debris flow and other disaster information provided by the emergency center; earthquake data such as magnitude and epicenter location provided by the China Earthquake Network; and weather forecast information such as temperature, humidity, rainfall, and solar radiation provided by the local meteorological station.

[0067] Step two: based on the basic operation data and the disaster environment data, the failure probability of each tower of the power distribution network under natural disasters is determined; the power failure diffusion path after the tower line breakage failure is simulated to obtain the overall failure probability of the power distribution network under natural disasters, and the power distribution network vulnerability curve is constructed based on the overall failure probability;

[0068] In step two, based on the basic operation data and the disaster environment data, the failure probability of each tower of the power distribution network under natural disasters is determined; including the method:

[0069] S21, based on the FEA analysis method, the inertial force converted from the earthquake acceleration time history curve is taken as the load of the tower under the earthquake, and the seismic inertial force F eq of the tower is calculated; df ;

[0070] The seismic inertial force F eq is:

[0071] ;

[0072] wherein: c pole represents the mass of the tower; t represents the time period of the seismic shock wave; a(t) represents the seismic acceleration time history; the impact force F df of the debris flow on the tower is:

[0073] ;

[0074] wherein: p df represents the bulk density of the debris flow; v df represents the flow velocity of the debris flow; b pole represents the drag coefficient of the tower; a pole represents the projected area of the tower in the impact direction of the debris flow.

[0075] After the seismic inertia force and the impact force of the debris flow on the tower are determined, the failure probability of a single tower can be determined based on the maximum action force of the tower.

[0076] S22, the maximum action force of the tower is analyzed based on the seismic inertia force F eq and the impact force F df of the debris flow, and the failure probability of the tower under natural disasters is determined according to the maximum action force:

[0077] ;

[0078] wherein: ∫ represents integration; S(F eq ) represents the failure probability of the tower under the seismic inertia force F eq ; S(F df ) represents the failure probability of the tower under the impact force F df of the debris flow; and S(collapse| F eq ,F df ) represents the conditional failure probability of the tower obtained by the engineering simulation method FEA.

[0079] Natural disasters such as earthquakes and debris flows mainly affect distribution network towers, which may lead to tower collapse, thereby causing line breakage between towers and ultimately causing power outages. The influence mechanism of earthquakes on the collapse of distribution network towers is that the earthquake transmits vibration energy to the tower through ground motion, causing the tower structure to vibrate. When the vibration load exceeds the structural bearing limit of the tower, the tower will collapse. The destruction of debris flow to the distribution network tower is due to the high-speed flow of the debris flow carrying a large amount of sand and stones: on the one hand, the direct impact of the flow will generate a strong action force on the main body of the tower; on the other hand, the scouring of the high-speed flow may damage the tower foundation and weaken its stability; when the combined effects of these actions break through the structural bearing limit of the tower, the tower will collapse.

[0080] FEA simulation method is a kind of engineering simulation method with numerical analysis ability, which can be used to solve the physical field problems of complex structure under earthquake inertia force and debris flow impact force; FEA simulation method discretizes the continuous physical environment into finite interconnected units, establishes the approximate equation of each unit, and solves the response of the entire distribution network to earthquake and debris flow natural disasters.

[0081] In step two, the overall failure probability of the distribution network under natural disasters is obtained by simulating the power outage diffusion path after the tower line fault, and the vulnerability curve of the distribution network is constructed based on the overall failure probability; including method:

[0082] G21, based on the fault propagation algorithm and the topological structure of the distribution network, the power outage diffusion path after the tower line fault under natural disasters is simulated;

[0083] G22, based on the action logic of the protection device of the distribution network, the overall failure probability S of the distribution network under natural disasters is obtained DN :

[0084] ;

[0085] Where, n DN represents the number of towers of the distribution network; S pole.i represents the failure probability of the i-th tower of the distribution network under earthquake and debris flow; γ i represents the FEA analysis method in which the tower line of the i-th tower is broken to cause the failure of the distribution network;

[0086] G23, based on the overall failure probability S DN , the vulnerability curve η of the distribution network is constructed fail :

[0087] ;

[0088] Where, Φ() represents the cumulative function of the standard normal distribution of the natural disaster failure probability of the distribution network; ln represents the natural logarithm function; ψ represents the natural disaster intensity, which is F eq in earthquake natural disasters, and F df in debris flow natural disasters; S DN (μ) represents the overall mean of the overall failure probability of the distribution network after multiple simulations; S DN (σ) represents the standard deviation of the overall failure probability of the distribution network after multiple simulations.

