Electric boiler energy storage control method suitable for frequency modulation and deep peak regulation

Abstract

The invention belongs to the technical field of electric boiler energy storage, and discloses an electric boiler energy storage control method suitable for frequency modulation and deep peak regulation. Comprising the steps of obtaining power grid operation parameters and historical data in real time, and then calculating the frequency modulation emergency degree and the peak regulation demand level based on a power grid frequency deviation value and load demand changes. And then, dynamically optimizing the power distribution of the energy storage system and the electric boiler by utilizing a load prediction model and frequency modulation efficiency evaluation in combination with a multi-dimensional weight factor. The method also adopts a rolling optimization algorithm to carry out iterative correction on power distribution, and carries out elastic compensation according to real-time frequency fluctuation and load deviation. In addition, according to the method, a switching mechanism under the abnormal condition of the energy storage system is designed, and reliable operation of the system is ensured. According to the method, collaborative optimization control over the electric boiler and the energy storage system is achieved, the frequency modulation and peak regulation capacity of the system is remarkably improved, the response capacity to rapid change of a power grid is enhanced, and the energy utilization efficiency is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric boiler energy storage, and more particularly to an electric boiler energy storage control method suitable for frequency modulation and deep peak regulation. BACKGROUND

[0002] With the development of clean and low-carbon power systems, the large-scale integration of renewable energy has brought great challenges to the stable operation of the power grid. The traditional generator unit has limited regulation capacity and is difficult to cope with the increasingly complex power grid fluctuations. In this context, the electric boiler energy storage system, as a flexible demand-side resource, has gradually become an important means to maintain the stability of the power grid.

[0003] Electric boilers have large thermal inertia and flexible start-stop characteristics, and can regulate power to smooth load fluctuations. Energy storage systems have the advantages of fast response and strong bidirectional regulation, and can quickly participate in frequency and peak regulation of the power grid. By combining the two, an electric boiler energy storage collaborative system can theoretically achieve more efficient and flexible power grid regulation.

[0004] However, existing electric boiler energy storage control technologies have many problems, which seriously restrict their application effect in power grid regulation. In actual operation, the electric boiler and the energy storage system often operate independently, lacking a unified and coordinated control strategy, resulting in low resource utilization efficiency and difficulty in fully utilizing the collaborative advantages. Traditional fixed power allocation schemes cannot adapt to complex and variable power grid environments, often resulting in delayed regulation and response lag, affecting the stability of the power grid. In addition, existing technologies lack the ability to predict power grid frequency fluctuations and load changes, making it difficult to make reasonable scheduling decisions in advance, often leading to over-regulation or under-regulation, increasing system operating costs. When the state of charge of the energy storage system is abnormal, there is a lack of flexible switching mechanism and power redistribution strategy, which can easily cause system failure or safety hazards. At the same time, existing evaluation methods are too simple and cannot fully reflect the regulation effect of the system, hindering the continuous optimization of performance. These problems are particularly prominent in the case of large-scale renewable energy integration and rapid fluctuations in electricity demand.

[0005] In view of this, the present application proposes an electric boiler energy storage control method suitable for frequency modulation and deep peak regulation to solve the above problems. SUMMARY

[0006] In order to overcome the above-mentioned defects of the prior art, in order to achieve the above-mentioned purpose, the present application provides the following technical scheme: an electric boiler energy storage control method suitable for frequency modulation and deep peak regulation, comprising: step 1, real-time acquisition of power grid frequency deviation value, load demand change gradient, energy storage system state of charge and electric boiler operating power parameters, synchronous acquisition of load peak-valley fluctuation characteristics and frequency modulation response success rate in the historical period; Step 2, calculate the current frequency modulation urgency based on the grid frequency deviation value, generate the deep peak shaving demand level combined with the load demand change gradient and the preset peak shaving threshold, and extract the remaining dispatchable capacity and the maximum charge and discharge rate of the energy storage system; Step 3, construct a load prediction model according to the load peak-valley fluctuation characteristics, output the dynamic peak shaving power estimation value in the future period, and establish a frequency modulation efficiency coefficient matrix based on the frequency modulation response success rate; Step 4, generate a multi-dimensional weight factor by combining the current frequency modulation urgency, the deep peak shaving demand level, the remaining dispatchable capacity and the maximum charge and discharge rate, and calculate the power allocation priority of the energy storage system and the electric boiler combined with the dynamic peak shaving power estimation value and the frequency modulation efficiency coefficient matrix; Step 5, use a dynamic rolling optimization algorithm to iteratively correct the power allocation priority, generate a charge and discharge instruction set according to the correction result, and perform elastic compensation on the charge and discharge instruction set based on the real-time frequency fluctuation rate and the load deviation degree to obtain the compensated instruction set; Step 6, activate the frequency modulation mode or the peak shaving mode of the energy storage system according to the compensated instruction set, switch the electric boiler to be the main auxiliary peak shaving unit when the state of charge exceeds the safety threshold, and reallocate the power attenuation coefficient according to the new power attenuation coefficient. Reallocate the peak shaving power of the energy storage system and the electric boiler.

[0007] Further, the process of calculating the current frequency modulation urgency based on the grid frequency deviation value is as follows: The grid frequency deviation value is processed in sections, and several frequency deviation intervals are divided; the frequency modulation urgency score of each frequency deviation interval is calculated according to the historical frequency modulation response time delay and the frequency fluctuation amplitude ; ; wherein, is the historical frequency modulation response time delay, is the frequency fluctuation amplitude, and is the weight coefficient; the current frequency modulation urgency is calculated according to the frequency deviation interval to which the current grid frequency deviation value belongs, combined with the preset urgency weight coefficient .

