Distributed real-time state estimation method for collaborative aggregation of electric heat gas comprehensive energy system
By combining Kalman filtering and Monte Carlo simulation methods, the nonlinear characteristics and asynchronous information exchange of electric-thermal-gas systems are processed, real-time state estimation of electric-thermal-gas systems is realized, solving the inability of traditional methods to adapt to multi-energy The challenge of flow coupling improves the accuracy of state estimation and system robustness.
Patent Information
- Application Number
- CN202411210956.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-30
- Publication Date
- 2024-11-22
AI Technical Summary
The traditional state estimation method based on power systems does not adapt to the development trend of multi-energy flow coupling, and cannot effectively carry out real-time state estimation of electric-thermal-gas systems, resulting in challenges in collaborative scheduling of integrated energy systems.
A real-time state estimation method that comprehensively considers the information data of power, thermal and natural gas systems is proposed. Based on the Kalman filtering algorithm and Monte Carlo simulation, combined with the volume Kalman filtering and distributed optimization algorithm, it handles the non-electrical-thermal-gas system Linear features and asynchronous information exchange to realize global state estimation.
Improves the accuracy and efficiency of state estimation of electric-thermal-gas systems, reduces the computational complexity, and improves the system robustness. It is suitable for distributed asynchronous dynamic estimation of electric-thermal-gas systems, and supports collaborative planning on multiple time scales. and unified scheduling.
Smart Images

Figure CN119004723A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of integrated energy information-physical fusion, and specifically relates to a distributed real-time state estimation method for collaborative aggregation of an integrated energy system of electricity, heat and gas. Background Art
[0002] Integrated energy systems are an important part of the current energy revolution. To carry out efficient coordinated control and unified dispatch of integrated energy systems, state estimation, as the foundation and core of integrated energy systems, provides an essential data basis for subsequent operations such as dispatcher flow, safety and stability analysis. However, the traditional state estimation method based on power systems is not suitable for the energy development trend of multi-energy flow coupling. The electric-heat-gas combined system is the most widely used integrated energy system, which includes power systems, thermal systems, natural gas systems and energy conversion units. The online control and management of the electric-heat-gas combined system requires reliable and real-time system operation status. However, due to the lack of real-time measurement and the obstacles to information exchange between different energy systems, the measurement information collected from the measurement device is incomplete, and there are inevitable noise and errors. Although the state estimation of power systems has been widely studied, there is still little research on the electric-heat-gas system. Since power grids, thermal energy grids, natural gas grids, etc. are independent energy networks that have emerged successively with the continuous development of society, they did not consider the coupling with other networks during planning and construction, resulting in problems in the coordinated planning and unified scheduling of multi-energy flow networks. Under the pressure of the above various energy problems, building a heterogeneous and multi-energy integrated energy network with unified planning and design and overall coordinated scheduling, that is, an integrated energy system, has become an important direction of today's energy reform.
[0003] The measurement of the external input and output of the integrated energy system, as well as the description of the operating state and dynamic laws of the integrated energy system, are usually not directly obtained by the measurement system. State estimation is a method that relies on the measurement data obtained by the measurement device or measurement system to estimate the internal state of the system. Therefore, state estimation is of great significance for understanding the internal state of the system and then realizing the regulation and control of the system. State estimation algorithms can be divided into dynamic state estimation and static state estimation from the calculation method. Dynamic state estimation is to estimate the state of the system based on the prediction of the system state and the obtained measurement data at the same time, while static state estimation only determines the state at a certain moment based on the measurement data at that moment. At present, the dynamic state estimation method mainly uses the Kalman filter algorithm and its improved algorithm, among which the most commonly used is the unscented Kalman filter algorithm. The Sage-Husa noise statistical estimator with fading memory exponential weighting is introduced into the unscented Kalman filter algorithm to estimate the mean and variance of the time-varying system noise, which can effectively improve the accuracy of the algorithm. However, since the unscented transformation process is not changed, the calculation efficiency is not improved. Summary of the invention
[0004] Based on the above technical deficiencies, the present invention is based on the hybrid detection scenario of information exchange of the electric-heat-gas combination in the existing field of integrated energy information-physical fusion technology, and proposes a method for real-time state estimation that comprehensively considers the information data of the power, heat, and natural gas systems. The state estimation of the integrated energy system is measured based on the Kalman filter algorithm, and the Monte Carlo simulation is used to analyze the effect of the state estimation. The Monte Carlo experiment uses the flow calculation result as the true value, and superimposes Gaussian white noise on the true value as the measurement value, and then performs state estimation. The proposed method provides a solution for the adaptive distributed asynchronous dynamic estimation of the electric-heat-gas system, which can improve the efficiency and accuracy of the electric-heat-gas system for the development of the measurement information system, and has certain practical significance.
[0005] To achieve the above object, the present invention provides the following technical solutions:
[0006] A distributed real-time state estimation method for collaborative aggregation of electric, thermal and gas integrated energy systems, including:
[0007] Establish a thermal system network model based on thermodynamic difference within the system framework; further co-generate the power system steady-state model with the thermal system model; combine it with the power-heat cogeneration model to establish an electricity-heat-gas system model;
[0008] Based on the electric-heat-gas system model, after fully analyzing the working modes of various energy conversion hubs, a real-time state estimation method based on volumetric Kalman filtering is used to estimate the system state. The state is classified using the observation data and models of each subsystem. A distributed optimization algorithm is used to deal with the nonlinear characteristics of the electric-heat-gas system and to correct, iterate and update the asynchronous information exchange between the electric-heat-gas systems. A collaborative optimization calculation is performed in the whole system to obtain a comprehensive global state estimation.
