A method for adjusting load in response to live conditions in combination with virtual power plant

Abstract

The application provides a method for adaptive adjustment of load in response to live conditions in combination with a virtual power plant, comprising: real-time monitoring of current steam extraction amount of a combined heat and power unit and feeder impedance, analyzing a mapping relationship between the steam extraction amount and the feeder impedance, and extracting interactive process characteristics of the steam extraction amount and the feeder impedance according to the mapping relationship; analyzing a dynamic response degree of the feeder impedance when the steam extraction amount changes according to the interactive process characteristics, and determining an impedance change trend according to the dynamic response degree; if a peak value of the impedance change trend exceeds a preset peak threshold value, obtaining unit output and steam pressure of the current combined heat and power unit, and identifying an interactive amplitude of the steam extraction amount and the feeder impedance; adjusting a load instruction parameter of the combined heat and power unit according to the interactive amplitude, and generating an adaptive load instruction sequence; and processing virtual power plant operation scene data through a simulation verification method according to the optimized load instruction of the combined heat and power unit, to obtain safe and efficient operation indexes.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cogeneration unit load regulation, and particularly relates to a method for adaptively adjusting load in combination with virtual power plant response live. BACKGROUND

[0002] Under the background of rapid development of energy internet, as the core technical architecture of new power system, virtual power plant undertakes the important mission of overall planning and coordinating distributed energy resources and realizing power supply and demand balance optimization. With the continuous deepening of the application of virtual power plant technology, cogeneration units as an important component unit face increasingly complex technical problems in the process of dynamic load adjustment and operation. Especially under the double technical constraints of meeting the accurate control of steam extraction amount of the thermal system and maintaining the stability of the feeder impedance characteristics of the power system, the real-time dynamic response performance of the unit system has become a key technical bottleneck restricting the optimization of the overall operation efficiency of the virtual power plant. Taking a typical operating scenario as an example, when the system dispatching center issues a load adjustment instruction, the fluctuation of the feeder system impedance characteristic parameters caused by the significant change of the steam extraction amount of the cogeneration unit may cause the actual operating condition of the unit to deviate seriously from the preset target state, ultimately resulting in a significant reduction in the reliability and economy of the overall dispatching execution scheme of the virtual power plant.

[0003] At present, the technical scheme of load regulation of cogeneration units widely used in the industry has obvious technical limitations in actual engineering application. The existing control strategy often uses a single energy output oriented static optimization method, or only analyzes and processes the steady-state characteristics under a specific operating condition, seriously ignoring the strong coupling dynamic correlation mechanism between the steam extraction amount adjustment process and the change of the feeder impedance characteristics. The steam extraction amount, as the core control parameter of the thermal output regulation of the cogeneration unit, its value change not only directly determines the thermal supply capacity of the unit, but also has a significant impact on the impedance characteristic parameters of the feeder system. In turn, the dynamic fluctuation of the feeder impedance characteristics will form a reverse constraint on the power output regulation capacity of the unit. Specifically, when the virtual power plant dispatching system issues an incremental regulation instruction of the steam extraction amount of the unit based on the overall optimization target, the rising effect of the feeder system impedance parameters caused by the regulation action will directly weaken the actual power output regulation margin of the unit, and further cause a significant deviation between the actual operating response of the unit and the expected dispatching target, seriously affecting the accurate execution effect of the dispatching plan. SUMMARY

[0004] The present application provides a method for adaptively adjusting load in combination with virtual power plant response live, mainly comprising:

[0005] The steam extraction amount of a combined heat and power unit and the feeder impedance are monitored in real time, the mapping relationship between the steam extraction amount change and the feeder impedance change is analyzed, and the interactive process characteristics of the steam extraction amount and the feeder impedance are extracted; the dynamic response degree of the feeder impedance when the steam extraction amount changes is analyzed according to the interactive process characteristics, the impedance change trend is determined, the unit output and the steam pressure of the combined heat and power unit are obtained when the impedance change trend peak value exceeds a preset peak threshold value, the interactive amplitude of the steam extraction amount and the feeder impedance is identified, the load instruction parameters of the combined heat and power unit are adjusted according to the interactive amplitude, the adaptive load instruction sequence is generated, the adaptive load instruction sequence is input into a virtual power plant control system, actual adjustment feedback data of the combined heat and power unit are obtained, the actual adjustment feedback data and the adaptive load instruction sequence are compared, and the limiting factor of the combined heat and power unit adjustment is determined; the flow change and the pressure fluctuation of the steam extraction amount are obtained, the load instruction of the combined heat and power unit is optimized, and the optimized load instruction is generated; the virtual power plant operation scene data are processed according to the optimized load instruction, and the operation index is obtained.

[0006] Further, the real-time monitoring of the steam extraction amount of the combined heat and power unit and the feeder impedance, the analysis of the mapping relationship between the steam extraction amount change and the feeder impedance change, and the extraction of the interactive process characteristics of the steam extraction amount and the feeder impedance include:

[0007] The steam extraction amount measurement value of the combined heat and power unit and the real-time data of the feeder impedance are obtained, the difference value of the steam extraction amount measurement values at adjacent time points is calculated to obtain the steam flow change rate, the sliding window statistics of the feeder impedance real-time data is performed, the difference value of the maximum value and the minimum value in the preset time window is calculated to obtain the impedance fluctuation amplitude, the time sequence mapping relationship is established according to the steam flow change rate and the impedance fluctuation amplitude, the correlation value of the steam flow change rate sequence and the impedance fluctuation amplitude sequence is calculated, when the correlation value is greater than a preset threshold value, the thermal load value and the electric load value of the combined heat and power unit are extracted, the sum of the thermal load value and the electric load value is calculated as a total load value, the thermal load value is divided by the total load value to obtain the thermal load distribution ratio, and the steam flow change rate, the thermal load distribution ratio and the impedance fluctuation amplitude are combined to form a data combination representing the interactive process characteristics of the steam extraction amount and the feeder impedance.

[0008] Further, the analysis of the dynamic response degree of the feeder impedance when the steam extraction amount changes according to the interactive process characteristics and the determination of the impedance change trend include:

[0009] According to the steam flow rate of change, the heat and power load distribution ratio and the impedance fluctuation amplitude in the interactive process characteristics, the contribution weight of each characteristic component to the impedance change is calculated to generate a weight coefficient vector; the weight coefficient vector and the interactive process characteristics are subjected to dot product operation to obtain a dynamic response degree value; according to the dynamic response degree value, the sensitivity level of the feeder impedance to the steam extraction amount change is judged; when the dynamic response degree value exceeds a preset sensitivity threshold, the impedance change sequence under similar working conditions in the historical operation data is extracted; the impedance change sequence is subjected to time series decomposition to separate the trend component, and a fitting method is used to determine the impedance change trend.

