A method for cutting load of a hydrogen production unit based on wind power ramping layering

Abstract

The present application relates to hydrogen production control technology field, especially in a kind of based on wind power climbing layered hydrogen production unit cut load method, comprising: by continuous descending interval, slope turning position and back up alternative position cut climbing segment, and according to descending form division climbing level, make wind power downlink process by single point power judgment into segment identification;By with the same climbing level under historical termination segment form matching, determine remaining life label, obtain the determination basis of subsequent downlink period after climbing;Again in combination with hydrogen production unit load shedding lag state and stable load boundary screening candidate cut load unit group and distribution cut load advance, avoid load shedding lag and low load out of bounds;Finally according to load shedding feedback state and dc bus power deviation correction cut load advance, form the cut load sequence of closed-loop adjustment.Therefore, the present application improves the cut load advance of wind power hydrogen production system, power balance stability and hydrogen production unit operation safety.

Description

Technical Field

[0001] This invention relates to the field of hydrogen production control technology, and in particular to a method for load shedding of hydrogen production units based on wind power ramping and stratification. Background Technology

[0002] In the high-fluctuation green energy hydrogen production scenario of wind power bases equipped with atmospheric pressure alkaline electrolyzer clusters, the wind farm supplies power to the hydrogen production DC bus after collection and rectification. Multiple atmospheric pressure alkaline hydrogen production units are connected in parallel to the hydrogen production DC bus and jointly absorb wind power. Due to factors such as sudden drops in wind speed, gust attenuation, wind turbine power-limited operation, grid connection restrictions, or power regulation on the site side, the wind power decline process may exhibit different climbing patterns, such as continuous gradual decline, short-term rapid decline, step drop, and repeated rise and fall after decline. During the load reduction process, the atmospheric pressure alkaline hydrogen production units also experience current decline. Operational constraints such as response time reduction, stable load boundary, gas-liquid circulation stability, and load reduction feedback lag mean that if hydrogen production units are cut off based solely on the current wind power gap or fixed power threshold, problems such as delayed load shedding, improper selection of load shedding targets, some hydrogen production units operating below the stable load boundary, increased DC bus power deviation, and frequent start-up and shutdown of hydrogen production units can easily occur. Therefore, in high-fluctuation wind power hydrogen production scenarios, existing hydrogen production load regulation technologies still struggle to generate a reasonable hydrogen production unit load shedding sequence in advance based on the continuity and morphological differences of the downward wind power ramp-up.

[0003] CN111463826A discloses a method for configuring and optimizing the control of alkaline electrolyzer arrays for wind power-to-hydrogen production. This method determines the maximum power that the electrolyzer array needs to absorb by using historical wind turbine output power data, and determines the total number of individual electrolyzers. It then classifies the operating states of the individual electrolyzers into rated power operation, fluctuating power operation, overload power operation, and shutdown states, allocating the operating states of the individual electrolyzers based on the real-time value of the wind turbine output power. This scheme can improve the ability of wind power-to-hydrogen production systems to absorb fluctuating power from the perspectives of array capacity configuration and electrolyzer state rotation. However, the key technical focus of this scheme is... The method disclosed in CN111463826A still has problems when dealing with wind power downturns in atmospheric pressure alkaline electrolyzer clusters. It does not divide the climbing segments according to the continuous decline interval, slope turning point and rebound alternation position during the wind power downturn process, nor does it divide the climbing level according to the decline pattern of the climbing segment. It is difficult to carry out differentiated load shedding treatment for different downturn patterns such as rapid wind power drop, step drop and oscillating fall. Therefore, when dealing with rapid downturn power, the load shedding triggering basis of CN111463826A still has problems such as bias towards the current power value, lack of judgment on the climbing continuity, and difficulty in determining the load shedding advance of hydrogen production unit.