[0008] Further, the combination of the load demand change gradient and the preset peak shaving threshold to generate the deep peak shaving demand level comprises: Perform time series analysis on the load demand change gradient to extract the change rate and duration of the load demand change gradient; compare the change rate and duration with the preset peak shaving threshold to divide the peak shaving demand level; correct the initially determined peak shaving demand level combined with the historical peak shaving response success rate Rp to obtain the final peak shaving demand level , wherein, for the correction coefficient, for the preliminary determination of the peak shaving demand level.

[0009] Further, the manner of constructing a load prediction model according to the load peak-valley fluctuation characteristics and outputting a dynamic peak shaving power estimation value in a future period includes: performing feature decomposition on the load peak-valley fluctuation characteristics in the historical period to obtain decomposition features; the decomposition features include peak time, valley time and fluctuation period; collecting a historical load data sequence ; wherein, a long short-term memory network (LSTM) is used to construct a load prediction model, the decomposition features are used as corresponding labels, the input is the historical load data sequence , and the output is a load change curve in the future period ; according to , in combination with a peak shaving power allocation ratio , a dynamic peak shaving power estimation value is calculated; the number of successful times and the number of failed times of each frequency modulation response in the historical period are counted, and then a frequency modulation response success rate is calculated; according to the frequency modulation response success rate, in combination with a frequency deviation amplitude of the frequency modulation event and a response time length , a frequency modulation efficiency evaluation function is constructed; wherein, , and are weight coefficients; the values of the frequency modulation efficiency evaluation functions of all frequency modulation events are recorded as frequency modulation efficiency coefficients, and are arranged in time sequence to obtain a frequency modulation efficiency coefficient matrix Mf.

[0010] Further, the manner of generating a multi-dimensional weight factor includes: normalizing the current frequency modulation urgency and the peak shaving demand level to obtain corresponding standardized values; according to the remaining dispatchable capacity and the maximum charging and discharging rate , a dispatch flexibility coefficient is calculated; ; wherein, is a standardized rate; based on the standardized value of the current frequency modulation urgency, the standardized value of the peak shaving demand level and the dispatch flexibility coefficient , a multi-dimensional weight factor W is generated by using a weighted average method; according to the dynamic peak shaving power estimation value, an initial power allocation ratio of the energy storage system and the electric boiler is calculated; wherein and The estimated peak shaving power of the energy storage system and the electric boiler respectively; based on the frequency modulation efficiency coefficient matrix Mf, the frequency modulation efficiency coefficient of the current period is extracted , based on the frequency modulation efficiency coefficient of the current period , the initial proportion is corrected, and the formula for correction is: ; wherein, is the correction coefficient, is the corrected proportion, the power allocation priority p1 of the energy storage system and the power allocation priority p2 of the electric boiler are calculated in combination with the corrected proportion, and then the priority sequence Pp=[p1, p2] is obtained.

[0011] Further, the iteration correction method of the power allocation priority includes: A dynamic rolling optimization model is constructed with the priority sequence Pp as the optimization target, and the optimization time window Tw and the rolling step are set; in each optimization time window, the constraint conditions of the dynamic rolling optimization model are updated based on the current grid frequency deviation value and the load demand change gradient; the objective function of the dynamic rolling optimization model is defined as: ; wherein, is the target total power; The expression of the constraint condition is: ; wherein, is a preset priority threshold; the dynamic rolling optimization model is iterated, and the corrected priority sequence, i.e. the correction result, is output; the charge and discharge instruction set of the energy storage system is formed according to the corrected priority sequence .

[0012] Further, the iteration method of the dynamic rolling optimization model includes: A main film is created, containing a plurality of sub-films, each sub-film representing a potential corrected priority sequence, and objects are placed in each sub-film, which represent the values of the power allocation priority; The evolution rules include dissolution rules, splitting rules and communication rules; The dissolution rule is that when the corrected priority sequence in a certain sub-film does not meet the constraint condition, it is dissolved; the splitting rule is that according to the value of the objective function, the sub-film with smaller value is selected for splitting to generate new potential solutions; the communication rule is that objects are randomly exchanged between different sub-films; the maximum iteration number N_max and the convergence threshold are set; for each iteration: the dissolution rule and the splitting rule are applied in turn, and the communication rule is applied at the same time, and whether the maximum iteration number N_max is reached is checked; in each iteration, the potential corrected priority sequence is fine-tuned according to the value of the objective function, and the formula for fine-tuning is: ;in, is the priority sequence before fine-tuning, is the fine-tuned priority sequence, is the current best solution, is the global optimal solution; For the K sub-membranes with the smallest objective function values, simulated annealing or hill climbing algorithm is used for local optimization. If the maximum number of iterations N_max is reached, the iteration is stopped; the potential modified priority sequence corresponding to the smallest objective function value is selected as the final modified priority sequence.

[0013] Furthermore, the charge and discharge instruction set is elastically compensated based on the real-time frequency fluctuation rate and load deviation, and the compensated instruction set is obtained in the following manner: Calculate the real-time frequency volatility of the current period and load deviation ; ;in, is the range of time periods, is the gradient of frequency fluctuation, is the preset fluctuation detection threshold; ; Respectively with the preset volatility threshold and deviation threshold To compare; if or , then calculate the elastic compensation coefficient ; in, is the initial compensation coefficient of the threshold; 、 and is the weight of the corresponding item, is the target state of charge; Based on elastic compensation coefficient , dynamically adjust the power value in the charge and discharge instruction set: ;in, It is the dynamically adjusted charge and discharge instruction set, that is, the compensated instruction set.