[0010] Further, when the impedance change trend peak value exceeds the preset peak threshold value, the unit output and the steam pressure of the combined heat and power unit are obtained, and the interactive amplitude of the steam extraction amount and the feeder impedance is identified, including:

[0011] When the impedance change trend peak value exceeds the preset peak threshold value, the unit output data and the steam pressure data of the combined heat and power unit are obtained, the steam turbine admission valve opening signal is read, the steam flow rate is determined according to the valve flow rate characteristic curve; based on the steam flow rate and the unit output data, a time sequence corresponding data set is constructed, and the correlation coefficient of the unit output data and the impedance change trend value is calculated; the steam pressure data is normalized to construct a mapping relationship matrix, extract the principal components of the matrix, and obtain the characteristic value representing the coupling strength; according to the characteristic value and the steam extraction amount, the ratio of the steam extraction amount change to the impedance change amount is calculated to obtain the interactive amplitude.

[0012] Further, the interactive amplitude is used to adjust the load instruction parameters of the combined heat and power unit to generate an adaptive load instruction sequence, including:

[0013] According to the interactive amplitude, the steam flow rate upper and lower limit values and the thermal power regulation range are read from the control station of the combined heat and power unit, the adjustment records under the same load level in the historical database are queried, the adjustment amount of the steam flow rate set point and the change value of the thermal power distribution ratio are extracted; the adjustment step of the steam flow rate set point is determined according to the adjustment amount, and the new thermal power distribution ratio is calculated; the adjustment step and the thermal power distribution ratio are used to construct a time sequence instruction data set to generate a smooth changing instruction sequence; the instruction sequence is sampled according to the control period, and the execution time is added to generate the adaptive load instruction sequence.

[0014] Further, the adaptive load instruction sequence is input into the virtual power plant control system to obtain actual adjustment feedback data of the combined heat and power unit, the actual adjustment feedback data and the adaptive load instruction sequence are compared, and the limiting factor of the combined heat and power unit adjustment is determined, including:

[0015] The adaptive load instruction sequence is transmitted to a control execution module through a communication protocol, actual steam extraction amount, actual feeder impedance value and thermal power load distribution real-time data of a combined heat and power unit are collected, deviation of the actual steam extraction amount and the adaptive load instruction sequence is calculated, when the deviation exceeds a set value, the steam pressure data is read, whether the steam pressure reaches a limit boundary is judged, and the steam pipeline pressure limit is determined as the limiting factor, if the steam pressure is in a normal range, a thermal power load actual adjustment rate is calculated, compared with a design rate, and the thermal power load adjustment rate limit is determined as the limiting factor.

[0016] Further, the flow variation and pressure fluctuation of the steam extraction amount are acquired, and the combined heat and power unit load instruction is optimized to generate an optimized load instruction.

[0017] Real-time flow data of the steam extraction amount and the steam pressure data are acquired from a steam flow meter and a pressure sensor, a variation rate of the flow data and a fluctuation amplitude of the pressure data are calculated, time correlation of the flow variation rate and the limiting factor occurrence time and time correlation of the pressure fluctuation amplitude and the limiting factor occurrence time are calculated respectively, and an association strength coefficient is recorded, an instruction execution deviation of the combined heat and power unit actual adjustment feedback data is extracted according to the association strength coefficient, a deviation trend is calculated, the adaptive load instruction sequence is corrected according to the deviation trend, a variation rate of a steam flow set point or an instruction point time interval is adjusted, and the optimized load instruction is generated.

[0018] Further, the virtual power plant operation scene data is processed according to the optimized load instruction to obtain an operation index, including:

[0019] The virtual power plant operation scene simulation environment is constructed according to the optimized load instruction, load demand prediction data and power grid impedance distribution data are extracted, different load levels and impedance states are combined, the optimized load instruction is executed, a thermal power unit output power time sequence is recorded, power fluctuation amplitude and frequency spectrum characteristics are calculated, and unit operation stability is judged, thermal power unit thermal parameters are extracted, thermal power and electric power are calculated, and total efficiency is obtained, and a plurality of working condition efficiency values are weighted and averaged to obtain thermal power conversion efficiency and unit operation stability.

[0020] The technical scheme provided by the embodiment of the application can include the following beneficial effects:

[0021] The application discloses a method for adaptively adjusting load in combination with virtual power plant response live, and solves the problems of uneven thermal power load distribution and insufficient operation stability in the business scenario under the interactive influence of steam extraction amount and feeder impedance. BRIEF DESCRIPTION OF DRAWINGS

[0022] Fig. 1 A flow chart of the method for adaptively adjusting load in combination with virtual power plant response live.

[0023] Fig. 2 A schematic diagram of the method for adaptively adjusting load in combination with virtual power plant response live. DETAILED DESCRIPTION

[0024] In order to further understand the content of the application, the application will be described in detail in conjunction with the drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the related application, and not to limit the application. In addition, in order to facilitate the description, only the parts related to the application are shown in the drawings.

[0025] As Figs. 1-2 The method for adaptively adjusting load in combination with virtual power plant response live can specifically include:

[0026] In step S101, the current steam extraction amount and feeder impedance of the combined heat and power unit are monitored in real time, the mapping relationship between the steam extraction amount change and the feeder impedance change is analyzed, and the interactive process characteristics of the steam extraction amount and the feeder impedance are extracted according to the mapping relationship.

[0027] The steam extraction amount measurement value and the feeder impedance real-time data of the combined heat and power unit at the current time are acquired, the steam extraction amount measurement values at adjacent time points are processed by using a difference calculation method to obtain a steam flow rate change rate, and the feeder impedance real-time data is statistically processed by using a sliding window to obtain a difference between a maximum value and a minimum value of the impedance values in a preset time window, thereby obtaining an impedance fluctuation amplitude. According to the steam flow rate change rate and the impedance fluctuation amplitude, a time sequence mapping relationship therebetween is established, a correlation value between the steam flow rate change rate sequence and the impedance fluctuation amplitude sequence is calculated by using a Pearson correlation coefficient, and if the correlation value is greater than a preset threshold value, it is determined that a strong coupling relationship exists. Based on the strong coupling relationship, a current thermal load value and an electric load value are extracted from the operation data of the combined heat and power unit, a sum of the thermal load value and the electric load value is calculated as a total load value, a thermal-electric load distribution ratio is obtained by dividing the thermal load value by the total load value, and the steam flow rate change rate, the thermal-electric load distribution ratio and the impedance fluctuation amplitude are combined to form a data combination representing a characteristic of an interaction process between the steam extraction amount and the feeder impedance.