[0004] CN114990600A discloses a new energy hydrogen production system and its control method, which includes a new energy input module, a power conversion module, and multiple electrolyzers. The control system controls the power conversion module based on the power of the new energy input module, so that the multiple electrolyzers operate within their load range during operation. When the new energy input power decreases, the system gradually controls each electrolyzer to reduce its load range until shutdown. When the power fluctuation exceeds the tolerance range of a single electrolyzer, at least two electrolyzers bear the power fluctuation. This scheme can perform load sharing and load reduction control on multiple electrolyzers when the new energy input power changes. However, this scheme still relies heavily on new energy. The method primarily relies on input power variations and the electrolyzer's load range for control, failing to perform morphological matching between the current ramp segment and historical termination segments at the same ramp level, and also failing to generate a remaining lifetime label reflecting the remaining descent period. Furthermore, the method does not incorporate the hydrogen production unit's load reduction lag state, stable load boundary, load reduction feedback state, and DC bus power deviation into the load shedding sequence generation process. Therefore, the method disclosed in CN114990600A still suffers from problems such as insufficient advance load shedding, inaccurate selection of load shedding targets, and difficulty in timely correction of DC bus power deviation in scenarios where wind power continuously declines and the actual load reduction of the hydrogen production unit is delayed.

[0005] Given that existing wind power-to-hydrogen load regulation technologies suffer from problems such as overemphasizing current power gaps or load-bearing interval control, lack of stratification of wind power downhill ramping patterns, difficulty in determining the remaining downhill period, and failure to incorporate the load shedding lag state of hydrogen production units and stable load boundaries into the load shedding sequence, this invention proposes a hydrogen production unit load shedding method based on wind power ramping stratification. This method solves the problem of how to pre-select suitable hydrogen production units for load shedding and generate a load shedding sequence that can be corrected by feedback under highly fluctuating wind power supply conditions, based on the wind power downhill ramping pattern and its remaining duration. Summary of the Invention

[0006] The purpose of this section is to outline some aspects of the embodiments of the present invention and to briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section, as well as in the abstract and title of the present application, to avoid obscuring the purpose of this section, the abstract and title of the invention. Such simplifications or omissions shall not be used to limit the scope of the present invention.

[0007] In view of the aforementioned existing problems, the present invention is proposed.

[0008] To solve the above-mentioned technical problems, the present invention provides the following technical solution: the climbing segment is divided according to the continuous descent interval, slope inflection point and rise alternation point in the downward process of wind power, and the climbing level is divided according to the descent pattern of the climbing segment; The remaining lifetime label of the climbing segment is determined by morphological matching with the historical termination segment under the same climbing level. Based on the remaining descent period defined by the remaining lifetime tag, the lag state of the hydrogen production unit and the stable load boundary are sorted and screened to obtain candidate load shedding unit groups and load shedding advance amounts. The load shedding advance amount is adjusted based on the load reduction feedback status of the candidate load shedding unit group and the DC bus power deviation to generate the load shedding sequence of the hydrogen production unit.

[0009] The beneficial effects of this invention are as follows: This invention segments the climbing process by dividing it into continuous descent intervals, slope inflection points, and alternating rise points, and classifies the climbing levels according to the descent pattern, transforming the wind power descent process from single-point power judgment to segmented identification; by matching the pattern with historical termination segments under the same climbing level, the remaining lifetime label is determined, providing a basis for determining the subsequent descent period of the climbing process; furthermore, by combining the load reduction lag state of the hydrogen production unit with the stable load boundary, candidate load shedding unit groups are screened and load shedding advance is allocated to avoid load reduction lag and low load exceeding the boundary; finally, the load shedding advance is corrected based on the load reduction feedback state and DC bus power deviation, forming a load shedding sequence that can be adjusted in a closed loop; therefore, this invention improves the load shedding advance, power balance stability, and operational safety of the hydrogen production unit in the wind power hydrogen production system. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a schematic flowchart of the hydrogen production unit load shedding method based on wind power ramping and stratification as shown in this invention. Detailed Implementation

[0011] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0012] Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort should fall within the scope of protection of this invention.