[0014] Furthermore, the step 6 further includes: Real-time monitoring of the state of charge of the energy storage system, comparing with the preset safety threshold; if the state of charge exceeds the safety threshold, the auxiliary peak shaving mode of the electric boiler is activated, and the available peak shaving capacity of the electric boiler is calculated; according to the available peak shaving capacity of the electric boiler and the current power distribution ratio of the energy storage system, the power attenuation coefficient is recalculated, and the power distribution scheme is updated.

[0015] The technical effect and advantage of the electric boiler energy storage control method suitable for frequency modulation and deep peak shaving of the present application are: The present application realizes the collaborative optimization control of the electric boiler and the energy storage system, greatly improves the frequency modulation and peak shaving capability and operation efficiency of the overall system, significantly enhances the response speed of the electric boiler and the energy storage system to the rapid changes of the power grid, effectively reduces the frequency fluctuation and load fluctuation, and provides a strong guarantee for the safe and stable operation of the power grid. Through dynamic optimization of the power distribution of the electric boiler and the energy storage system, the flexibility and accuracy of resource scheduling are greatly improved, the operation cost is effectively reduced, and the energy utilization efficiency is improved. In addition, the adaptability and reliability of the electric boiler energy storage system are also enhanced, which can flexibly cope with various complex working conditions, significantly improve the anti-interference ability and self-healing ability of the system; not only optimizes the operation index of the electric boiler and the energy storage system, but also promotes the power demand side response and source-grid-load-storage coordinated interaction. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 A schematic diagram of an electric boiler energy storage control method suitable for frequency modulation and deep peak shaving of the present application; Figure 2 A schematic diagram of an electric boiler energy storage control system suitable for frequency modulation and deep peak shaving of the present application. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0018] Embodiment 1 Please refer to Figure 1 The electric boiler energy storage control method suitable for frequency modulation and deep peak shaving described in this embodiment includes: Step 1, real-time acquisition of power grid frequency deviation value, load demand change gradient, energy storage system state of charge and electric boiler operation power parameters, synchronous acquisition of load peak valley fluctuation characteristics and frequency modulation response success rate in historical period; Step 2, calculate the current frequency modulation urgency based on the grid frequency deviation value, combine the load demand change gradient and the preset peak shaving threshold to generate the deep peak shaving demand level, and extract the remaining dispatchable capacity and the maximum charge and discharge rate of the energy storage system; Step 3, construct a load prediction model according to the load peak-valley fluctuation characteristics, output the dynamic peak shaving power estimate value in the future period, and establish a frequency modulation efficiency coefficient matrix based on the frequency modulation response success rate; Step 4, generate multi-dimensional weight factors by combining the current frequency modulation urgency, deep peak shaving demand level, remaining dispatchable capacity and maximum charge and discharge rate, and calculate the power allocation priority of the energy storage system and the electric boiler by combining the dynamic peak shaving power estimate value and the frequency modulation efficiency coefficient matrix; Step 5, use a dynamic rolling optimization algorithm to iteratively correct the power allocation priority, generate a charge and discharge instruction set according to the correction result, and perform elastic compensation on the charge and discharge instruction set based on the real-time frequency fluctuation rate and load deviation; obtain the compensated instruction set; Step 6, activate the frequency modulation mode or the peak shaving mode of the energy storage system according to the compensated instruction set, switch the electric boiler to be the main auxiliary peak shaving unit when the state of charge exceeds the safety threshold, and redistribute the power attenuation coefficient, and redistribute the peak shaving power of the energy storage system and the electric boiler according to the new power attenuation coefficient.

[0019] Wherein, it is necessary to explain the frequency modulation and peak shaving mentioned in the present application: frequency modulation refers to quickly adjusting the power output of the energy storage system or the electric boiler to respond to the real-time fluctuation of the grid frequency to maintain the stability of the grid frequency; peak shaving refers to adjusting the power output of the energy storage system or the electric boiler to smooth the peak-valley changes of the load and optimize the operation efficiency of the grid; the present application aims to dynamically optimize the power allocation of the energy storage system and the electric boiler to achieve efficient frequency modulation and deep peak shaving.

[0020] In traditional grid regulation, the energy storage system and the electric boiler are often controlled independently, making it difficult to achieve coordinated optimization of resources. This method collects the operating parameters of multiple systems at the same time, laying a data foundation for subsequent coordinated control.

[0021] Specifically, for real-time acquisition of power grid frequency deviation value, load demand change gradient, energy storage system state of charge and electric boiler operating power parameters, synchronous collection of load peak valley fluctuation characteristics and frequency modulation response success rate in historical period, the process is, specifically, through real-time monitoring system to obtain the running state data of power grid. Specifically, the power grid frequency deviation value is collected by the frequency sensor in real time, the unit is Hz; the load demand change gradient is calculated by the load monitoring device, the unit is MW / s; the state of charge SOC of the energy storage system is obtained by the battery management system (BMS), the range is 0-100%; the electric boiler operating power is collected by the power sensor, the unit is MW. At the same time, the load peak valley fluctuation characteristics (including peak time, valley time and fluctuation period) and the frequency modulation response success rate (such as the number of successful times and the number of failed times of each frequency modulation event) in the historical period (such as the past 24 hours) are collected synchronously; the collection period can be adjusted according to actual needs, and the collection period is 1 minute and the sampling interval is 1 second in the embodiment.