[0089] The fault propagation algorithm FPA simulates the power outage diffusion path after the tower line fault based on the topological structure of the distribution network, determines the island region of the distribution network according to the action logic of the protection device such as pole circuit breaker and drop type fuse, and forms the overall failure probability of the distribution network.

[0090] Step three: evaluate the disaster load loss of the power distribution network based on the physical propagation process under natural disasters and the vulnerability curve of the power distribution network; this step specifically includes the method:

[0091] Disperse the earthquake and debris flow into a space-time dynamic field respectively, and combine the power distribution network vulnerability curve to obtain the natural disaster intensity; the natural disaster diffusion intensity includes the earthquake diffusion impact force and the debris flow diffusion impact force eq The earthquake diffusion impact force and the debris flow diffusion impact force are respectively: df

[0092] The earthquake diffusion impact force and the debris flow diffusion impact force are respectively: eq The earthquake diffusion impact force and the debris flow diffusion impact force are respectively: df

[0093]

[0094] Wherein, F eq.max represents the inertial force of the earthquake epicenter; exp() represents the exponential function; L eq.rad represents the influence radius of the earthquake; L DNeq represents the distance from the tower to the earthquake epicenter; F df.max represents the debris flow impact force of the natural disaster observation point; g df represents the gravity acceleration of the debris flow; L DNdf represents the distance from the tower to the natural disaster observation point; ω represents the slope of the debris flow to the tower channel; δ represents the friction coefficient of the debris flow and the ground.

[0095] In the dynamic evaluation of the disaster damage risk of the power distribution network, the user load loss can be used as the basis for ensuring power supply when formulating the disaster damage emergency recovery scheme of the power distribution network.

[0096] According to the natural disaster diffusion intensity, the disaster load loss P loss of the power distribution network is obtained:

[0097]

[0098]

[0099] Wherein, k represents the set of load nodes in the power distribution network; P DN.j represents the jth load node in the power distribution network; ψ std represents the impact force causing the tower to collapse; f () represents the intermediate function.

[0100] Step four: predict the load gap of the power distribution network based on the VQNN model; this step specifically includes the method:

[0101] S41, construct a VQNN model including a quantum feature mapping part, a variational quantum circuit part and a joint distribution modeling part; specifically,​​​​​

[0102] S411, Let the load state vector of the distribution network be P. ld For: P ld =(p ld.1 , p ld.2 ,…, p ld.n ), where n represents the number of load states, and the load state vector of the distribution network is transformed into a quantum state |φ(P) in Hilbert space through a quantum feature mapping function. ld )>:

[0103] ;

[0104] Where m represents the number of qubits; β y Z represents the mapping coefficient of the y-th qubit; Z represents the normalization coefficient; m bv Represents the computational vector.

[0105] The quantum feature mapping function is learned through the parameters of the VQNN model. The simplified mapping is obtained using angular encoding. The angular encoding R of the VQNN model... ld for:

[0106]

[0107] In the formula: θ represents the quantum rotation angle in the VQNN mapping; cos() represents the cosine function; sin() represents the sine function.

[0108] The quantum rotation angle θ maps the load state vector of the distribution network to a quantum rotation angle:

[0109]

[0110] In the formula: π represents the value of a circle.

[0111] S412, Tensor product of each quantum bit state in the user load |φ(P) ld1 >For:

[0112]

[0113] In the formula, n AE P represents the dimension of the mapping in VQNN; ld1.i Represents the i-th dimension of the user load state vector; Represents the i-th n AE Controlled NOT gate; Represents n AE The controlled NOT gate; the above process will n AE Feature mapping to Hilbert space enables high-dimensional feature expansion.

[0114] In the variational quantum circuit part, the nonlinear features are extracted by designing a parameterized quantum layer, and the parameterized quantum layer U(θ) is:

[0115] ;

[0116] In the formula, U ent represents the quantum entanglement layer of the VQNN, n L represents the number of layers of the quantum entanglement layer; n ql represents the number of quantum bits; R zd represents the rotation gate of the VQNN; θ k.j represents the quantum rotation angle of the kth quantum bit in the jth quantum entanglement layer.