[0028] In an embodiment, the steam extraction amount monitoring of the combined heat and power unit is realized by a vortex flowmeter installed at a steam extraction port of a steam turbine. The flowmeter collects 10 flow data per second to form a continuous time sequence. The feeder impedance real-time data are acquired by a synchronous phasor measurement device of a power distribution network, and the measurement frequency is consistent with that of the flowmeter. The difference calculation method specifically adopts a first-order difference, that is, a steam extraction amount measurement value at a current time is subtracted from a measurement value at a previous time, and then divided by a time interval to obtain an instantaneous steam flow rate change rate.

[0029] The time window for the sliding window statistics is set to 5 minutes, and 300 impedance sampling points are continuously recorded in the window. The calculation process of the impedance fluctuation amplitude is to find the maximum value and the minimum value of all sampling values in the window, and the difference between the two values is the impedance fluctuation amplitude of the time period. This processing method can effectively reflect the change characteristics of the feeder impedance in a short time.

[0030] Specifically, the calculation of the Pearson correlation coefficient needs to first standardize the steam flow rate change rate sequence and the impedance fluctuation amplitude sequence. The covariance of the two sequences is divided by the product of the standard deviations to obtain the correlation coefficient value. The preset threshold value is usually set to 0.7. When the calculated correlation coefficient exceeds the threshold value, it indicates that there is a significant linear correlation between the steam extraction amount change and the feeder impedance change, that is, a strong coupling relationship.

[0031] In a possible implementation, the acquisition of the thermal-electric load distribution ratio involves real-time operation parameters of the unit. The thermal load value is calculated by measuring the flow rate and enthalpy value of the heating steam, and the electric load value is directly read from a power meter at the outlet of the generator.

[0032] For example, when the thermal load is 120 MW and the electric load is 80 MW at a certain moment, the total load is 200 MW, and the thermal-electric load distribution ratio is 0.6.

[0033] Preferably, the data combination of the interactive process characteristics is stored in the form of a three-dimensional vector, wherein the first dimension is the normalized steam flow rate change rate, the second dimension is the thermal-electric load distribution ratio, and the third dimension is the standardized impedance fluctuation amplitude.

[0034] In step S102, the dynamic response degree of the feeder impedance to the change in the steam extraction amount is analyzed according to the interactive process characteristics, and the impedance change trend is determined according to the dynamic response degree.

[0035] According to the steam flow rate change rate, the thermal-electric load distribution ratio, and the impedance fluctuation amplitude in the interactive process characteristics, the contribution weight of each characteristic component to the impedance change is calculated, the weight coefficient vector is obtained through normalization processing, the dot product operation of the weight coefficient vector and the interactive process characteristics is performed, and the dynamic response degree value is obtained. Based on the dynamic response degree value, the sensitivity level of the feeder impedance to the change in the steam extraction amount is judged. If the dynamic response degree value exceeds the preset sensitivity threshold, the impedance change sequence under similar working conditions is extracted from the historical operation data. The trend component and the periodic component are separated by performing time series decomposition on the impedance change sequence, the trend component is fitted by using the least square method, and the impedance change trend is determined.

[0036] In an embodiment, the calculation of the weight coefficient determines the relative importance of each characteristic component by using the analytic hierarchy process.

[0037] Specifically, the weight of the steam flow rate change rate is usually set to 0.5, the weight of the thermal-electric load distribution ratio is 0.3, and the weight of the impedance fluctuation amplitude is 0.2. This weight distribution reflects that the direct influence of the steam flow rate change on the feeder impedance is the largest. The normalization processing ensures that the sum of the weight coefficients is 1, forming a standard weight coefficient vector.

[0038] The specific implementation of the dot product operation is to multiply each component of the weight coefficient vector with the corresponding component of the interaction process feature, and then sum up to obtain a single dynamic response degree value, which ranges from 0 to 1, and the larger the value, the more intense the response of the feeder impedance to the change in steam extraction. The preset sensitivity threshold is generally set to 0.6, and when the dynamic response degree value exceeds this threshold, it indicates that it is in a high sensitivity state. Similar working conditions can be extracted from historical operation data. By comparing the current unit load level, environmental temperature and steam parameters, search for historical working conditions with a matching degree of more than 85% in the historical database, and extract the recorded impedance change sequence under these working conditions as reference data. The impedance change sequence is decomposed into a trend component reflecting the long-term change direction of the impedance, a seasonal component embodying the periodic fluctuation characteristics and a random component containing irregular disturbances by using the additive decomposition method. The decomposition process first extracts the trend component by moving average method, and the window length is selected as an integer multiple of the data sampling period. Then, the trend component is subtracted from the original sequence to obtain a detrended sequence, and the seasonal component is extracted by period average method.

[0039] In one possible implementation, when the least squares method fits the trend component, a linear or quadratic polynomial is selected as the fitting function. By minimizing the sum of the squares of the vertical distances between the actual data points and the fitted curve, the polynomial coefficients are solved. The fitted trend equation can predict the direction and amplitude of the impedance change in the future period, providing a basis for subsequent load adjustment decisions.

[0040] In step S103, if the impedance change trend peak value exceeds the preset peak threshold value, the unit output and steam pressure of the current combined heat and power unit are obtained, and the interaction amplitude of the steam extraction and the feeder impedance is identified.

[0041] If the impedance change trend peak value exceeds the preset peak threshold value, the current unit output data and the main steam pipe pressure data are obtained from the distributed control station of the combined heat and power unit, the steam turbine admission regulating valve opening signal is read through the data acquisition card, and the current steam flow rate is determined according to the pre-calibrated regulating valve flow rate characteristic curve. Based on the steam flow rate and the unit output data, a time sequence corresponding data set is established, the dimensionless correlation coefficient of the unit output value at each time and the impedance change trend value is calculated by using the grey correlation degree method, the average value of the correlation coefficient sequence is taken as the comprehensive correlation degree, and the main steam pipe pressure data is normalized. According to the comprehensive correlation degree and the normalized pressure data, a mapping relationship matrix is constructed, the rows of the matrix correspond to the unit output values at different times, the columns correspond to the pressure values, and the matrix elements are the corresponding impedance change values. The first principal component of the matrix is extracted by principal component analysis to obtain a characteristic value representing the coupling strength. The characteristic value and the current steam extraction are used to calculate the ratio of the extraction change amount to the impedance change amount by using the difference method, and the interaction amplitude of the steam extraction and the feeder impedance is obtained.

[0042] In an embodiment, the determination of the peak of the impedance change trend adopts a sliding window peak detection method. In a continuous impedance change data sequence, a sliding window with a time length of 10 minutes is set, and the maximum value in the window is calculated as a local peak. A preset peak threshold is determined according to a percentage of the rated impedance of the feeder, and is usually set to 1.3 times of the rated value. When a peak exceeding the threshold is detected, a data acquisition program is triggered to extract unit output and steam pressure parameters from a real-time database of the distributed control station.