[0022] The single frequency modulation or peak regulation strategy is difficult to cope with the complex and changeable power grid operation condition, especially in the case of coexistence of deep peak regulation demand and frequent frequency modulation events; The concepts of frequency modulation urgency and deep peak regulation demand level are introduced, so that the system can flexibly adjust the response strategy according to the actual demand of the power grid.

[0023] In specific implementation, the power grid frequency deviation value is processed in sections, and a plurality of frequency deviation intervals (such as [-0.2, -0.1], [-0.1, 0], [0, 0.1], [0.1, 0.2]) are divided; According to the historical frequency modulation response time delay and the frequency fluctuation amplitude, the frequency modulation urgency score of each frequency deviation interval is calculated The frequency modulation urgency score calculation formula of the frequency deviation interval is: ; wherein, is the historical frequency modulation response time delay, is the frequency fluctuation amplitude, and is the weight coefficient; for balancing the contribution of historical frequency modulation response time delay and frequency fluctuation amplitude to the urgency score, adjusts the influence weight of historical frequency modulation response time delay; the larger value means that the response time delay has greater influence on the urgency, adjusts the influence weight of frequency fluctuation amplitude; the larger value means that the frequency fluctuation amplitude has greater influence on the urgency.

[0024] According to the frequency deviation interval to which the current power grid frequency deviation value belongs, combined with the preset urgency weight coefficient , which is used to further adjust the importance of different frequency deviation intervals; the current frequency modulation urgency​ For example, different weightings can be assigned to different frequency deviation intervals: [-0.2, -0.1] Hz: weighting 1.5; [-0.1, 0] Hz: weighting 1.0; [0, 0.1] Hz: weighting 1.0; [0.1, 0.2] Hz: weighting 1.5; Such a setting reflects that a larger frequency deviation (whether positive or negative) can require a more urgent response.

[0025] The way to generate the deep peak regulation demand level includes: Time series analysis is performed on the load demand change gradient to extract the change rate and duration of the load demand change gradient; the change rate and duration are compared with the preset peak regulation threshold L_th2 to divide the peak regulation demand level (such as low, medium, and high levels, which correspond to fixed numerical values); for example, if the change rate is greater than L_th2 and the duration is greater than 10 s, the corresponding peak regulation demand level is preliminarily determined as high; the peak regulation demand level preliminarily determined is corrected in combination with the historical peak regulation response success rate Rp to obtain the final peak regulation demand level wherein, is the correction coefficient, is the peak regulation demand level preliminarily determined.

[0026] According to the state of charge of the energy storage system (such as 80%), the remaining dispatchable capacity is calculated. wherein, is the total capacity of the energy storage system, is the lower limit of safety; the maximum charge and discharge rate R_max is determined by the specifications of the energy storage system, and the unit is MW / s.

[0027] The traditional method lacks the ability to predict future load changes, making it difficult to respond in advance; step 3 has the ability to respond proactively by building a load prediction model, improving the accuracy and timeliness of peak regulation; this is particularly useful in renewable energy grid regulation, better accommodating fluctuating renewable energy and improving grid stability.

[0028] In specific implementation, the following steps are included: The peak-to-valley fluctuation characteristics of the historical period load are decomposed to obtain the decomposition characteristics; the decomposition characteristics include the peak time, the valley time, and the fluctuation period; the historical load data sequence is collected; wherein, is the index of the sampling point; a load prediction model is constructed using a long short-term memory network (LSTM), with the decomposed features as corresponding labels, and the input being a historical load data sequence , and the output being a load change curve for a future period (e.g., the next 1 hour) .

[0029] According to , in combination with the peak shaving power allocation ratio , a dynamic peak shaving power estimate is calculated; wherein is the average load, indicating the average power consumption of the power system within a certain time period (e.g., a day, a week, or a month); a sliding window is used to continuously update the average load value to reflect the latest power consumption trend.

[0030] The number of successful times and the number of failed times of each frequency modulation response in the historical period are counted, and then the frequency modulation response success rate is calculated; according to , in combination with the frequency deviation amplitude of the frequency modulation event and the response time t_r, a frequency modulation effectiveness evaluation function is constructed; wherein , and are weight coefficients; the frequency modulation effectiveness evaluation function can comprehensively evaluate the frequency modulation performance of the system under different conditions; this multi-dimensional evaluation provides a panoramic view of the system performance.

[0031] The values of the frequency modulation effectiveness evaluation function of all frequency modulation events are recorded as the frequency modulation effectiveness coefficient, and arranged in time sequence to obtain the frequency modulation effectiveness coefficient matrix Mf; arranging the frequency modulation effectiveness coefficient in time sequence into a matrix enables the system to analyze the trend of the frequency modulation effectiveness over time; this helps to identify periodic patterns or long-term trends, providing a basis for optimizing the frequency modulation strategy.

[0032] The fixed power allocation scheme lacks flexibility and cannot be dynamically adjusted according to the real-time grid state and equipment operation, step 4 realizes the coordinated control of the energy storage system and the electric boiler by fusing multi-dimensional data, and improves the overall efficiency of frequency modulation and peak shaving. This method can realize intelligent scheduling of energy equipment in intelligent microgrid operation and improve the autonomy of the system.