[0117] The quantum state |φ(P ld1 , θ) > output by the user load is:

[0118] ;

[0119] Similarly, the output quantum state of the PV output is calculated.

[0120] S42, based on the VQNN model, the user load P ld1 and the distributed PV output P pv of the power distribution network are predicted; the VQNN model obtains the predicted values f(P ld1 , θ) and f(P pv , θ) of the user load and the PV output through the Pauli measurement operator:

[0121] ;

[0122] In the formula, O ld1 , O pv respectively represent the Pauli measurement operator of the user load and the PV output; |φ(P pv ) > represents the tensor product of the quantum bit state of the PV output; |φ(P pv , θ) > represents the quantum state of the PV output; P hist.ld1 , P hist.pv respectively represent the historical user load and PV output sample data.

[0123] The joint distribution modeling part of the VQNN model outputs the conditional probability distribution of the user load and the distributed PV output of the power distribution network, and the user load and the distributed PV output satisfy a two-dimensional Gaussian distribution, and the probability density function is:

[0124] ;

[0125] Where, σ ld1 , σ pvrespectively represent the standard deviation of user load and the standard deviation of PV output; λ ld1 , λ pv respectively represent the mean of user load and the mean of PV output; represent the correlation coefficient of user load and PV output; p(P ld1 , P pv ) represents the probability density function of user load and distributed photovoltaic;

[0126] The negative log-likelihood loss τ of the minimization of the joint probability density function is:

[0127]

[0128] In the formula: n T represents the emergency power supply period under natural disasters; P ld1.t , P pv.t respectively represent the value of user load and PV output at the t-th moment. The parameter is updated by quantum-classical hybrid optimization to update the value of θ, so as to improve the prediction accuracy of user load and PV output.

[0129] S43, according to user load P ld1 and distributed photovoltaic output P pv , the load gap P GAP of the power distribution network is obtained:

[0130] ;

[0131] Wherein, max() represents the function of finding the maximum; P es represents the power provided by the energy storage of the power distribution network.

[0132] Step five: based on the disaster load loss and the load gap, taking the disaster load loss and the power balance of the emergency power supply resource and the emergency power supply resource capacity as constraint conditions, and taking the minimum scheduling cost as the target, the dynamic emergency resource deployment is carried out. This step specifically includes the method:

[0133] Taking the load gap ≤ the power provided by the emergency power supply resource, the power supply time ≤ the maximum power supply time of the emergency power supply resource, and the distance from the emergency power supply resource to the fault point of the power distribution network ≤ the maximum distance that the emergency power supply resource can reach as constraint conditions, and taking the minimum scheduling cost as the objective function, in the offline planning stage, the globally optimal emergency resource deployment scheme is generated based on the NSGA-II algorithm, and in the online adjustment stage, the emergency resource deployment scheme is dynamically adjusted based on the PPO algorithm according to the real-time monitoring of the natural disaster evolution data.

[0134] Specifically, the emergency power supply resources of the power grid company mainly include mobile power generation vehicles and fixed diesel generators. The mobile power generation vehicles can directly drive to the DN fault point for power supply, while the fixed diesel generators need to be transported to the distribution network fault point by truck for power supply. The global positioning system (GPS) and Beidou positioning device are used to locate the positions of the two types of emergency power supply resources. The Internet of Things sensors are used to collect the remaining fuel quantity and real-time output power of the emergency power supply resources during power supply. The scheduling of the emergency power supply resources is completed by comprehensively considering the distance, fuel quantity and other factors.

[0135] The power supply time T of the emergency power supply resource SUP is:

[0136]

[0137] wherein, H SUP represents the remaining fuel quantity of the emergency power supply resource; η SUP represents the fuel efficiency coefficient of the emergency power supply resource; P SUP represents the power generation power of the emergency power supply resource; T INR represents the maintenance interval of the emergency power supply resource during power supply.

[0138] The state vector E of the emergency power supply resource is constructed as:

[0139]

[0140] wherein, x k , y k represent the geographical information abscissa and ordinate of the kth emergency power supply resource; P SUP.k represents the power provided by the kth emergency power supply resource; T SUP.k represents the power supply time provided by the kth emergency power supply resource.