[0009] Through Monte Carlo experiments, the effectiveness of the real-time state estimation method based on cubature Kalman filtering in improving estimation accuracy, reducing computational complexity and enhancing system robustness is verified;
[0010] This embodiment discloses that a thermal system network model based on thermodynamic differential is established within the system framework. The iterative equation of the thermal system needs to consider water flow continuity, loop pressure, thermal power, temperature mixing and heat loss. Thermodynamics is described by a set of differential equations. The specific process is as follows:
[0011] Step 1: Considering the continuity of water flow, the mass flow rate of water is consistent with the mass flow rate of water entering the bus, which is equal to the mass flow rate of water flowing out of the bus, which is equal to the sum of the water consumption of the bus. The formula is as follows:
[0012] Am=m q (1)
[0013] Where A is the network incidence matrix between the busbar and the branch in the thermal system; m is the mass flow vector in each pipeline, m q is the mass flow vector of the heat load injected into each bus;
[0014] Step 2: Due to the friction inside the pipe, the pressure head of the water flow will drop, and the pressure difference between the busbars will drive the flow of water in turn. Similar to Kirchhoff's voltage law, the head loss around the closed hydraulic circuit must be zero. The circuit pressure calculation formula is as follows:
[0015] BK fric m*|m|=0 (2)
[0016] Where B is the loop incident matrix between the closed loop and each pipeline; K fric is the resistance coefficient of each pipeline; m*|m| is the absolute value of the matrix; * represents the product;
[0017] Step 3: The thermal power generated by the heat source can be calculated as follows:
[0018] φ=C p m q (T s -T r ) (3)
[0019] Where φ is the thermal power of each bus; C p is the specific heat of water; m q is a single variable; T S and T r are the node supply temperature and return temperature respectively;
[0020] Step 4: Heat loss calculation. The relationship between the start and end temperatures of the pipeline can be expressed as:
[0021]
[0022] Among them, T end represents the pipe end temperature, T start is the temperature at the beginning of the pipe, T a is the ambient temperature, m is a single variable, λ is the heat transfer coefficient per unit length, and L is the pipe length;
[0023] Step 5. The water temperature flowing out of the pipe can be calculated by the following formula:
[0024] (Σm out )T out =∑(m in T in ) (5)
[0025] Among them, m out and m in are the mass flow rates of water leaving and entering the pipe, T out is the water temperature entering the pipe head, T in is the water temperature entering the end of the pipe;
[0026] Step 6. Heat energy is slowly transferred through the hot water in the pipe. The change of the pipe inlet temperature will affect the outlet temperature after the drainage time. Considering that the thermal system operates under constant flow and variable temperature, the hydrodynamics can be ignored. Based on the thermodynamics, the focus is mainly on heat transfer delay and heat loss. Based on the law of conservation of heat, the partial differential equation can be obtained. The formula is as follows:
[0027]
[0028] T represents the temperature at any point, S represents the cross-sectional area, t and x represent the start time and the length of the pipeline, C p is the specific heat of water, λ is the heat transfer coefficient per unit length, m is the mass flow rate of water in the pipe, T a is the ambient temperature;
[0029] Step 7: Discretize the partial differential equation based on the Lax-Wendroff method. The proposed difference-based model only requires the last historical state of the node supply and returns its characteristic information, where k represents the instantaneous time and T i and T j The start time and end time are respectively, a is the ambient temperature, and the formula is as follows:
[0030]
[0031] This embodiment discloses that the thermal system model and the power system steady-state model are further co-generated, and the specific process is as follows:
[0032] Step 1: Establish the power system model, P i and Q i is the active power and reactive power, V i is the voltage amplitude, g ij is the conductivity, b ij is the magnetic susceptibility, θ ij is the phase angle difference;
[0033]
[0034] Step 2: Further combine the thermal system model with the power system steady-state model, where φ CHP is the heat output, P CHP is the power output, C mis the heat rate ratio, η S is the ratio of power-heat cogeneration output power, P cap is the maximum output power with zero heat output, η e is the efficiency of the combined heat and power generation unit, and the specific formula is as follows:
[0035]
[0036]
[0037] P cap =η e G in (11)
[0038] This embodiment discloses a natural gas system model, which is combined with a power-heat cogeneration model to establish an electricity-heat-gas system model. The specific process is as follows:
[0039] Step 1: Establish a natural gas model. The transient characteristics of the natural gas system are mainly reflected in the pipeline flow. The following energy conservation law ignoring gravity, inertia and transient pressure can be obtained. The formula is as follows:
[0040]
[0041]
[0042]
[0043]
[0044] Where NP is a set of EPS measurements, Π l,t Represents the pressure of the pipeline length l at time t, m l,t represents the flow rate of the pipe length l at time t, D represents the inner diameter of the pipeline of the natural gas system, R represents the gas constant, T represents the average temperature of the gas in the pipeline, Z represents the average compression coefficient of natural gas, ρ 0 represents the natural gas density under standard conditions, F represents the friction factor, and time t is divided into NT time points according to the time period, where Each node in the natural gas system must maintain mass flow. Node K means that the natural gas inflow is equal to the outflow. They represent the sum of all natural gas flows in and out of the pipeline at node K, Σm k-source and Σm k-load loads represent the natural gas source and natural gas load at node K, respectively. and They represent the flow rate of the booster station and the natural gas flow rate consumed by the booster station, τc represents the mass flow rate of gas consumed by the compressor, S crepresents the flow direction of compressor c relative to node Q, which is at the end of compressor S c =1, and when K is at the compressor head, S c =-1, the node mass flow balance equation is as follows:
[0045]
[0046] Step 2: Combined with the electricity-heat cogeneration model, establish an electricity-heat-gas system model with the following formula:
[0047] h h =f h (h i ,T s ,T r )
[0048] h e =f e (V i ,θ j )
[0049]
[0050] g(h i ,T s ,T r ,V i ,θ j )=0 (17)
[0051] Among them, h h is the true value of the heat network, f h is the heat network flow equation, h e is the real value of the power grid, f is the power flow equation of the power grid, h gas is the true value of the gas grid, f gas is the gas grid power flow equation, g is the coupling element constraint condition, V i is the voltage amplitude in the power grid, θ j is the phase angle, represents the upstream pressure of the natural gas system at time t, represents the pressure downstream of the natural gas system at time t; T s and T r are the node supply temperature and return temperature respectively;
[0052] The present embodiment discloses a real-time state estimation method based on a volumetric Kalman filter, based on an electric-heat-gas system model, and on the basis of fully analyzing various energy conversion hub working modes, estimates the system state, uses the observation data and models of each subsystem to classify the state, adopts a distributed optimization algorithm, handles the nonlinear characteristic problem of the electric-heat-gas system, and corrects, iterates, and updates the asynchronous information exchange between the electric-heat-gas systems, performs collaborative optimization calculations in the entire system, and obtains a comprehensive global state estimation. The specific steps are:
[0053] Step 1: The quality of nonlinear function estimation has a great impact on the performance of the power-heat-gas system. Due to the nonlinear characteristics of power-heat cogeneration, especially the nonlinear characteristics of the power system, the method based on cubature Kalman filtering can improve the state estimation accuracy of the power-heat-gas system. i Generate 2n volume points to capture the mean of the state estimate and the covariance of the two random variables, as follows:
[0054]
[0055]
[0056]
[0057]
[0058] Among them, S k represents the square root matrix at time k, P k is the estimated covariance; is the i-th volume point; ξ i is the basic data point set;[e] i is the i-th column of the identity matrix; w c and w m are the weights of the mean and covariance, respectively;
[0059] Step 2: Time update is performed through a nonlinear function. The following formula can be used to calculate the predicted state after each time update. The formula is as follows:
[0060]
[0061]
[0062]
[0063] Among them, the covariance matrix is represented by Q k-1 express, is the predicted state of a set of volume points;
[0064] Step 3: Using the predicted state, we can also generate a set of Based on the asynchronous distributed dynamic estimation algorithm, the state of the power system, thermal system and natural gas system is estimated at the same time;
[0065] Step 4, after asynchronous distributed state estimation, under normal circumstances, the distributed state execution cycle of the electrical system is 15 minutes, and the execution cycle of the natural gas system is 60 minutes. The dynamic process of the thermal system is much slower than that of the power system, and the state adjustment cannot be performed at the same frequency. Therefore, the three different subsystems should be performed at different frequencies. However, in the electric-gas system, the natural gas system can be considered as an independent component in the electrical system model. The following only considers the state estimation of the power-heat cogeneration system;
[0066] Step 5, when performing the local state estimation of the combined power and heat system, the asynchronous state estimation of the local Kalman filter is performed separately in the combined power and heat system. When the asynchronous state estimation of the Kalman filter converges, the power and heat output of the combined power and heat system will be calculated using the measured state estimation value. The output at this time will be converted into an equivalent output in another energy form, and then transmitted to the estimator in the other system as an enhanced measurement result, which can be used for the next local estimation iteration. The global convergence criterion formula of the two systems is as follows;
[0067]
[0068]
[0069] Among them, φ CHP is the heat output, P CHP is the power output, ε e and ε h are the critical values of the difference between the power generation and the calorific value of the combined power-heat generation system under different voltages;
[0070] This embodiment discloses a Monte Carlo experiment to verify the effect of the real-time state estimation method based on cubature Kalman filtering in improving estimation accuracy, reducing computational complexity and improving system robustness. The specific steps are as follows:
[0071] Due to the randomness of local state estimation, Monte Carlo simulation is used to analyze the effect of state estimation. The Monte Carlo experiment takes the power flow calculation result as the true value, superimposes Gaussian white noise on the true value as the measurement value, and then performs state estimation. The statistical value of the measurement error S M And the estimated error statistic S H The formula is as follows:
[0072]
[0073]
[0074] Where M is the number of tests, N is the number of measurements, and S after joint state estimation is much better than that of individual state estimation. H / S M The error is reduced and the overall error is reduced, which shows that joint state estimation can improve the average effect of state estimation;
[0075] The present invention carried out 1000 Monte Carlo simulations and obtained a series of statistical results. M and the estimated error statistic S H To evaluate the state estimate, S H / S M It can be used to evaluate the filtering effect of state estimation. A value less than 1 indicates that the effect of state estimation is better than that of stateless estimation. The real-time state estimation method based on cubature Kalman filtering is better than that of stateless estimation in improving estimation accuracy, reducing computational complexity and improving system robustness. The smaller the value, the better the estimation effect.