[0033] In specific implementation, the following steps are included: The current frequency modulation urgency and peak shaving demand level are normalized to obtain corresponding standardized values; according to the remaining dispatchable capacity and the maximum charge / discharge rate , a dispatch flexibility coefficient is calculated; ; wherein, is the normalized rate, which is to convert the actual maximum charge and discharge rate into a standardized value, usually ranging from 0 to 1; make the charge and discharge rate of energy storage systems of different scales or types comparable.

[0034] Based on the normalized value of the current frequency modulation urgency, the normalized value of the peak regulation demand level and the scheduling flexibility coefficient , a multi-dimensional weight factor W is generated by using the weighted average method.

[0035] According to the dynamic peak regulation power estimation value, the initial proportion of power distribution of the energy storage system and the electric boiler is calculated ; wherein and are the estimated peak regulation power of the energy storage system and the electric boiler respectively; based on the frequency modulation efficiency coefficient matrix Mf, the frequency modulation efficiency coefficient of the current period is extracted, based on the frequency modulation efficiency coefficient of the current period, the initial proportion is corrected, and the formula for correction is: ; wherein, is the correction coefficient, is the corrected proportion, combining the corrected proportion, the power distribution priority p1 of the energy storage system and the power distribution priority p2 of the electric boiler are calculated; and then the priority sequence Pp=[p1,p2] is obtained.

[0036] Specifically, based on the corrected proportion , a 2x2 grey judgment matrix A is constructed, and the elements in the grey judgment matrix are the relative importance between the energy storage system and the electric boiler; for example, the element , is the uncertainty parameter.

[0037] The grey number addition operation is performed on each row of the grey judgment matrix A to obtain the row vector sum Ri, and the corresponding grey weight is obtained after normalization processing; it needs to be explained that the grey number addition operation is: .

[0038] The whitening value of each grey weight is calculated, and this whitening value is taken as the corresponding power distribution priority ; wherein, is the lower bound of the grey weight, representing the minimum value or lower limit of the interval; is the upper bound of the grey weight; this operation converts the grey interval into a specific numerical value.

[0039] In the case of considering uncertainty, the power distribution priority of the energy storage system and the electric boiler is more accurately calculated.

[0040] The iteration method for power allocation priority includes: A dynamic rolling optimization model is constructed with the priority sequence Pp as the optimization target, an optimization time window Tw (such as 5 minutes) and a rolling step (such as 1 minute) are set; in each optimization time window, the constraint conditions of the dynamic rolling optimization model are updated based on the current grid frequency deviation value and the load demand change gradient; the objective function of the dynamic rolling optimization model is defined as: ; wherein, is the target total power (i.e. the peak shaving demand to be met); The expression of the constraint condition is: ; wherein, is a preset priority threshold value; The dynamic rolling optimization model is iterated to output the corrected priority sequence; specifically, a main film is created, containing a plurality of sub-films, each sub-film representing a potential corrected priority sequence (encoded), and objects are placed in each sub-film, which represent the values of the power allocation priority.

[0041] The definition of the evolution rule includes a dissolution rule, a splitting rule and a communication rule; The dissolution rule is that when the corrected priority sequence in a certain sub-film does not meet the constraint condition, it is dissolved; the splitting rule is that according to the value of the objective function, the sub-film with a smaller value is selected for splitting to generate a new potential solution; the communication rule is that objects are randomly exchanged between different sub-films to promote the diversity of solutions.

[0042] The maximum number of iterations N_max and the convergence threshold are set; for each iteration: the dissolution rule and the splitting rule are applied in turn, and the communication rule is applied at the same time, and it is checked whether the maximum number of iterations N_max is reached; in each iteration, the potential corrected priority sequence is fine-tuned according to the value of the objective function, and the formula for fine-tuning is: ; wherein, is the priority sequence before fine-tuning, is the priority sequence after fine-tuning, is the current best solution, is the global best solution.

[0043] The K sub-films with the smallest value of the objective function are locally optimized using the simulated annealing or hill climbing algorithm, and if the maximum number of iterations N_max is reached, the iteration is stopped; the potential corrected priority sequence corresponding to the smallest value of the objective function is selected as the final corrected priority sequence.

[0044] According to the modified priority sequence, a charge-discharge instruction set of the energy storage system is constituted For example, the charge-discharge instruction set includes a power value of charge-discharge and a corresponding time stamp, and the time resolution of the instruction set is set, for example, 5 minutes; a data structure is created to store the instruction set, including the time stamp, the charge-discharge power value, and the operation mode (charging / discharging / standby).

[0045] According to the current state of charge (SOC) of the energy storage system, the available capacity is calculated, and based on the priority p1 and the total peak shaving demand, the power P_storage assigned to the energy storage system is calculated as P_total×(p1 / (p1+p2)); A time sequence is created for a predefined time range (such as the next 4 hours), and each time point is spaced at a fixed time; for each time point, the power demand at the current time is calculated, the charging or discharging mode is determined according to the positive or negative power, it is checked whether the maximum charge-discharge power limit is exceeded, it is verified whether the cumulative energy exceeds the available capacity, and an instruction is generated: {time stamp, power value, mode}; according to the modified priority sequence, a comprehensive, safe and efficient charge-discharge instruction set of the energy storage system is constituted, and it is ensured that the system can optimally perform the peak shaving task.