[0141] The NSGA-II algorithm is a multi-objective optimization algorithm, which can optimize multiple conflicting objectives simultaneously and provide a set of non-dominated optimal solutions. Therefore, the NSGA-II algorithm is used for static scheduling of the emergency power supply resources.

[0142] The minimum scheduling cost optimization objective f FPL calculated by the NSGA-II algorithm is:

[0143] ;

[0144] wherein, P GAP.t represents the distribution network load gap at the tth moment; P loss.t represents the distribution network disaster load loss at the tth moment; P SUP.t represents the power provided by the emergency power supply resource at the tth moment; d SUPrepresents the distance from the emergency power supply resource to the fault point of the distribution network; w SUP represents the weight of the load of the distribution network.

[0145] The constraint condition expression is:

[0146]

[0147] In the formula, T max represents the maximum power supply duration of the emergency resource; d SUP.max represents the maximum distance that the emergency resource can reach.

[0148] In the offline planning stage, the NSGA-II generates a globally optimal emergency power supply resource scheduling scheme; in the online adjustment stage, the PPO algorithm dynamically adjusts the resource allocation according to the real-time monitoring of the earthquake and debris flow disaster evolution data, and continuously optimizes the scheduling strategy.

[0149] The PPO algorithm is a reinforcement learning algorithm based on policy gradient, which limits the step length of policy update to ensure training stability. The algorithm uses an online learning mechanism, which can dynamically iterate the policy according to the changes in power supply demand caused by earthquakes, mudslides and other disasters, and continuously optimize the policy without the need for offline retraining. This feature makes it suitable for dynamic adjustment scenarios of emergency power supply resources.

[0150] The update policy function of the PPO algorithm is:

[0151]

[0152] In the formula, s SMS represents the deployment path strategy of the emergency resource; represents the emergency resource deployment strategy network; V t represents the value strategy at the t time, represents the strategy network at the t time.

[0153] Embodiment 2

[0154] The embodiment provides a distribution network disaster emergency resource deployment system based on a VQNN model, which is used to implement the distribution network disaster emergency resource deployment method based on the VQNN model in Embodiment 1; as shown in Figure 2 The system comprises:

[0155] The acquisition module is configured to acquire basic operation data and disaster environment data of the distribution network.

[0156] The first calculation module is configured to determine the fault probability of each tower of the distribution network under natural disasters based on the basic operation data and the disaster environment data; simulate the power outage diffusion path after the tower disconnection fault to obtain the overall fault probability of the distribution network under natural disasters, and construct a distribution network vulnerability curve based on the overall fault probability.​

[0157] an evaluation module, configured to evaluate disaster loss load loss of the power distribution network in combination with a physical propagation process under the natural disaster and a vulnerable curve of the power distribution network;

[0158] a second calculation module, configured to predict load gap of the power distribution network based on the VQNN model;

[0159] a deployment module, configured to perform dynamic emergency resource deployment with a minimum cost as a target, based on the disaster loss load loss and the load gap, and with a disaster loss load loss and emergency power supply resource power balance and an emergency power supply resource capability as constraint conditions.

[0160] Embodiment 3

[0161] The embodiment provides a computer readable medium, and a computer program is stored on the computer readable medium, wherein the computer program is executed by a processor to implement the method for disaster loss emergency resource deployment of the power distribution network based on the VQNN model in the embodiment 1; and the following steps are specifically executed:

[0162] Step 1: Collecting basic operation data and disaster environment data of the power distribution network;

[0163] Step 2: Determining a fault probability of each tower of the power distribution network under the natural disaster based on the basic operation data and the disaster environment data; simulating a power cut diffusion path after a tower disconnection fault to obtain an overall fault probability of the power distribution network under the natural disaster, and constructing a vulnerable curve of the power distribution network based on the overall fault probability;

[0164] Step 3: Evaluating the disaster loss load loss of the power distribution network in combination with the physical propagation process under the natural disaster and the vulnerable curve of the power distribution network;

[0165] Step 4: Predicting the load gap of the power distribution network based on the VQNN model;

[0166] Step 5: Performing dynamic emergency resource deployment with a minimum cost as a target, based on the disaster loss load loss and the load gap, and with a disaster loss load loss and emergency power supply resource power balance and an emergency power supply resource capability as constraint conditions.