[0043] The pre-calibration of the flow characteristic curve of the regulating valve is obtained through actual flow test. In the unit debugging stage, the opening degree of the regulating valve is gradually changed, and the corresponding steam flow is measured by using an ultrasonic flowmeter, and the opening degree and flow data pairs are recorded. A polynomial fitting method is adopted to establish the functional relationship between the opening degree and the flow, and a cubic polynomial is used for fitting. The fitted polynomial can be:

[0044]

[0045] Q represents the flow of the regulating valve, θ represents the opening degree of the regulating valve, a0 represents the zero-order coefficient, a1 represents the first-order coefficient, a2 represents the second-order coefficient, and a3 represents the third-order coefficient. The formula establishes a cubic polynomial relationship between the opening degree and the flow, and can describe the nonlinear characteristics of the regulating valve. The calibrated characteristic curve is stored in the database of the control station, and the corresponding flow value is queried according to the current opening degree value in real-time operation.

[0046] Specifically, the calculation process of the grey correlation degree includes two stages of data preprocessing and correlation coefficient calculation. In the data preprocessing stage, the unit output sequence and the impedance change trend sequence are dimensionless processed, and the initial value method is adopted, that is, the first data value of each sequence is removed from the entire sequence to eliminate the influence of dimension. When calculating the correlation coefficient, first, the absolute difference values corresponding to the time of the two sequences are calculated, and the minimum and maximum values of all the difference values are found. Then, according to the grey correlation coefficient formula, the correlation coefficient at each time is calculated, wherein the resolution coefficient is usually taken as 0.5. Finally, the correlation coefficients at all times are averaged to obtain the comprehensive correlation degree value. The comprehensive correlation degree reflects the overall correlation degree between the unit output and the impedance change, and the closer the value is to 1, the stronger the correlation is.

[0047] The main steam pipeline pressure data is normalized by using the maximum and minimum value standardization method. The maximum and minimum value range of the pressure is obtained from the historical operation data, and the current pressure value is mapped to the interval of 0 to 1. This processing method eliminates the influence of the absolute value of the pressure, so that it is compared with other parameters in the same order of magnitude.

[0048] ​In a possible implementation, the mapping relationship matrix is constructed based on time-aligned data sequences. Each row of the matrix represents the unit output value at a certain time, each column represents the corresponding pressure value, and the matrix element stores the impedance change value under the working condition. Since the unit output and pressure are continuously changing in actual operation, the continuous data needs to be discretized. The unit output is divided into intervals of 10 MW, and the pressure is divided into intervals of 0.5 MPa to form discrete working condition points. For data falling within a certain working condition interval, the average value of the impedance change value is taken as the value of the matrix element. This discretization process not only retains the main mapping relationship characteristics but also reduces the complexity of data processing.

[0049] Exemplarily, the specific implementation process of principal component analysis in extracting the mapping matrix features is as follows. First, the mapping relationship matrix is subjected to centering processing, the mean value of each column is calculated, and the mean value is subtracted from the original data. Then, the covariance matrix is calculated, and the eigenvalues and eigenvectors are obtained through eigenvalue decomposition. The eigenvalues are sorted in descending order, and the first principal component corresponds to the eigenvector of the largest eigenvalue. The modulus of the eigenvector is the eigenvalue representing the coupling strength, which reflects the comprehensive influence of the unit operating parameters on the impedance change. Further, the implementation details of the differential method for calculating the interaction amplitude involve a small perturbation analysis. In the vicinity of the current operating point, a small steam extraction amount change is artificially given, usually 5% of the current value, and the impedance change caused thereby is observed. The ratio of the two is the local interaction amplitude coefficient. To improve the calculation accuracy, perturbation tests are performed in both positive and negative directions, and the average value of the two calculation results is taken as the interaction amplitude.

[0050] It can be understood that the physical meaning of the interaction amplitude is the impedance change caused by a unit change in the steam extraction amount, and its numerical value directly affects the load adjustment strategy of the unit. When the interaction amplitude is large, it indicates that it is in a high sensitivity state, and a small adjustment in the steam extraction amount will cause significant impedance fluctuations, and therefore a more cautious control strategy needs to be adopted.

[0051] For example, in the actual application of a certain 300 MW cogeneration unit, when the impedance peak value reaches 1.35 times the rated value, the system automatically obtains that the current unit output is 250 MW and the main steam pressure is 16.5 MPa. The interaction amplitude calculated by the above method is 0.8 Ω / (t / h), indicating that for every 1 ton / hour increase in the steam extraction amount, the feeder impedance will increase by 0.8 ohms.

[0052] Step S104, adjusting the cogeneration unit load instruction parameter according to the interaction amplitude to generate an adaptive load instruction sequence.

[0053] According to the interaction amplitude value, the upper and lower limits of steam flow and the thermal power regulation range under the current working condition are read from the cogeneration unit control station, the adjustment records under the same load level in the historical database are queried, and the single adjustment amount of steam flow and the power distribution proportion change value in the records are extracted. Based on the single adjustment amount, the adjustment step of the current steam flow set point is determined, if the interaction amplitude is greater than the preset sensitivity threshold, half of the adjustment amount is used as the actual step, otherwise the adjustment amount is directly used, and according to the power distribution proportion change value and the current thermal power load demand, a new thermal power distribution proportion is calculated. The adjustment step and the new thermal power distribution proportion are used to construct the instruction data set in the form of time series, and the transition value is generated between adjacent instruction points by linear interpolation method, forming a smooth change instruction sequence. According to the instruction sequence, the flow set value and the power distribution value of each sampling point are sampled at equal intervals according to the control period, and the corresponding execution time is attached to generate an adaptive load instruction sequence.

[0054] In an embodiment, the application of the interaction amplitude value involves a multi-level parameter adjustment mechanism. The real-time database of the control station stores the operating boundary conditions of the unit under different working conditions, including the physical limit of steam flow and the regulation ability of thermal power. The upper limit of steam flow is usually determined by the through-flow capacity of the steam turbine, and the lower limit is limited by the minimum flow requirement for stable operation of the unit. The thermal power regulation range reflects the flexibility margin of the unit under the current working condition, which is dynamically adjusted with the changes of environmental temperature, condenser vacuum and circulating water temperature.

[0055] The adjustment records in the historical database are stored according to load level. Each record contains operating parameters before and after adjustment, time series data of adjustment process and adjustment effect evaluation index. When querying, according to the closeness of the current load and the historical load, the historical records with load deviation within ±5% are selected. The single adjustment amount extracted reflects the successful adjustment experience under similar working conditions, and the power distribution proportion change value reflects the coordination relationship between thermal and electric loads. The preset sensitivity threshold is set according to the unit capacity and the stability requirement of the power grid. For a 300MW unit, the threshold is usually set to 0.5.