[0046] The real-time frequency fluctuation rate of the current period is calculated And the load deviation degree ; ; wherein, is the range of the time period, is the gradient of the frequency fluctuation, is the preset fluctuation detection threshold value; ; respectively compared with the preset fluctuation rate threshold value and the deviation threshold value ; if or , the elastic compensation coefficient is calculated; ; wherein, is the initial compensation coefficient of the threshold value, which is usually greater than 1; , and are the weights of the corresponding items, is the target state of charge (such as 50%); Based on the elastic compensation coefficient , the power value in the charge-discharge instruction set is dynamically adjusted: ; wherein, The dynamic adjustment of the charge and discharge instruction set is the compensation instruction set. The dynamic rolling optimization and elastic compensation mechanism improves the adaptability and response speed of the system to power grid fluctuations. The dynamic rolling optimization and elastic compensation mechanism improves the adaptability and response speed of the system to power grid fluctuations; in the demand side response, it can help to implement the demand side response strategy more accurately.

[0047] The charge and discharge power value and timestamp in the compensation instruction set are analyzed, and it is judged whether it is a charging or discharging instruction according to the positive and negative of the charge and discharge power value; that is, if the charge and discharge power value is greater than 0, it is a charging instruction, if the charge and discharge power value is less than 0, it is a discharging instruction, and if the charge and discharge power value is equal to 0, it is a standby instruction; A preset power change frequency threshold, power change amplitude threshold and observation time window are set; in the observation time window, the power change frequency is calculated, and the maximum power change amplitude is recorded; if the power change frequency is greater than the power change frequency threshold, and the maximum power change amplitude is less than the power change amplitude threshold, the frequency modulation mode of the energy storage system is activated; otherwise, the peak regulation mode of the energy storage system is activated; the corresponding mode activation instruction and power set value are sent to the energy storage system controller.

[0048] The state of charge SOC of the energy storage system is monitored in real time; the upper limit SOC_high (such as 90%) and the lower limit SOC_low (such as 10%) of the safety threshold are set; when it is detected that SOC is greater than SOC_high or SOC is less than SOC_low, the switching mechanism is triggered.

[0049] When the switching mechanism is triggered, the electric boiler is set as the main peak regulation device, the switching instruction is sent to the electric boiler control system, the peak regulation function of the electric boiler is activated, and the power output of the electric boiler is adjusted to make it bear the main peak regulation task.

[0050] The power attenuation coefficient of the energy storage system is defined And the power attenuation coefficient of the electric boiler ; in the initial state, Greater than , indicating that the energy storage system bears the main peak regulation task, after switching, the power attenuation coefficient is recalculated: ; ; wherein, The median value of the safe operation interval, The current state of charge, The new power attenuation coefficient of the energy storage system, The new power attenuation coefficient of the electric boiler; According to the new power attenuation coefficient, the peak regulation power of the energy storage system and the electric boiler is redistributed: ; ; wherein, is the total peak power demand, and are the power allocated to the energy storage system and the electric boiler, respectively.

[0051] The updated power instructions are sent to the energy storage system and the electric boiler, respectively, the actual output power of the two systems is monitored to ensure compliance with the instructions, and key parameters during the switching process, such as switching time and power change, are recorded; the state of charge of the energy storage system is continuously monitored; when the SOC returns to the safe range (SOC_low < SOC < SOC_high), it is evaluated whether the original power allocation scheme needs to be restored; if so, the power decay coefficient is gradually adjusted to make the energy storage system become the main peak shaving device again.

[0052] The energy storage system has limited capacity, and it is difficult to meet the long-time and high-power peak shaving demand solely relying on the energy storage system; step 6 effectively solves the capacity limitation problem of the energy storage system through the electric boiler auxiliary peak shaving mechanism, realizing more persistent peak shaving capability; this method can coordinate the electric boiler and the thermal energy storage system in the regional heating system regulation, realizing the collaborative optimization of heating load and grid regulation; through the above detailed steps, flexible switching to the electric boiler auxiliary peak shaving is realized when the state of charge of the energy storage system is abnormal, and the power distribution is dynamically adjusted to ensure the continuous and stable operation and efficient peak shaving and frequency modulation of the system.

[0053] This embodiment realizes the collaborative optimization control of the electric boiler and the energy storage system, greatly improves the frequency modulation and peak shaving capability and operation efficiency of the overall system, significantly enhances the response speed of the electric boiler and the energy storage system to the rapid changes of the power grid, effectively reduces the frequency fluctuation and load fluctuation, and provides a strong guarantee for the safe and stable operation of the power grid. Through dynamic optimization of the power distribution of the electric boiler and the energy storage system, the flexibility and accuracy of resource scheduling are greatly improved, the operation cost is effectively reduced, and the energy utilization efficiency is improved. In addition, the adaptability and reliability of the electric boiler and the energy storage system are also enhanced, which can flexibly cope with various complex working conditions, significantly improve the anti-interference ability and self-healing ability of the system; not only optimizes the operation indicators of the electric boiler and the energy storage system, but also promotes the power demand side response and the coordination and interaction of source, grid, load and storage.