[0076] Positive effects:
[0077] The present invention provides a distributed real-time state estimation method for the coordinated aggregation of an electric, thermal and gas integrated energy system. Based on the existing distributed state estimation method for an electric-heat-gas combined system, combined with asynchronous information exchange, the system is flexibly adjusted according to the actual state and requirements of each component, so that the system is more adaptable to variable working conditions and environments. At the same time, in order to achieve coordinated planning and unified scheduling of electricity, gas, heat and other networks, it is necessary to build an energy management system for the integrated energy system. State estimation is the core and foundation of the integrated energy system. The specific beneficial technical effects are as follows:
[0078] According to the current problems faced by power system state estimation, an improved Kalman filter state estimation method based on adaptive hybrid prediction is proposed, which effectively improves the computational efficiency and accuracy of the state estimation algorithm.
[0079] According to the characteristics of different sub-networks, the state estimation method based on the power system is extended to the natural gas network and thermal energy network, laying the foundation for the state estimation of the power-gas-heat coupled network.
[0080] (3) According to the dynamic characteristics of different sub-networks, an asynchronous distributed state estimation method based on multi-time scale electric-gas-thermal coupled network is constructed, which effectively improves the efficiency of the algorithm while ensuring the dynamic characteristics of the estimation results;
[0081] (4) Based on the interaction of coupling units in the electric-gas-thermal coupling network, a globally consistent algorithm for state estimation of the electric-gas-thermal coupling network is proposed. Through the constraints of the coupling units, the distributed state estimation is unified, and the dynamic state estimation of the electric-gas-thermal coupling network is realized.
[0082] (5) The present invention conducted 1000 Monte Carlo simulations and obtained a series of statistical results. M and the estimated error statistic S H To evaluate the state estimate, S H / S M It can be used to evaluate the filtering effect of state estimation. A value less than 1 indicates that the effect of state estimation is better than that of stateless estimation, and the smaller the value, the better the estimation effect.
[0083] In summary, the distributed real-time state estimation method for the coordinated aggregation of the electric, heat and gas integrated energy system proposed in the present invention constructs a cogeneration system model, further fully considers the operating characteristics of the electric-heat-gas system, and improves the efficiency of distributed dynamic estimation technology in energy system optimization. It has good development prospects and certain research significance. It is suitable for application as a distributed real-time state estimation method for the coordinated aggregation of the electric, heat and gas integrated energy system. BRIEF DESCRIPTION OF THE DRAWINGS
[0084] Figure 1 This is the overall flow chart of the embodiment of this application;
[0085] Figure 2 A flow chart of a thermal system model based on thermodynamic differential is established for the embodiment of the present application;
[0086] Figure 3 This is a flow chart of dynamic state estimation of the electric heating and gas system of the embodiment of the present application;
[0087] Figure 4 This is a flow chart of the execution cycle of state estimation of three subsystems in an embodiment of the present application. DETAILED DESCRIPTION
[0088] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Although the embodiments of the present invention have been shown and described, it can be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the attached claims and their equivalents. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative work are within the scope of protection of the present invention.
[0089] Distributed real-time state estimation method for collaborative aggregation of electric, heat and gas integrated energy systems, such as Figure 1 As shown, the following steps are included:
[0090] A thermal system network model based on thermodynamic differential is established within the system framework; the steady-state model of the power system is further co-generated with the thermal system model; an electric-heat-gas system model is established in combination with the electric-heat cogeneration model; based on the electric-heat-gas system model, after fully analyzing the working modes of various energy conversion hubs, the real-time state estimation method based on the volumetric Kalman filter is used to estimate the system state, and the state is classified using the observed data and models of each subsystem. A distributed optimization algorithm is used to deal with the nonlinear characteristic problems of the electric-heat-gas system and to correct, iterate and update the asynchronous information exchange between the electric-heat-gas systems. Collaborative optimization calculations are performed within the entire system to obtain a comprehensive global state estimation; the Monte Carlo experiment is used to verify the effectiveness of the real-time state estimation method based on the volumetric Kalman filter in improving estimation accuracy, reducing computational complexity and enhancing system robustness;
[0091] A thermal system network model based on thermodynamic difference is established within the system framework. Figure 2 As shown, the iterative equations of the thermal system need to consider water flow continuity, circuit pressure, thermal power, temperature mixing and heat loss. Thermodynamics is described by a set of differential equations. The specific process is:
[0092] Step 1: Considering the continuity of water flow, the mass flow rate of water is consistent with the mass flow rate of water entering the bus, which is equal to the mass flow rate of water flowing out of the bus, which is equal to the sum of the water consumption of the bus. The formula is as follows:
[0093] Am=m q (1)
[0094] Where A is the network incidence matrix between the busbar and the branch in the thermal system; m is the mass flow vector in each pipeline, m q is the mass flow vector of the heat load injected into each bus;
[0095] Step 2: Due to the friction inside the pipe, the pressure head of the water flow will drop, and the pressure difference between the busbars will drive the flow of water in turn. Similar to Kirchhoff's voltage law, the head loss around the closed hydraulic circuit must be zero. The circuit pressure calculation formula is as follows:
[0096] BK fric m*|m|=0 (2)