[0056] When the interaction amplitude exceeds this threshold, it indicates that it is in a high sensitivity state, at this time the conservative adjustment strategy is adopted, and the historical adjustment amount is halved as the actual step to avoid excessive adjustment amplitude causing system oscillation. On the contrary, when the interaction amplitude is lower than the threshold, the response is relatively gentle, and the historical adjustment amount can be directly used to speed up the adjustment speed. The new thermal power allocation ratio is calculated combined with the current heat load demand and the electric load demand. The heat load demand comes from the real-time monitoring data of the heat supply pipe network including heat supply flow, temperature and pressure parameters. The load instruction issued by the power grid dispatching center determines the electric load demand. In the calculation process, the minimum extraction steam quantity required to meet the heat load demand is determined first, and then the electric power output is optimized in the remaining adjustment range. The allocation ratio is adjusted in accordance with the thermal power decoupling principle, that is, the power generation efficiency is maximized under the premise of meeting the heat supply demand.

[0057] In one possible implementation, each data point of the instruction data set in the form of time series contains three elements, execution time, steam flow set value and power allocation ratio. The execution time is determined according to the scanning period of the control system, and the typical value is 1 second. The time span between the initial instruction point and the target instruction point is determined according to the adjustment amplitude, and a large amplitude adjustment needs a longer transition time to ensure system stability.

[0058] The specific implementation of the linear interpolation method is as follows: let the flow set value at the initial time be F0, the flow set value at the target time be F1, and the transition time be T, then the interpolation formula at any intermediate time t is F(t) = F0 + (F1-F0) x t / T. For the power allocation ratio, the same interpolation method is used to ensure the coordinated change of thermal and electric loads. The density of the interpolation points depends on the control accuracy requirement, and one interpolation point is generated every second to form a continuous instruction curve. Further, the control period of the equidistant sampling implementation is determined by the configuration of the distributed control system, and the common period is 100 milliseconds to 1 second. The sampling process extracts discrete points from the continuous curve generated by interpolation according to the control period, and each sampling point retains complete parameter information. The selection of sampling frequency needs to balance the control accuracy and system burden, and too high sampling frequency will increase the calculation load of the controller, and too low will affect the smoothness of adjustment.

[0059] The additional execution time provides accurate time identification for each instruction point. The time identification adopts absolute time format, including year, month, day, hour, minute, second and millisecond information, which ensures the time sequence relationship of the instruction sequence. This time identification mechanism supports the caching and delay execution of instructions, and when there is a short interruption in network communication, the controller can continue to execute the predetermined instruction sequence according to the time identification.

[0060] For example, in a certain load adjustment process, the interaction amplitude is 0.7, which exceeds the sensitivity threshold of 0.5. The historical record shows that the single adjustment amount under similar working conditions is 20 t / h, so the actual step size is 10 t / h. The initial flow setting value is 150 t / h, the target value is 160 t / h, and the transition time is set to 30 seconds. Through linear interpolation, 30 intermediate instruction points are generated, and then sampled according to a control period of 1 second, finally forming an adaptive load instruction sequence containing 31 instruction points, realizing smooth load adjustment.

[0061] In step S105, the adaptive load instruction sequence is input into the virtual power plant control system, and actual adjustment feedback data of the combined heat and power unit is obtained. By comparing the actual adjustment feedback data with the adaptive load instruction sequence, a comparison result is obtained as a limiting factor of the combined heat and power unit adjustment.

[0062] The adaptive load instruction sequence is transmitted to the control execution module through the virtual power plant control communication protocol. The control execution module sequentially issues adjustment instructions according to the time stamps in the instruction sequence, and simultaneously collects real-time actual steam extraction amount, actual feeder impedance value and combined heat and power load distribution live data from the flow meter, impedance measuring device and power transducer of the combined heat and power unit. The deviation of the actual value at each time from the instruction setting value is calculated. If the steam extraction amount deviation exceeds five percent of the set value and lasts for more than three sampling periods, the steam pipe pressure sensor data is read to determine whether the current pressure reaches the upper limit of the pipe design pressure or is lower than the minimum inlet steam pressure of the steam turbine. If the pressure reaches the limit boundary, the steam pipe pressure limit is determined as the limiting factor. If the pressure is within the normal range, the actual adjustment rate is calculated by dividing the difference between the actual values of the combined heat and power load at adjacent time points by the time interval. The actual adjustment rate is compared with the unit design adjustment rate. If the actual rate is less than seventy percent of the design rate, the combined heat and power load adjustment rate limit is determined as the limiting factor. The determination results are combined to obtain the limiting factor containing the steam pipe pressure limit or the combined heat and power load adjustment rate limit.

[0063] In an embodiment, the virtual power plant control communication protocol adopts the IEC61850 standard or the Modbus TCP / IP protocol to realize reliable transmission of control instructions. The control execution module serves as an intermediate layer, responsible for parsing the instruction sequence and converting it into control signals recognizable by the unit. The instruction sequence is transmitted to the control execution module through Ethernet, and each instruction contains three fields: time stamp, parameter type and parameter value. The control execution module has a built-in clock synchronization mechanism to ensure that the instructions are executed accurately at the predetermined time.

[0064] The flow meter is a vortex shedding meter or a differential pressure flow meter installed on the steam extraction pipeline, with a measurement accuracy of 0.5 level. The impedance measurement device is integrated in the feeder protection device, which obtains electrical quantities through voltage and current transformers and calculates the feeder impedance value in real time. The power transducer monitors the thermal power and electric power, respectively. The thermal power is calculated by the steam flow and enthalpy value, and the electric power is directly measured at the generator outlet. The sampling frequency of these measurement devices is uniformly set to 100 Hz to ensure the time synchronization of the data.

[0065] Specifically, the deviation calculation adopts a combination of absolute deviation and relative deviation. The absolute deviation is the difference between the actual value and the set value, and the relative deviation is the ratio of the absolute deviation to the set value. For steam extraction quantity, when the relative deviation exceeds 5%, it is considered that significant deviation occurs. The sampling period is determined according to the scanning period of the control system, and the typical value is 1 second. The judgment of three consecutive sampling periods avoids the misjudgment of transient disturbance and improves the accuracy of the limiting factor identification. The duration of the deviation is counted from the first time the limit is exceeded until the deviation returns to the allowed range.