[0054] Embodiment 2 Please refer to Figure 2 , some parts of which are not described in detail in Embodiment 1, a kind of electric boiler energy storage control system suitable for frequency modulation and deep peak shaving is provided, comprising: A data acquisition module is used to acquire the power grid frequency deviation value, load demand change gradient, energy storage system state of charge and electric boiler operation power parameters in real time, and synchronously acquire load peak valley fluctuation characteristics and frequency modulation response success rate in historical period. The frequency modulation peak shaving demand evaluation module calculates the current frequency modulation urgency based on the grid frequency deviation value, generates the deep peak shaving demand level in combination with the load demand change gradient and the preset peak shaving threshold, and extracts the remaining dispatchable capacity and the maximum charge and discharge rate of the energy storage system; The load prediction and performance evaluation module is used for constructing a load prediction model according to the load peak-valley fluctuation characteristics, outputting a dynamic peak shaving power prediction value in a future period, and establishing a frequency modulation performance coefficient matrix based on the frequency modulation response success rate; The power distribution priority calculation module is used for generating a multi-dimensional weight factor by fusing the current frequency modulation urgency, the deep peak shaving demand level, the remaining dispatchable capacity and the maximum charge and discharge rate, and calculating the power distribution priority of the energy storage system and the electric boiler in combination with the dynamic peak shaving power prediction value and the frequency modulation performance coefficient matrix; The dynamic optimization and instruction generation module is used for iteratively correcting the power distribution priority by using a dynamic rolling optimization algorithm, generating a charge and discharge instruction set according to the correction result, and performing elastic compensation on the charge and discharge instruction set based on the real-time frequency fluctuation rate and the load deviation degree; and obtaining the compensated instruction set. The mode switching and power redistribution module is used for synchronously activating the frequency modulation mode or the peak shaving mode of the energy storage system according to the compensated instruction set, switching the electric boiler to be the auxiliary peak shaving main body and redistributing the power attenuation coefficient when detecting that the state of charge exceeds a safety threshold, and redistributing the peak shaving power of the energy storage system and the electric boiler according to the new power attenuation coefficient. The modules are connected through wired and / or wireless modes to realize data transmission between the modules.

[0055] The above formulas are dimensionless values calculated, the formulas are obtained by collecting a large amount of data to simulate a formula of the latest real situation, and the preset parameters and threshold values in the formulas are set by a person skilled in the art according to the actual situation.

[0056] The above description is only the preferred embodiments of the present application, and the protection scope of the present application is not limited to the above embodiments. Any technical solution falling within the concept of the present application shall fall within the protection scope of the present application. It should be noted that, for ordinary technical users in the technical field, some improvements and refinements without departing from the principles of the present application shall also be considered as the protection scope of the present application.

Claims

1. A method for controlling electric boiler energy storage adapted to frequency modulation and deep peak regulation, characterized in that: include: Step 1: Obtain the grid frequency deviation value, load demand change gradient, energy storage system charge state and electric boiler operating power parameters in real time, and simultaneously collect the load peak and valley fluctuation characteristics and frequency regulation response success rate within the historical period; Step 2: Calculate the current frequency regulation urgency based on the grid frequency deviation value, generate a deep peak regulation demand level based on the load demand change gradient and the preset peak regulation threshold, and extract the remaining dispatchable capacity and maximum charge and discharge rate of the energy storage system; Step 3: Build a load forecasting model based on the load peak-valley fluctuation characteristics, output the estimated value of dynamic peak-shaving power in the future period, and establish a frequency regulation efficiency coefficient matrix based on the frequency regulation response success rate; Step 4: The current frequency regulation urgency, deep peak regulation demand level, remaining dispatchable capacity, and maximum charge and discharge rate are integrated to generate a multi-dimensional weight factor. Combined with the dynamic peak regulation power estimate and the frequency regulation efficiency coefficient matrix, the power allocation priority of the energy storage system and the electric boiler is calculated. Step 5: Use a dynamic rolling optimization algorithm to iteratively correct the power allocation priority, generate a charge and discharge instruction set based on the correction result, and elastically compensate the charge and discharge instruction set based on the real-time frequency fluctuation rate and load deviation to obtain a compensated instruction set; Step 6: Synchronously activate the frequency regulation mode or peak regulation mode of the energy storage system according to the compensated instruction set. When it is detected that the state of charge exceeds the safety threshold, switch the electric boiler to the auxiliary peak regulation body and redistribute the power attenuation coefficient. Based on the new power attenuation coefficient, redistribute the peak regulation power of the energy storage system and the electric boiler.

2. The electric boiler energy storage control method adapted to frequency modulation and deep peak regulation according to claim 1 is characterized in that: The process of calculating the current frequency regulation urgency based on the grid frequency deviation value is as follows: The grid frequency deviation value is segmented and divided into several frequency deviation intervals; the frequency regulation urgency score of each frequency deviation interval is calculated based on the historical frequency regulation response delay and frequency fluctuation amplitude. ; ;in, is the historical FM response delay, is the frequency fluctuation amplitude, and is the weight coefficient; according to the frequency deviation interval to which the current grid frequency deviation value belongs, combined with the preset emergency weight coefficient , calculate the current frequency modulation urgency .

3. The electric boiler energy storage control method adapted to frequency modulation and deep peak regulation according to claim 2 is characterized in that: The step of generating a deep peak shaving demand level by combining the load demand change gradient and the preset peak shaving threshold comprises: Perform time series analysis on the load demand gradient to extract the rate of change and duration of the load demand gradient; compare the rate of change and duration with the preset peak-shaving threshold to divide the peak-shaving demand level; combine the historical peak-shaving response success rate Rp to correct the initially determined peak-shaving demand level to obtain the final peak-shaving demand level ,in, is the correction factor, This is the preliminary determined peak load demand level.