[0167] The above detailed description further describes the purpose, technical solution and beneficial effects of the present application, and it should be understood that the above description is only a specific implementation of the present application and is not used to limit the protection scope of the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A power distribution network disaster loss emergency resource deployment method based on a VQNN model, characterized in that, The application relates to a method for dynamically deploying emergency resources of a power distribution network. The method comprises the following steps: collecting basic operation data and disaster environment data of the power distribution network; determining a failure probability of each tower of the power distribution network under natural disasters based on the basic operation data and the disaster environment data; simulating a power outage diffusion path after tower disconnection failure to obtain an overall failure probability of the power distribution network under natural disasters, and constructing a power distribution network vulnerability curve based on the overall failure probability; evaluating disaster load loss of the power distribution network by combining a physical propagation process under natural disasters and the power distribution network vulnerability curve; predicting a load gap of the power distribution network based on a VQNN model; deploying emergency resources of the power distribution network based on the disaster load loss and the load gap, with the disaster load loss and emergency power supply resource power balance and the emergency power supply resource capacity as constraint conditions and the minimum scheduling cost as an objective. The method for predicting the load gap of the power distribution network based on the VQNN model comprises the following steps: constructing a VQNN model comprising a quantum feature mapping part, a variational quantum circuit part and a joint distribution modeling part; The user load P of the power distribution network is predicted based on the VQNN model ld1 and the distributed photovoltaic output P pv ; According to the user load P ld1 and the distributed photovoltaic output P pv the load gap P GAP of the distribution network is obtained ； where max() denotes the maximum function; P es denotes the power provided by the energy storage of the distribution grid.

2. The VQNN model-based power distribution network disaster loss emergency resource deployment method according to claim 1, characterized in that, the basic operation data comprises power grid basic data and operation state data of the power distribution network; the power grid basic data comprises a power distribution network topology, line archives and power distribution equipment archives; the operation state data comprises power distribution network operation history records, line load data, power distribution transformer load data, user historical power load and historical distributed photovoltaic power generation curves; and the disaster environment data comprises disaster power outage ranges, debris flow condition data, earthquake condition data and meteorological data.

3. The VQNN model-based power distribution network disaster loss emergency resource deployment method according to claim 1, characterized in that, The method for determining the failure probability of each tower of the power distribution network under natural disasters based on the basic operation data and the disaster environment data comprises the following steps: Based on the FEA analysis method, the inertia force converted from the time history curve of earthquake acceleration is taken as the load of the tower under earthquake, and the seismic inertia force F of the tower is calculated eq ; and the impact force F of the debris flow on the tower is calculated based on hydrodynamics df ; Based on seismic inertial force F eq and impact force F df The maximum force of the tower is analyzed, and the failure probability of the tower under natural disasters is determined according to the maximum force: ； Where ∫ denotes the integral; S(F eq The symbol F represents the tower under seismic inertial force F. eq The failure probability is as follows: S(F) df The symbol indicates the tower's resistance to the impact force F of a debris flow. df The probability of failure is as follows: S(collapse|F eq ,F df ) represents the probability of conditional failure of the tower obtained through the engineering simulation method FEA.

4. The VQNN model-based power distribution network disaster loss emergency resource deployment method according to claim 3, characterized in that, the method for simulating the power outage diffusion path after tower disconnection failure to obtain the overall failure probability of the power distribution network under natural disasters, and constructing the power distribution network vulnerability curve based on the overall failure probability; The method comprises the following steps: simulating a power outage diffusion path after tower disconnection failure under natural disasters based on a failure propagation algorithm and a power distribution network topology; Based on the action logic of the power distribution network protection device, the overall failure probability S of the power distribution network under natural disasters is obtained DN : ； where n DN represents the number of towers of the distribution network; S pole.i represents the failure probability of the i-th tower of the distribution network under earthquake and debris flow; γ i represents that the disconnection of the i-th tower in the FEA analysis method leads to the failure of the distribution network; Based on the overall failure probability S DN The power distribution network vulnerability curve η is constructed fail : ； Wherein, Φ() represents the cumulative function of the standard normal distribution of the power distribution network natural disaster failure probability; ln represents the natural logarithm function; ψ represents the natural disaster intensity, which is F eq in the earthquake natural disaster df ; S DN (μ) represents the overall mean of the overall failure probability of the power distribution network after being constructed by multiple simulations; S DN (σ) represents the standard deviation of the overall failure probability of the power distribution network after being constructed by multiple simulations.