[0013] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0014] In a preferred embodiment, a method for cutting and loading hydrogen production units based on wind power ramp stratification is applied to the high-fluctuation green power hydrogen production scenario of an atmospheric-pressure alkaline electrolyzer cluster supporting a wind power base; wherein, the wind power base includes a wind farm power acquisition device, a DC collection bus, a hydrogen production power conversion device, an atmospheric-pressure alkaline electrolyzer cluster, and a hydrogen production energy management controller; the atmospheric-pressure alkaline electrolyzer cluster includes multiple hydrogen production units operating in parallel, and each hydrogen production unit includes an electrolyzer body, a rectifier power supply, an alkali liquid circulation device, a gas-liquid separation device, and a unit controller; the hydrogen production energy management controller receives the grid-connected side power of the wind power base, the actual input power of the DC bus, the current load of each hydrogen production unit, the load reduction feedback status of each hydrogen production unit, and the stable load boundary of each hydrogen production unit, and forms the cutting and loading sequence of the hydrogen production units according to the decline form and the remaining decline period in the wind power downward process.

[0015] Preferably, the wind power sampling value is taken from the DC collection side of the wind power base or the active power measurement point of the wind farm, and the sampling interval is 1 s; to reduce the influence of wind power sampling jitter on the ramp segment segmentation result, the wind power sampling value is first processed by a 3-point sliding median. The 3-point sliding median processing means arranging the 3 power values at the same sampling position and its adjacent sampling positions before and after in size, and taking the middle power value as the wind power sampling value at this sampling position; for the positions at the beginning and end of the sampling sequence, the adjacent existing sampling positions are taken to participate in the arrangement; the wind power sampling value after 3-point sliding median processing enters the ramp segment segmentation process of S1.

[0016] According to the embodiments of the present invention, combined with Figure 1 the flowchart shown, a method for cutting and loading hydrogen production units based on wind power ramp stratification specifically includes the following steps: S1. Cut the ramp segments according to the continuous decline intervals, slope turning positions, and recovery alternating positions in the wind power downward process, and divide the ramp levels according to the decline form of the ramp segments. It should be noted in this step that: S1.1. Calculate the difference between adjacent power sampling values in the wind power downward process according to the sampling order to obtain a power change sequence, and merge the sampling positions with continuously downward directions in the power change sequence into continuous decline intervals.

[0017] As an example, when calculating the difference, the wind power sampling value at the latter sampling position is subtracted from the wind power sampling value at the previous sampling position to obtain the power change value at the corresponding sampling position, and its calculation formula is as follows: wherein, is the power change value corresponding to the th sampling position; is the wind power sampling value at the th sampling position; is the wind power sampling value at the th sampling position; is the sampling position number, and the value starts from the 2nd sampling position.

[0018] In this embodiment, the sampling position with a power change value less than or equal to -50kW is determined as the sampling position in the downward direction, the sampling position with a power change value greater than or equal to 50kW is determined as the sampling position in the rising direction, and the sampling position with a power change value greater than -50kW and less than 50kW is determined as the sampling position in the stable direction; among them, 50kW is taken from the larger value of the active power measurement resolution of the power conversion device for hydrogen production supporting the wind power base and the short-term jitter amount of the wind farm; when the installed capacity of the wind power base is 50MW, 50kW corresponds to 0.1% of the installed capacity, which can distinguish the effective power downward and the measurement noise.

[0019] It should be noted that the continuous downward interval is an interval where at least 3 sampling positions in the downward direction appear continuously in the sampling order, and the cumulative downward amount between the first sampling position in the downward direction and the last sampling position in the downward direction within the continuous downward interval is not less than 150kW; 3 sampling positions in the downward direction correspond to a continuous downward process of 3s, which can cover the minimum recognition window before the load reduction control of the rectifier power supply of the atmospheric pressure alkaline electrolyzer; 150kW corresponds to the downward amount of at least 50kW for each of the 3 sampling positions, which can avoid misjudging a single-point jitter as a wind power downward process.