[0066] Preferably, the steam pipeline pressure limit is determined by two boundary conditions. The upper limit of the pipeline design pressure is determined by the strength of the pipeline material and the safety factor, which is usually 1.25 times the rated pressure. When the actual pressure approaches this upper limit, continued increase of steam flow will cause pipeline overpressure, threatening equipment safety. The minimum inlet steam pressure of the steam turbine is determined by the design characteristics of the steam turbine. Below this pressure, the efficiency of the steam turbine decreases sharply, and even the blade may vibrate. The pressure sensor uses a diffused silicon pressure transmitter with a range of 0 to 1.5 times the rated pressure, and the measurement accuracy reaches 0.25%.

[0067] In one possible implementation, since the original load data has measurement noise, direct difference calculation will amplify the noise effect. Therefore, the load data is first filtered by moving average, and the window length is selected as 5 sampling points. The filtered data is used to calculate the load change at adjacent time, and the time interval is divided to obtain the instantaneous regulation rate. A plurality of instantaneous rates are statistically averaged to obtain a stable rate value. The design regulation rate of the unit is specified in the technical specification of the unit, which reflects the dynamic response capability of the unit.

[0068] Exemplarily, the comparison of the actual adjustment rate and the design adjustment rate adopts a percentage threshold judgment. The threshold of 70% is selected based on engineering experience and safety margin considerations. When the actual rate is less than 70% of the design rate, it indicates that the adjustment capacity of the unit is limited, and the possible reasons include slow response of the actuator, improper control loop parameter setting, or mechanical component wear. Such rate limitation will cause the unit to fail to respond to the load instruction in time, affecting the frequency regulation and power balance of the power grid. Further, the comprehensive judgment of the limiting factors considers various possible combinations. In some operating conditions, there may be both pressure limitation and rate limitation. By comparing the actual deviation caused by the two limitations, the dominant limiting factor is identified, and the limitation with a larger deviation is identified as the dominant factor. The identification result of the limiting factor is output in a structured data form, including the limitation type, limitation degree, and occurrence time, etc.

[0069] It can be understood that the accurate identification of the limiting factor is of great significance for subsequent control optimization. The pressure limitation usually requires adjusting the operation mode of the steam system, such as changing the extraction port position or adjusting the turbine back pressure. The rate limitation requires optimizing the control parameters or overhauling the actuator. By identifying the limiting factor in real time, the virtual power plant control system can adaptively adjust the control strategy to maximize the adjustment capacity of the unit under the premise of ensuring safety.

[0070] For example, in a certain load adjustment process, the instruction requires the steam extraction amount to increase from 100 t / h to 120 t / h. In the actual execution process, the extraction amount stops increasing at 115 t / h, with a deviation of 5 t / h, which exceeds the set value of 4.2%. At this time, the steam pipe pressure reaches 16.8 MPa, close to the design upper limit of 17 MPa, and the system determines that it is a steam pipe pressure limitation.

[0071] In step S106, the flow change of the steam extraction amount and the pressure fluctuation are obtained simultaneously, and it is analyzed whether the limiting factor is related to the flow change of the steam extraction amount or the pressure fluctuation. If it is related, the cogeneration unit load instruction is optimized according to the actual adjustment feedback data of the cogeneration unit, and the optimized cogeneration unit load instruction is obtained.

[0072] Real-time flow data and pipeline pressure data of steam extraction are obtained from a steam flow meter and a pressure sensor, and a differential method is used to calculate the flow rate of change at adjacent time, and the standard deviation of pressure data is calculated by a sliding window to obtain the pressure fluctuation amplitude. According to the flow rate of change and the pressure fluctuation amplitude, the time correlation of the flow rate of change and the time correlation of the pressure fluctuation amplitude with the occurrence time of the limiting factor are calculated, respectively. If the time difference between the flow mutation time and the occurrence time of the limiting factor is less than the preset time delay threshold, it is determined that there is a causal relationship, and the correlation strength coefficient is recorded. Based on the correlation strength coefficient, the command execution deviation before and after the occurrence of the limiting factor is extracted from the actual adjustment feedback data of the combined heat and power unit, the deviation mean and the deviation trend are calculated, and the compensation direction and the compensation amplitude are determined according to the deviation trend. The original load command is corrected through the compensation direction and the compensation amplitude. If it is determined that it is related to pressure limitation, the flow rate of change in the command is reduced to a preset proportion of the original value, and if it is determined that it is related to speed limitation, the time interval of adjacent command points is increased, and the optimized load command of the combined heat and power unit is obtained.

[0073] In an embodiment, the real-time flow data of steam extraction is obtained by a vortex flow meter installed on the extraction pipeline, and the sampling frequency is set to 10 Hz to ensure that the rapid change of flow can be captured. The pipeline pressure data is measured by a piezoresistive pressure sensor, and the installation position is selected in the straight pipe section behind the extraction regulating valve to avoid the influence of valve disturbance on pressure measurement. When calculating the flow rate of change by the differential method, the central difference formula is used, that is, the flow difference between the current time and each sampling point is divided by the time interval. This method has higher accuracy than the forward difference.

[0074] The pressure fluctuation amplitude is calculated by a sliding window method, and the window length is set to 30 seconds, containing 300 sampling points. The standard deviation of pressure data is calculated in each window, which reflects the fluctuation degree of pressure in this period. The calculation process of the standard deviation includes calculating the average value of pressure in the window, calculating the deviation square sum of each sampling point from the average value, and then taking the square root after dividing by the number of sampling points. The sliding step is set to 1 second, realizing continuous monitoring of pressure fluctuation.

[0075] The identification of flow mutation adopts a threshold detection method, and when the flow rate of change exceeds 3 times the standard deviation of the normal fluctuation range, it is determined that flow mutation occurs. The occurrence time of the limiting factor is obtained through the limiting factor identification result in the foregoing steps. The preset time delay threshold is determined according to the dynamic characteristics of the unit, and is usually set to 5 to 10 seconds. When the time difference between the flow mutation and the occurrence of the limiting factor is less than this threshold, it indicates that there is a causal relationship between them. The correlation strength coefficient is normalized, which maps the reciprocal of the time difference to the interval of 0 to 1. The smaller the time difference is, the greater the correlation strength is.

[0076] The time period before the restriction factor occurs is selected as the first 30 seconds, and the time period after the restriction factor occurs is selected as the last 60 seconds. This asymmetric time period selection takes into account the lasting influence of the restriction factor. In each time period, the difference sequence between the actual value and the instruction value is calculated, and the bias mean value is obtained by arithmetic averaging. The determination of the bias change trend adopts a linear regression method, and a first-order polynomial fitting is performed on the bias sequence. A positive slope indicates an increasing bias trend, and a negative slope indicates a decreasing bias trend.