[0097] Where B is the loop incident matrix between the closed loop and each pipeline; K fric is the resistance coefficient of each pipeline; m*m| is the absolute value of the matrix; * represents the product;
[0098] Step 3: The thermal power generated by the heat source can be calculated as follows:
[0099] φ=C p m q (T s -T r ) (3)
[0100] Where φ is the thermal power of each bus; C p is the specific heat of water; m q is a single variable; T S and T r are the node supply temperature and return temperature respectively;
[0101] Step 4: Heat loss calculation. The relationship between the start and end temperatures of the pipeline can be expressed as:
[0102]
[0103] Among them, T end represents the pipe end temperature, T start is the temperature at the beginning of the pipe, T a is the ambient temperature, m is a single variable, λ is the heat transfer coefficient per unit length, and L is the pipe length;
[0104] Step 5. The water temperature flowing out of the pipe can be calculated by the following formula:
[0105] (Σm out )T out =∑(m in T in ) (5)
[0106] Among them, m out and m in are the mass flow rates of water leaving and entering the pipe, T out is the water temperature entering the pipe head, T in is the water temperature entering the end of the pipe;
[0107] Step 6. Heat energy is slowly transferred through the hot water in the pipe. The change of the pipe inlet temperature will affect the outlet temperature after the drainage time. Considering that the thermal system operates under constant flow and variable temperature, the hydrodynamics can be ignored. Based on the thermodynamics, the focus is mainly on heat transfer delay and heat loss. Based on the law of conservation of heat, the partial differential equation can be obtained. The formula is as follows:
[0108]
[0109] T represents the temperature at any point, S represents the cross-sectional area, t and x represent the start time and the length of the pipeline, C pis the specific heat of water, λ is the heat transfer coefficient per unit length, m is the mass flow rate of water in the pipe, T a is the ambient temperature;
[0110] Step 7: Discretize the partial differential equation based on the Lax-Wendroff method, where k represents the instantaneous time, T i and T j The start time and end time are respectively, a is the ambient temperature, and the formula is as follows:
[0111]
[0112] The thermal model based on thermodynamic difference proposed in the present invention only needs the last historical state supplied by the node and returns its temperature information. Therefore, only a small amount of historical information is needed, which simplifies the dynamic characteristic prediction state of the thermal system;
[0113] The thermal system model is further combined with the power system steady-state model. The specific process is as follows:
[0114] Step 1: Establish the power system model, P i and Q i is the active power and reactive power, V i is the voltage amplitude, g ij is the conductivity, b ij is the magnetic susceptibility, θ ij is the phase angle difference;
[0115]
[0116] Step 2: Further combine the thermal system model with the power system steady-state model, where φ CHP is the heat output, P CHP is the power output, C m is the heat rate ratio, η S is the ratio of power-heat cogeneration output power, P cap is the maximum output power with zero heat output, η e is the efficiency of the combined heat and power generation unit, and the specific formula is as follows:
[0117]
[0118] P cap =η e G in (11)
[0119] Establish a natural gas system model, combine it with the power-heat cogeneration model, and establish an electricity-heat-gas system model. The specific process is as follows:
[0120] Step 1: Establish a natural gas model. The transient characteristics of the natural gas system are mainly reflected in the pipeline flow. The following energy conservation law ignoring gravity, inertia and transient pressure can be obtained. The formula is as follows:
[0121]
[0122]
[0123]
[0124]
[0125] Where NP is a set of EPS measurements, Π l,t Represents the pressure of the pipeline length l at time t, m l,t represents the flow rate of the pipe length l at time t, D represents the inner diameter of the pipeline of the natural gas system, R represents the gas constant, T represents the average temperature of the gas in the pipeline, Z represents the average compression coefficient of natural gas, ρ 0 represents the natural gas density under standard conditions, F represents the friction factor, and time t is divided into NT time points according to the time period, where Each node in the natural gas system must maintain mass flow. Node K means that the natural gas inflow is equal to the outflow. They represent the sum of all natural gas flows in and out of the pipeline at node K, ∑m k-source and ∑m k-load loads represent the natural gas source and natural gas load at node K, respectively. and They represent the flow rate of the booster station and the natural gas flow rate consumed by the booster station, τc represents the mass flow rate of gas consumed by the compressor, S c represents the flow direction of compressor c relative to node Q, which is at the end of compressor S c =1, and when K is at the compressor head, S c =-1, the node mass flow balance equation is as follows:
[0126]
[0127] Step 2: Combined with the electricity-heat cogeneration model, establish an electricity-heat-gas system model with the following formula:
[0128] h h =f h (h i ,T s ,T r )
[0129] h e =f e (Vi ,θ j )
[0130]
[0131] g(h i ,T s ,T r ,V i ,θ j )=0 (17)
[0132] Among them, h h is the true value of the heat network, f h is the heat network flow equation, h e is the real value of the power grid, f is the power flow equation of the power grid, h gas is the true value of the gas grid, f gas is the gas grid power flow equation, g is the coupling element constraint condition, V i is the voltage amplitude in the power grid, θ j is the phase angle, represents the upstream pressure of the natural gas system at time t, represents the pressure downstream of the natural gas system at time t; T s and T r are the node supply temperature and return temperature respectively;
[0133] Based on the electric-heat-gas system model, after fully analyzing the working modes of various energy conversion hubs, the real-time state estimation method based on volumetric Kalman filtering is used to estimate the system state, and the state classification is carried out using the observation data and models of each subsystem. The distributed optimization algorithm is used to deal with the nonlinear characteristic problems of the electric-heat-gas system and to correct, iterate and update the asynchronous information exchange between the electric-heat-gas systems. The coordinated optimization calculation is carried out in the whole system to obtain a comprehensive global state estimation, such as Figure 3 As shown, the specific steps are:
[0134] Step 1: The quality of nonlinear function estimation has a great impact on the performance of the power-heat-gas system. Due to the nonlinear characteristics of power-heat cogeneration, especially the nonlinear characteristics of the power system, the method based on cubature Kalman filtering can improve the state estimation accuracy of the power-heat-gas system. i Generate 2n volume points to capture the mean of the state estimate and the covariance of the two random variables, as follows:
[0135]
[0136]
[0137]
[0138]
[0139] Among them, S k represents the square root matrix at time k, P k is the estimated covariance; is the i-th volume point; ξ i is the basic data point set;[e] i is the i-th column of the identity matrix; w c and w m are the weights of the mean and covariance, respectively;
[0140] Step 2: Time update is performed through a nonlinear function. The following formula can be used to calculate the predicted state after each time update. The formula is as follows:
[0141]
[0142]
[0143]
[0144] Among them, the covariance matrix is represented by Q k-1 express, is the predicted state of a set of volume points;
[0145] Step 3: Using the predicted state, we can also generate a set of Based on the asynchronous distributed dynamic estimation algorithm, the state of the power system, thermal system and natural gas system is estimated at the same time;
[0146] Step 4, after asynchronous distributed state estimation, under normal circumstances, the distributed state execution cycle of the electrical system is 15 minutes, and the execution cycle of the natural gas system is 60 minutes. The dynamic process of the thermal system is much slower than that of the power system, and the state adjustment cannot be performed at the same frequency. Therefore, the three different subsystems should be performed at different frequencies. The present invention proposes an asynchronous parallel distributed solution. In the electric-gas system, the natural gas system can be considered as an independent component in the electrical system model and further decomposed into smaller local areas;
[0147] Step 5: When performing local state estimation of the combined power and heat system, the asynchronous state estimation of the local Kalman filter is performed by
[0148] The combined power and heat system is executed separately. When the asynchronous state estimation of the Kalman filter converges, the power and heat output of the combined power and heat system will be calculated using the measured state estimation value. The output at this time will be converted into an equivalent output in another energy form, and then transmitted to the estimator in the other system as an enhanced measurement result, which can be used for the next local estimation iteration. The global convergence criterion formula of the two systems is as follows;
[0149]
[0150]
[0151] Among them, φ CHP is the heat output, P CHP is the power output, ε e and ε h are the critical values of the difference between the power generation and the calorific value of the combined power-heat generation system under different voltages;
[0152] Through Monte Carlo experiments, the effectiveness of the real-time state estimation method based on cubature Kalman filtering in improving estimation accuracy, reducing computational complexity, and improving system robustness is verified. The specific steps are as follows:
[0153] Due to the randomness of local state estimation, Monte Carlo simulation is used to analyze the effect of state estimation. The Monte Carlo experiment takes the power flow calculation result as the true value, superimposes Gaussian white noise on the true value as the measurement value, and then performs state estimation. The statistical value of the measurement error S M And the estimated error statistic S H The formula is as follows:
[0154]
[0155]
[0156] Where M is the number of tests, N is the number of measurements, and S after joint state estimation is much better than that of individual state estimation. H / S M The error is reduced and the overall error is reduced, which shows that the joint state estimation can improve the average effect of state estimation.
[0157] The present invention carried out 1000 Monte Carlo simulations and obtained a series of statistical results. M and the estimated error statistic S H To evaluate the state estimate, S H / S MIt can be used to evaluate the filtering effect of state estimation. A value less than 1 indicates that the effect of state estimation is better than that of stateless estimation. The real-time state estimation method based on cubature Kalman filtering is better than that of stateless estimation in improving estimation accuracy, reducing computational complexity and improving system robustness. The smaller the value, the better the estimation effect.
[0158] Features of the present invention:
[0159] The present invention proposes a distributed real-time state estimation method for the collaborative aggregation of an integrated energy system of electricity, heat and gas. Based on the Kalman filter algorithm and combined with the characteristics of different sub-networks, the traditional dynamic state estimation method of the power system is extended to the natural gas network and the thermal energy network. A dynamic state estimation method for the natural gas network and the thermal energy network based on Kalman filtering is constructed, the state equations in the state estimation of each network are derived, and the dynamic state of the thermal energy network is estimated based on the characteristics of the thermal energy network. The proposed method provides a solution for the adaptive distributed real-time state estimation of the electricity-heat-gas system, which has certain practical significance.
[0160] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A distributed real-time state estimation method for the coordinated aggregation of electric, thermal and gas integrated energy systems, characterized in that: include: Establish a thermal system network model based on thermodynamics differences within the system framework; Further co-generation of the power system steady-state model with the thermal system model; Combined with the electricity-heat cogeneration model, an electricity-heat-gas system model is established; Based on the electric-heat-gas system model, after fully analyzing the working modes of various energy conversion hubs, a real-time state estimation method based on volumetric Kalman filtering is used to estimate the system state. The state is classified using the observation data and models of each subsystem. A distributed optimization algorithm is used to deal with the nonlinear characteristics of the electric-heat-gas system and to correct, iterate and update the asynchronous information exchange between the electric-heat-gas systems. A collaborative optimization calculation is performed in the whole system to obtain a comprehensive global state estimation. Through Monte Carlo experiments, the effectiveness of the real-time state estimation method based on cubature Kalman filtering in improving estimation accuracy, reducing computational complexity and enhancing system robustness is verified.