[0020] Exemplarily, the wind power sampling values after 3-point sliding median processing are 28600kW, 28480kW, 28340kW, 28220kW, 28190kW, 28020kW, 27820kW, 27760kW in sequence; the power change values obtained according to the above difference calculation method are -120kW, -140kW, -120kW, -30kW, -170kW, -200kW, -60kW in sequence; among them, the 2nd to 4th sampling positions are continuously sampling positions in the downward direction, the 5th sampling position is a sampling position in the stable direction, and the 6th to 8th sampling positions are continuously sampling positions in the downward direction; therefore, the 2nd to 4th sampling positions form 1 continuous downward interval, and the 6th to 8th sampling positions form another continuous downward interval.

[0021] S1.2. Within the continuous downward interval, determine the position where the direction of the difference between adjacent power change values changes as the slope turning position, and divide the continuous downward interval into multiple candidate climbing segments according to the slope turning position.

[0022] Specifically, when determining the slope inflection point within a continuous decreasing interval, the decrease magnitude of adjacent power change values ​​is first compared according to the sampling order. Since the power change values ​​within the continuous decreasing interval are all negative, this embodiment uses the absolute value of the power change value as the power decrease amount. When the power decrease amount at the later sampling position is greater than the power decrease amount at the previous sampling position, it is recorded as the direction of deepening decrease. When the power decrease amount at the later sampling position is less than the power decrease amount at the previous sampling position, it is recorded as the direction of slowing down decrease. If the direction of two adjacent comparisons changes from the direction of deepening decrease to the direction of slowing down decrease, or from the direction of slowing down decrease to the direction of deepening decrease, then the intermediate sampling position involved in both comparisons is determined as the slope inflection point.

[0023] When dividing a continuous descending interval according to the slope inflection point, the starting sampling position of the continuous descending interval is used as the starting boundary of the first candidate climbing segment, and the first slope inflection point is used as the ending boundary of the first candidate climbing segment. The sampling positions between two adjacent slope inflection points form the intermediate candidate climbing segment. The last slope inflection point to the ending sampling position of the continuous descending interval forms the final candidate climbing segment. If a candidate climbing segment with a sampling length of less than 3 sampling positions appears after segmentation, the candidate climbing segment is merged into the side with the closer power decrease among the adjacent candidate climbing segments. If the power decrease on both sides is the same, it is merged into the previous candidate climbing segment.

[0024] As an example, the power decreases within the continuous descent interval are sampled in the following order: 80kW, 120kW, 170kW, 130kW, 90kW, 150kW, and 210kW. The first three power decreases show an increasing downward trend, while the decreases from the third to the fifth power decrease show a decreasing downward trend. Therefore, the sampling position corresponding to the third power decrease is the slope inflection point. The decreases from the fifth to the seventh power decrease again show an increasing downward trend, so the sampling position corresponding to the fifth power decrease is another slope inflection point. The continuous descent interval is thus divided into three candidate climbing segments.

[0025] S1.3 Identify sampling position combinations that change from a descending direction to an ascending direction and then from an ascending direction to a descending direction between adjacent candidate climbing segments, and determine the sampling position combinations as ascending alternating positions.

[0026] Specifically, when identifying alternating uphill positions between adjacent candidate uphill segments, the sampling position between the termination boundary of the previous candidate uphill segment and the starting boundary of the next candidate uphill segment is used as the identification range. If, within the identification range, at least one uphill sampling position appears first in the sampling order, and at least one downhill sampling position appears after that uphill sampling position, then the combination of sampling positions formed by the last downhill sampling position, the subsequent uphill sampling position, and the next downhill sampling position is determined as the alternating uphill position. If there are stable sampling positions between the uphill sampling positions, and the number of stable sampling positions does not exceed two, then the stable sampling positions are classified as the intermediate transition part of the alternating uphill positions. If the number of stable sampling positions exceeds two, then they are used as the dividing boundary of the candidate uphill segments and are no longer classified into the same alternating uphill position.