[0077] In one possible implementation, the compensation direction and compensation amplitude are determined based on the bias analysis result. The compensation direction is opposite to the bias direction, that is, when the actual value is less than the instruction value, the compensation direction is to increase; otherwise, it is to decrease. The basic compensation amplitude is equal to the absolute value of the bias mean value, and the trend correction coefficient is determined according to the bias change slope. If the bias is increasing, the correction coefficient is greater than 1, and is usually taken as 1.2 to 1.5; if the bias is decreasing, the correction coefficient is less than 1, and is usually taken as 0.8 to 0.9. The final compensation amplitude is the product of the basic compensation amplitude and the correction coefficient.

[0078] Exemplarily, the modification strategy of the load instruction adopts different processing modes according to the type of the restriction factor. For the case related to pressure limitation, the main problem is that the flow changes too fast to cause pressure over-limiting, and therefore it is necessary to reduce the flow change rate. In specific implementation, the flow difference value between adjacent instruction points in the original instruction sequence is multiplied by a rate adjustment coefficient, and the preset proportion of the coefficient is usually 0.6 to 0.8. The adjusted instruction sequence maintains the change trend, but the slope is more gentle, leaving a buffer time for pressure regulation. This adjustment method is particularly suitable for scenes where the capacity of the steam pipe network is limited. Further, the optimization related to rate limitation adopts a time interval adjustment strategy. The time interval of the original instruction sequence is usually uniform, such as one instruction point per second. When the rate limitation is identified, the time interval is increased to 1.5 to 2 times of the original time interval, so that the adjustment amplitude in unit time is reduced. At the same time, a transition point is inserted between adjacent instruction points, and the parameter value of the transition point is determined by linear interpolation, ensuring the continuity of the instruction. This method effectively alleviates the problem of insufficient response speed of the actuator.

[0079] It can be understood that the optimized load instruction needs to be verified before being issued for execution. The verification process includes boundary check and change rate check. The boundary check ensures that all instruction values are within the safe operating range of the unit, and the change rate check ensures that the change between adjacent instructions does not exceed the adjustment capacity of the unit.

[0080] For example, in a certain adjustment process, it is detected that the steam flow suddenly increases from 100 t / h to 110 t / h within 2 seconds, and the pressure limit occurs after 3 seconds. It is determined that there is a causal relationship between the two, and the correlation strength coefficient is 0.67. Analysis finds that the instruction execution deviation mean is -3 t / h, and shows an increasing trend. Accordingly, it is determined that the compensation direction is to increase, and the compensation amplitude is 3.6 t / h. The flow rate is reduced to 0.7 times the original value, successfully avoiding the re-occurrence of pressure overrun.

[0081] In step S107, according to the optimized combined heat and power unit load instruction, the virtual power plant operation scene data is processed by a simulation verification method to obtain a safe and efficient operation index.

[0082] According to the optimized combined heat and power unit load instruction, a virtual power plant operation scene simulation environment is constructed, load demand prediction data sequences of future time periods and impedance distribution data of each node of the power grid are extracted from a historical database, different load levels and impedance states are randomly combined by using a Monte Carlo method to form multiple sets of operation conditions. Based on the operation conditions, a load instruction sequence is executed in the simulation environment, time sequence data of unit output power is recorded, power standard deviation is calculated as fluctuation amplitude, power spectrum is extracted by fast Fourier transform, if the fundamental component offset is less than a preset limit value and the fluctuation amplitude is within an allowed range, it is determined that the condition is stable. Through the simulation execution result, heating steam flow and enthalpy value are extracted from the unit thermal parameters to calculate thermal power, electric power output is obtained from the generator end, the sum of thermal power and electric power divided by the input heat value of fuel obtains total efficiency, the efficiency values of all conditions are weighted and averaged according to load distribution probability to obtain a safe and efficient operation index including heat and power conversion efficiency index and unit operation stability index.

[0083] In an embodiment, the virtual power plant operation scene simulation environment is constructed based on a MATLAB / Simulink platform, and includes three core modules of a combined heat and power unit dynamic model, a power grid equivalent model and a load model. Load demand prediction data is extracted from a historical database of an energy management system, a time series prediction method is used to obtain a load curve of 24 hours in the future, and the time resolution is 15 minutes. The impedance distribution data of the power grid is obtained by a power flow calculation software, and includes equivalent impedance values and line impedance parameters of each bus node.

[0084] The application process of the Monte Carlo method involves random number generation and probability distribution sampling. The load level is randomly fluctuated within ±10% of the predicted value according to a normal distribution, and the impedance state is discretely sampled according to the switching probability of the power grid operation mode. Each condition is composed of a load value at a specific time and a corresponding impedance distribution, 1000 different operation conditions are generated by random combination, covering various operation scenes that the virtual power plant may encounter.

[0085] Specifically, the optimized load instruction sequence is taken as input to drive the cogeneration unit model to adjust the output according to the instructions during simulation. The calculation of power standard deviation is based on a 30-minute sliding window, reflecting the degree of power output fluctuation. Fast Fourier transform converts the time-domain power signal to the frequency domain, and the fundamental frequency component is usually 50 Hz. When the frequency offset exceeds ± 0.2 Hz, the system is considered unstable.

[0086] Power is obtained by multiplying the mass flow rate of heating steam by the specific enthalpy, and electrical power is directly measured from the generator end. The fuel input heat value is determined based on fuel composition analysis and low heat value. The total efficiency is equal to the sum of the thermal power and the electrical power divided by the fuel input heat value, with typical values between 75% and 85%.

[0087] The load distribution probability is determined based on historical statistical data. By analyzing the load data of the past year, the frequency of each load level is counted, and the probability distribution is obtained after normalization. In weighted average, the weight of high probability working condition is larger, and the weight of low probability working condition is smaller, and the final efficiency index can better reflect the actual operation.

[0088] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for adaptive adjustment of load in response to live in conjunction with virtual power plant, characterized by, The method comprises the following steps: Real-time monitoring of steam extraction quantity and feeder impedance of a combined heat and power unit, analyzing the mapping relationship between steam extraction quantity change and feeder impedance change, and extracting the interactive process characteristics of the steam extraction quantity and the feeder impedance; analyzing the dynamic response degree of the feeder impedance when the steam extraction quantity changes according to the interactive process characteristics, determining the impedance change trend; when the peak value of the impedance change trend exceeds a preset peak threshold value, obtaining the unit output and steam pressure of the combined heat and power unit, identifying the interactive amplitude of the steam extraction quantity and the feeder impedance; adjusting the load instruction parameters of the combined heat and power unit according to the interactive amplitude, generating an adaptive load instruction sequence; inputting the adaptive load instruction sequence into a virtual power plant control system, obtaining actual adjustment feedback data of the combined heat and power unit, comparing the actual adjustment feedback data with the adaptive load instruction sequence, and determining the limiting factor of the adjustment of the combined heat and power unit; Obtaining the flow change and pressure fluctuation of the steam extraction quantity, optimizing the load instruction of the combined heat and power unit, and generating an optimized load instruction; processing virtual power plant operation scene data according to the optimized load instruction to obtain an operation index.