2. The distributed real-time state estimation method for the coordinated aggregation of the electric, heat and gas integrated energy system according to claim 1 is characterized by: A differential thermal system network model based on thermodynamics is established within the system framework. The iterative equations of the thermal system need to consider water flow continuity, circuit pressure, thermal power, temperature mixing and heat loss. Thermodynamics is described by a set of differential equations. The specific process is as follows: Step 1: Considering the continuity of water flow, the mass flow rate of water is consistent with the mass flow rate of water entering the bus, which is equal to the mass flow rate of water flowing out of the bus, which is equal to the sum of the water consumption of the bus. The formula is as follows: Am=m q (1) Where A is the network incidence matrix between the busbar and the branch in the thermal system; m is the mass flow vector in each pipeline, m q is the mass flow vector of the heat load injected into each bus; Step 2: Due to the friction inside the pipe, the pressure head of the water flow will drop, and the pressure difference between the busbars will drive the flow of water in turn. Similar to Kirchhoff's voltage law, the head loss around the closed hydraulic circuit must be zero. The circuit pressure calculation formula is as follows: BK fric m*|m|=0 (2) Where B is the loop incident matrix between the closed loop and each pipeline; K fric is the resistance coefficient of each pipeline; m*m| is the absolute value of the matrix; * represents the product; Step 3: The thermal power generated by the heat source can be calculated as follows: φ=C p m q (T s -T r ) (3) Where φ is the thermal power of each bus; C p is the specific heat of water; m q is a single variable; T S and T r are the node supply temperature and return temperature respectively; Step 4: Heat loss calculation. The relationship between the start and end temperatures of the pipeline can be expressed as: Among them, T end represents the pipe end temperature, T start is the temperature at the beginning of the pipe, T a is the ambient temperature, m is a single variable, λ is the heat transfer coefficient per unit length, and L is the pipe length; Step 5. The water temperature flowing out of the pipe can be calculated by the following formula: (Σm out )T out =Σ(m in T in ) (5) Among them, m out and m in are the mass flow rates of water leaving and entering the pipe, T out is the water temperature entering the pipe head, T in is the water temperature entering the end of the pipe; Step 6. Heat energy is slowly transferred through the hot water in the pipe. The change of the pipe inlet temperature will affect the outlet temperature after the drainage time. Considering that the thermal system operates under constant flow and variable temperature, the hydrodynamics can be ignored. Based on the thermodynamics, the focus is mainly on heat transfer delay and heat loss. Based on the law of conservation of heat, the partial differential equation can be obtained. The formula is as follows: T represents the temperature at any point, S represents the cross-sectional area, t and x represent the start time and the length of the pipeline, C p is the specific heat of water, λ is the heat transfer coefficient per unit length, m is the mass flow rate of water in the pipe, T a is the ambient temperature; Step 7: Discretize the partial differential equation based on the Lax-Wendroff method. The proposed difference-based model only requires the last historical state of the node supply and returns its characteristic information, where k represents the instantaneous time and T i and T j The start time and end time are respectively, a is the ambient temperature, and the formula is as follows:
3. The distributed real-time state estimation method for the coordinated aggregation of the electric, heat and gas integrated energy system according to claim 1 is characterized by: The thermal system model is further combined with the power system steady-state model. The specific process is as follows: Step 1: Establish the power system model, P i and Q i is the active power and reactive power, V i is the voltage amplitude, g ij is the conductivity, b ij is the magnetic susceptibility, θ ij is the phase angle difference; Step 2: Further combine the thermal system model with the power system steady-state model, where φ CHP is the heat output, P CHP is the power output, C m is the heat rate ratio, η S is the ratio of power-heat cogeneration output power, P cap is the maximum output power with zero heat output, η e is the efficiency of the combined heat and power generation unit, and the specific formula is as follows: P cap =the e G in (11) 4. The distributed real-time state estimation method for the coordinated aggregation of the electric, heat and gas integrated energy system according to claim 1 is characterized by: Establish a natural gas system model, combine it with the power-heat cogeneration model, and establish an electricity-heat-gas system model. The specific process is as follows: Step 1: Establish a natural gas model. The transient characteristics of the natural gas system are mainly reflected in the pipeline flow. The following energy conservation law ignoring gravity, inertia and transient pressure can be obtained. The formula is as follows: Where NP is a set of EPS measurements, Π l,t Represents the pressure of the pipeline length l at time t, m l,t represents the flow rate of the pipe length l at time t, D represents the inner diameter of the pipeline of the natural gas system, R represents the gas constant, T represents the average temperature of the gas in the pipeline, Z represents the average compression coefficient of natural gas, ρ0 represents the density of natural gas under standard conditions, F represents the friction factor, and time t is divided into NT time points according to the time period, where Each node in the natural gas system must maintain mass flow. Node K means that the natural gas inflow is equal to the outflow. They represent the sum of all natural gas flows in and out of the pipeline at node K, Σm k-source and Σm k-load loads represent the natural gas source and natural gas load at node K, respectively. and They represent the flow rate of the booster station and the natural gas flow rate consumed by the booster station, τc represents the mass flow rate of gas consumed by the compressor, S c represents the flow direction of compressor c relative to node Q, which is at the end of compressor S c =1, and when K is at the compressor head, S c =-1, the node mass flow balance equation is as follows: Step 2: Combined with the electricity-heat cogeneration model, establish an electricity-heat-gas system model with the following formula: h h =f h (h i ,T s ,T r ) h e =f e (V i ,θ j ) g(h i ,T s ,T r ,V i ,θ j )=0 (17) Among them, h h is the true value of the heat network, f h is the heat network flow equation, h e is the real value of the power grid, f is the power flow equation of the power grid, h gas is the true value of the gas grid, f gas is the gas grid power flow equation, g is the coupling element constraint condition, V i is the voltage amplitude in the power grid, θ j is the phase angle, represents the upstream pressure of the natural gas system at time t, represents the pressure downstream of the natural gas system at time t; T s and T r are the node supply temperature and return temperature, respectively.