4. The electric boiler energy storage control method adapted to frequency modulation and deep peak regulation according to claim 3 is characterized in that: Methods for building a load forecasting model based on load peak-valley fluctuation characteristics and outputting an estimated value of dynamic peak-shaving power in the future period include: Decompose the peak and valley fluctuation characteristics of the load in the historical period to obtain the decomposition characteristics; the decomposition characteristics include peak time, valley time and fluctuation period; collect historical load data series ;in, is the index of the sampling point; the long short-term memory network (LSTM) is used to build the load forecasting model, and the decomposition features are used as the corresponding labels, and the input is the historical load data sequence , the output is the load change curve of the future period ;according to , combined with the peak load power allocation ratio , calculate the estimated value of dynamic peak load power ; Count the number of successful frequency modulation responses in each historical period and number of failures , and then calculate the FM response success rate; according to the FM response success rate, combined with the frequency deviation amplitude of the FM event and response time , construct the frequency modulation efficiency evaluation function ;in, 、 and is the weight coefficient; the values ​​of the frequency modulation efficiency evaluation function of all frequency modulation events are recorded as frequency modulation efficiency coefficients, and arranged in time series to obtain the frequency modulation efficiency coefficient matrix Mf.

5. The electric boiler energy storage control method adapted to frequency modulation and deep peak regulation according to claim 4 is characterized in that: Methods for generating multi-dimensional weight factors include: Normalize the current frequency regulation urgency and peak regulation demand level to obtain the corresponding standardized value; according to the remaining dispatchable capacity and maximum charge and discharge rate , calculate the scheduling flexibility coefficient ; ;in, is the normalized rate; based on the normalized value of the current frequency regulation urgency, the normalized value of the peak regulation demand level, and the dispatch flexibility coefficient , the weighted average method is used to generate the multi-dimensional weight factor W; Calculate the initial power distribution ratio between the energy storage system and the electric boiler based on the estimated dynamic peak load power ;in and The estimated peak load of the energy storage system and the electric boiler respectively; based on the frequency regulation efficiency coefficient matrix Mf, the frequency regulation efficiency coefficient of the current period is extracted , based on the frequency modulation efficiency coefficient of the current period For the initial ratio Make corrections, and the correction formula is: ;in, is the correction factor, is the corrected ratio. Combined with the corrected ratio, the power allocation priority p1 of the energy storage system and the power allocation priority p2 of the electric boiler are calculated, and then the priority sequence Pp=[p1, p2] is obtained.

6. The electric boiler energy storage control method adapted to frequency modulation and deep peak regulation according to claim 5 is characterized in that: Methods for iteratively correcting power allocation priorities include: A dynamic rolling optimization model with the priority sequence Pp as the optimization target is constructed, and the optimization time window Tw and rolling step size are set. Within each optimization time window, the constraints of the dynamic rolling optimization model are updated based on the current grid frequency deviation value and load demand change gradient. The objective function of the dynamic rolling optimization model is defined as: ;in, is the target total power; The constraint expression is: ;in, is the preset priority threshold; the dynamic rolling optimization model is iterated to output the revised priority sequence, i.e., the revised result, and the charge and discharge instruction set of the energy storage system is constructed according to the revised priority sequence .

7. The electric boiler energy storage control method adapted to frequency modulation and deep peak regulation according to claim 6 is characterized in that: Ways to iterate on the dynamic rolling optimization model include: Create a main membrane containing several sub-membranes, each sub-membrane represents a potential revised priority sequence, and place objects in each sub-membrane, which represent the value of the power allocation priority; The evolution rules are defined including dissolution rules, splitting rules and communication rules; The dissolution rule is: when the modified priority sequence in a sub-membrane does not meet the constraints, it is dissolved; the splitting rule is: according to the value of the objective function, the sub-membrane with a smaller value is selected for splitting to generate a new potential solution; the communication rule is: randomly exchange objects between different sub-membranes; set the maximum number of iterations N_max and the convergence threshold; for each iteration: apply the dissolution rule and the splitting rule in sequence, and apply the communication rule at the same time to check whether the maximum number of iterations N_max is reached; in each iteration, the potential modified priority sequence is fine-tuned according to the value of the objective function. The fine-tuning formula is: ;in, is the priority sequence before fine-tuning, is the fine-tuned priority sequence, is the current best solution, is the global optimal solution; For the K sub-membranes with the smallest objective function values, simulated annealing or hill climbing algorithm is used for local optimization. If the maximum number of iterations N_max is reached, the iteration is stopped; the potential modified priority sequence corresponding to the smallest objective function value is selected as the final modified priority sequence.

8. The electric boiler energy storage control method adapted to frequency modulation and deep peak regulation according to claim 7 is characterized in that: Based on the real-time frequency fluctuation rate and load deviation, the charge and discharge instruction set is elastically compensated. The compensated instruction set is obtained in the following ways: Calculate the real-time frequency volatility of the current period and load deviation ; ;in, is the range of time periods, is the gradient of frequency fluctuation, is the preset fluctuation detection threshold; ； Respectively with the preset volatility threshold and deviation threshold To compare; if or , then calculate the elastic compensation coefficient ; in, is the initial compensation coefficient of the threshold; 、 and is the weight of the corresponding item, is the target state of charge; Based on elastic compensation coefficient , dynamically adjust the power value in the charge and discharge instruction set: ;in, It is the dynamically adjusted charge and discharge instruction set, that is, the compensated instruction set.

9. The electric boiler energy storage control method adapted to frequency modulation and deep peak regulation according to claim 8, characterized in that: The step 6 further includes: The state of charge of the energy storage system is monitored in real time and compared with the preset safety threshold. If the state of charge exceeds the safety threshold, the auxiliary peak-shaving mode of the electric boiler is activated and the available peak-shaving capacity of the electric boiler is calculated. Based on the available peak-shaving capacity of the electric boiler and the current power allocation ratio of the energy storage system, the power attenuation coefficient is recalculated and the power allocation plan is updated.

Images