5. The VQNN model-based power distribution network disaster loss emergency resource deployment method according to claim 4, characterized in that, The method for evaluating the disaster load loss of the power distribution network by combining the physical propagation process under natural disasters and the power distribution network vulnerability curve comprises the following steps: The earthquake and the debris flow are respectively dispersed into a time-space dynamic field, and the natural disaster intensity is obtained by combining the power distribution network vulnerable curve; the natural disaster diffusion intensity includes the earthquake diffusion impact force and ψ eq the debris flow diffusion impact force ψ df ; According to the intensity of natural disaster diffusion, the disaster loss load loss P of the distribution network is obtained loss : ； ； where k represents the set of load nodes in the distribution network; P DN.j represents the jth load node in the distribution network; ψ std represents the impact force causing the tower to collapse; f () represents an intermediate function.

6. The VQNN model-based power distribution network disaster loss emergency resource deployment method according to claim 1, characterized in that, the joint distribution modeling part outputs a conditional probability distribution of user load and distributed photovoltaic output of the power distribution network, and the user load and the distributed photovoltaic output satisfy a two-dimensional Gaussian distribution, and a probability density function is: ； where σ ld1 , σ pv denote the standard deviation of the user load and the standard deviation of the PV output, respectively; λ ld1 , λ pv denote the mean of the user load and the mean of the PV output, respectively; denotes the correlation coefficient of the user load and the PV output; p(P ld1 , P pv ) denotes the probability density function of the user load and the distributed photovoltaic. the joint distribution modeling part also obtains a user load prediction value and a PV output prediction value based on a Pauli measurement operator.

7. The VQNN model-based power distribution network disaster loss emergency resource deployment method according to claim 1, characterized in that, The method for deploying emergency resources of the power distribution network based on the disaster load loss and the load gap, with the disaster load loss and emergency power supply resource power balance and the emergency power supply resource capacity as constraint conditions and the minimum scheduling cost as an objective comprises the following steps: with the load gap being less than power provided by the emergency power supply resource, the power supply duration being less than the maximum power supply duration of the emergency power supply resource, and the distance from the emergency power supply resource to the fault point of the power distribution network being less than the maximum distance that can be reached by the emergency power supply resource as constraint conditions, and the minimum scheduling cost as an objective function, a globally optimal emergency resource deployment scheme is generated based on an NSGA-II algorithm in an offline planning stage, and the emergency resource deployment scheme is dynamically adjusted based on a PPO algorithm according to real-time monitoring of natural disaster evolution data in an online adjustment stage.

8. The power distribution network disaster loss emergency resource deployment system based on the VQNN model, characterized in that, The power distribution network disaster loss emergency resource deployment method based on the VQNN model of any one of claims 1-7 is implemented; the system comprises: A collection module for collecting basic operation data and disaster environment data of the power distribution network; A first calculation module for determining the failure probability of each tower of the power distribution network under natural disasters based on the basic operation data and disaster environment data; simulating the power outage diffusion path after the tower line breakage failure to obtain the overall failure probability of the power distribution network under natural disasters, and constructing a power distribution network vulnerability curve based on the overall failure probability; An evaluation module for evaluating the disaster loss load of the power distribution network in combination with the physical propagation process under natural disasters and the power distribution network vulnerability curve; A second calculation module for predicting the load gap of the power distribution network based on the VQNN model; A deployment module for deploying dynamic emergency resources based on the disaster loss load and the load gap, with the disaster loss load and the power balance of emergency power supply resources and the capacity of emergency power supply resources as constraint conditions, and the minimum scheduling cost as the target.

9. A computer readable medium having stored thereon a computer program, characterized in that The computer program executed by the processor can implement the power distribution network disaster loss emergency resource deployment method based on the VQNN model of any one of claims 1-7.

Images