5. The distributed real-time state estimation method for the coordinated aggregation of the electric, heat and gas integrated energy system according to claim 1 is characterized by: Based on the electric-heat-gas system model, after fully analyzing the working modes of various energy conversion hubs, a real-time state estimation method based on volumetric Kalman filtering is used to estimate the system state. The state classification is performed using the observation data and models of each subsystem. A distributed optimization algorithm is used to deal with the nonlinear characteristics of the electric-heat-gas system and to correct, iterate and update the asynchronous information exchange between the electric-heat-gas systems. A collaborative optimization calculation is performed in the whole system to obtain a comprehensive global state estimation. The specific steps are as follows: Step 1: The quality of nonlinear function estimation has a great impact on the performance of the power-heat-gas system. Due to the nonlinear characteristics of power-heat cogeneration, especially the nonlinear characteristics of the power system, the method based on cubature Kalman filtering can improve the state estimation accuracy of the power-heat-gas system. i Generate 2n volume points to capture the mean of the state estimate and the covariance of the two random variables, as follows: Among them, S k represents the square root matrix at time k, P k is the estimated covariance; is the i-th volume point; ξ i is the basic data point set; [e] i is the i-th column of the identity matrix; w c and w m are the weights of the mean and covariance, respectively; Step 2: Time update is performed through a nonlinear function. The following formula can be used to calculate the predicted state after each time update. The formula is as follows: Among them, the covariance matrix is represented by Q k-1 express, is the predicted state of a set of volume points; Step 3: Using the predicted state, we can also generate a set of Based on the asynchronous distributed dynamic estimation algorithm, the state of the power system, thermal system and natural gas system is estimated at the same time; Step 4, after asynchronous distributed state estimation, under normal circumstances, the distributed state execution cycle of the electrical system is 15 minutes, and the execution cycle of the natural gas system is 60 minutes. The dynamic process of the thermal system is much slower than that of the power system, and the state adjustment cannot be performed at the same frequency. Therefore, the three different subsystems should be performed at different frequencies. However, in the electric-gas system, the natural gas system can be considered as an independent component in the electrical system model. The following only considers the state estimation of the power-heat cogeneration system; Step 5: When performing local state estimation of the combined power and heat system, the asynchronous state estimation of the local Kalman filter is performed separately in the combined power and heat system. When the asynchronous state estimation of the Kalman filter converges, the power and heat output of the combined power and heat system will be calculated using the measured state estimation value. The output at this time will be converted into an equivalent output in another energy form, and then transmitted to the estimator in the other system as an enhanced measurement result, which can be used for the next local estimation iteration. The global convergence criterion formula of the two systems is as follows: Among them, φ CHP is the heat output, P CHP is the power output, ε e and εh are the critical values of the difference between the power generation and the calorific value of the combined power-heat system under different voltages.
6. The distributed real-time state estimation method for the coordinated aggregation of the electric, heat and gas integrated energy system according to claim 1 is characterized by: Through Monte Carlo experiments, the effectiveness of the real-time state estimation method based on cubature Kalman filtering in improving estimation accuracy, reducing computational complexity, and improving system robustness is verified. The specific steps are as follows: Due to the randomness of local state estimation, Monte Carlo simulation is used to analyze the effect of state estimation. The Monte Carlo experiment takes the power flow calculation result as the true value, superimposes Gaussian white noise on the true value as the measurement value, and then performs state estimation. The statistical value of the measurement error S M And the estimated error statistic S H The formula is as follows: Where M is the number of tests, N is the number of measurements, and S after joint state estimation is much better than that of individual state estimation. H / S M The error is reduced and the overall error is reduced, which shows that joint state estimation can improve the average effect of state estimation; After 1000 Monte Carlo simulations, a series of statistical results were obtained. M and the estimated error statistic S H To evaluate the state estimate, S H / S M It can be used to evaluate the filtering effect of state estimation. Its value is less than 1, which means that the effect of state estimation is better than that of stateless estimation. The real-time state estimation method based on cubature Kalman filtering is better than that of stateless estimation in improving estimation accuracy, reducing computational complexity and improving system robustness. H / S M The smaller the value, the better the estimated effect.
Citation Information
Cited By
Data-driven nonlinear system state estimation method
CN119596689A
Intelligent edge control method for information energy system based on self-adaptive dynamic programming
CN119960315A
Transformer current rising test data processing method fused with capacitance compensation
CN121562454A