2. The method of claim 1, wherein, The real-time monitoring of steam extraction quantity and feeder impedance of a combined heat and power unit, the analysis of the mapping relationship between steam extraction quantity change and feeder impedance change, and the extraction of the interactive process characteristics of the steam extraction quantity and the feeder impedance comprise the following steps: Obtaining the steam extraction quantity measurement value of a combined heat and power unit and real-time data of the feeder impedance, calculating the difference value of steam extraction quantity measurement values at adjacent time points to obtain a steam flow change rate; performing sliding window statistics on the real-time data of the feeder impedance, calculating the difference value between the maximum value and the minimum value in a preset time window to obtain an impedance fluctuation amplitude; establishing a time sequence mapping relationship according to the steam flow change rate and the impedance fluctuation amplitude, calculating the correlation value of the steam flow change rate sequence and the impedance fluctuation amplitude sequence; when the correlation value is greater than a preset threshold value, extracting the thermal load value and the electric load value of the combined heat and power unit, calculating the sum of the thermal load value and the electric load value as a total load value, and obtaining the thermal-electric load distribution ratio by dividing the thermal load value by the total load value; combining the steam flow change rate, the thermal-electric load distribution ratio, and the impedance fluctuation amplitude to form a data combination representing the interactive process characteristics of the steam extraction quantity and the feeder impedance.

3. The method of claim 2, wherein, The analysis of the dynamic response degree of the feeder impedance when the steam extraction quantity changes according to the interactive process characteristics, and the determination of the impedance change trend comprise the following steps: According to the steam flow rate of change, the heat and power load distribution ratio and the impedance fluctuation amplitude in the interactive process characteristics, the contribution weight of each characteristic component to the impedance change is calculated to generate a weight coefficient vector; the weight coefficient vector is subjected to dot product operation with the interactive process characteristics to obtain a dynamic response degree value; according to the dynamic response degree value, the sensitivity level of the feeder impedance to the steam extraction amount change is judged; when the dynamic response degree value exceeds a preset sensitivity threshold, the impedance change sequence under similar working conditions in the historical operation data is extracted; the impedance change sequence is subjected to time series decomposition to separate the trend component, and a fitting method is used to determine the impedance change trend.

4. The method of claim 1, wherein, When the impedance change trend peak value exceeds a preset peak threshold value, the unit output and the steam pressure of the combined heat and power unit are obtained, the interactive amplitude of the steam extraction amount and the feeder impedance is identified, including: When the impedance change trend peak value exceeds a preset peak threshold value, the unit output data and the steam pressure data of the combined heat and power unit are obtained, the steam turbine admission valve opening signal is read, and the steam flow rate is determined according to the valve flow rate characteristic curve; based on the steam flow rate and the unit output data, a time sequence corresponding data set is constructed, the correlation coefficient of the unit output data and the impedance change trend value is calculated; the steam pressure data is normalized to construct a mapping relationship matrix, the principal components of the matrix are extracted to obtain characteristic values representing the coupling strength; according to the characteristic values and the steam extraction amount, the ratio of the steam extraction amount change to the impedance change amount is calculated to obtain the interactive amplitude.

5. The method of claim 1, wherein, According to the interactive amplitude, the steam flow rate upper and lower limit values and the thermal power regulation range are read from the combined heat and power unit control station, the adjustment records under the same load level in the historical database are queried, the adjustment amount of the steam flow rate set point and the thermal power distribution ratio change value are extracted; the adjustment step of the steam flow rate set point is determined according to the adjustment amount, and the new thermal power distribution ratio is calculated; the adjustment step and the new thermal power distribution ratio are used to construct a time sequence instruction data set to generate a smooth changing instruction sequence; the instruction sequence is sampled according to the control period, and the execution time is added to generate the adaptive load instruction sequence. The adaptive load instruction sequence is input into the virtual power plant control system, the actual adjustment feedback data of the combined heat and power unit is obtained, the actual adjustment feedback data and the adaptive load instruction sequence are compared, and the limiting factor of the combined heat and power unit adjustment is determined, including:

6. The method of claim 4, wherein, The adaptive load instruction sequence is input into the virtual power plant control system, the actual adjustment feedback data of the combined heat and power unit is obtained, the actual adjustment feedback data and the adaptive load instruction sequence are compared, and the limiting factor of the combined heat and power unit adjustment is determined, including: The adaptive load instruction sequence is transmitted to a control execution module through a communication protocol, actual steam extraction amount, actual feeder impedance value and thermal power load distribution real-time data of a combined heat and power unit are collected, deviation of the actual steam extraction amount and the adaptive load instruction sequence is calculated, when the deviation exceeds a set value, the steam pressure data is read, whether the steam pressure reaches a limit boundary is judged, and whether the steam pipeline pressure limit is the limiting factor is determined; if the steam pressure is in a normal range, a thermal power load actual adjustment rate is calculated, compared with a design rate, and whether the thermal power load adjustment rate limit is the limiting factor is determined.

7. The method of claim 6, wherein, The flow variation and pressure fluctuation of the steam extraction amount are obtained, and the combined heat and power unit load instruction is optimized to generate an optimized load instruction, which comprises: Real-time flow data of the steam extraction amount and the steam pressure data are obtained from a steam flow meter and a pressure sensor, a variation rate of the flow data and a fluctuation amplitude of the pressure data are calculated, time correlation of the flow variation rate and a time when the limiting factor occurs and time correlation of the pressure fluctuation amplitude and the time when the limiting factor occurs are calculated respectively, and an association strength coefficient is recorded; according to the association strength coefficient, an instruction execution deviation of the combined heat and power unit actual adjustment feedback data is extracted, a deviation trend is calculated, the adaptive load instruction sequence is corrected according to the deviation trend, a variation rate of a steam flow set point or an instruction point time interval is adjusted, and the optimized load instruction is generated.

8. The method of claim 1, wherein, According to the optimized load instruction, virtual power plant operation scene data is processed to obtain an operation index, which comprises: According to the optimized load instruction, a virtual power plant operation scene simulation environment is constructed, load demand prediction data and power grid impedance distribution data are extracted, different load levels and impedance states are combined, the optimized load instruction is executed, a thermal power unit output power time sequence is recorded, power fluctuation amplitude and frequency spectrum characteristics are calculated, and unit operation stability is judged; thermal power unit thermal parameters are extracted, thermal power and electric power are calculated, and total efficiency is obtained; a plurality of working condition efficiency values are weighted and averaged to obtain thermal power conversion efficiency and unit operation stability.

Images