[0027] For example, the power change values ​​between adjacent candidate climbing segments are -110kW, 70kW, 40kW, and -90kW respectively. Since -110kW corresponds to the sampling position in the descending direction, 70kW corresponds to the sampling position in the ascending direction, 40kW corresponds to the sampling position in the steady direction and there is only one such sampling position, and -90kW corresponds to the sampling position in the descending direction, the above sampling positions are combined to form one ascending alternating position.

[0028] S1.4. Based on the slope inflection point and the rise alternation point, perform boundary correction on the candidate climbing segments to obtain the climbing segments.

[0029] Specifically, when performing boundary correction on candidate climbing segments, the boundary of the candidate climbing segment formed by the slope inflection point is first used as the basis, and then the segment boundary is adjusted according to the alternating rise position. If the alternating rise position is located between two candidate climbing segments, and the power rise at the sampling position in the rise direction is less than 20% of the cumulative decrease in the previous candidate climbing segment, then the alternating rise position is incorporated into the connection area between the previous and subsequent candidate climbing segments, and the previous and subsequent candidate climbing segments are merged into a single climbing segment. If the power rise at the sampling position in the rise direction is greater than or equal to 20% of the cumulative decrease in the previous candidate climbing segment, then the sampling position in the rise direction at the alternating rise position is taken as the termination boundary of the previous climbing segment, and its... The descent sampling position is used as the starting boundary of the next climbing segment. If there are two consecutive descent sampling positions before the starting boundary of the candidate climbing segment, and the power drop at both of these descent sampling positions is not less than 50% of the average power drop within the candidate climbing segment, then the starting boundary of the candidate climbing segment is moved forward to the first of these two descent sampling positions. If there are two consecutive descent sampling positions after the ending boundary of the candidate climbing segment, and the power drop at both of these descent sampling positions is not less than 50% of the average power drop within the candidate climbing segment, then the ending boundary of the candidate climbing segment is moved backward to the last of these two descent sampling positions. The climbing segment is obtained after the above boundary correction.

[0030] Furthermore, based on the descent pattern of the climbing segment, the climbing levels are divided, including the following steps: The number of consecutive descent sampling positions within the climbing segment is counted to obtain the descent duration; the power difference between the start and end positions of the climbing segment is determined as the total descent of the segment, and the maximum power decrease between adjacent sampling positions within the climbing segment is determined as the maximum descent in a single step; Calculate the ratio of the maximum single-step descent to the total descent of the segment to obtain the single-step descent percentage; when the single-step descent percentage is greater than or equal to two-thirds, and there are descent sampling positions before and after the sampling position where the maximum single-step descent occurs, the climbing segment is divided into a step drop layer. When the climbing segment is not classified as a step drop layer, the power drop of two adjacent segments is compared in the sampling order. When the ratio of the number of times the power drop of the latter segment is greater than the power drop of the former segment to the total number of adjacent comparisons is greater than or equal to two-thirds, and the duration of the drop is greater than or equal to the sampling length required for one load reduction response of the hydrogen production unit, the climbing segment is classified as a rapid descent climbing layer. When the climbing segment is not divided into a step - drop layer or a sharp - drop climbing layer, count the number of rising - alternating positions within the climbing segment; when the number of rising - alternating positions is greater than or equal to 2, and the termination power value between every two adjacent rising - alternating positions is lower than the starting power value, divide the climbing segment into an oscillating - decline layer; When the climbing segment is not divided into a step - drop layer, a sharp - drop climbing layer or an oscillating - decline layer, divide the climbing segment into a slow - decline climbing layer.

[0031] Preferably, through step S1, it is possible to distinguish single - point drops, continuous sharp drops, oscillating declines, and slow descents in the high - fluctuation green power input of the wind power base, solving the problem of premature or lagging load shedding of the hydrogen production unit due to judging load shedding only based on the instantaneous power difference during the wind power descent process; this step enables the subsequent remaining - life label and load - shedding lead to be based on the decline pattern of the climbing segment, improving the stability of load - shedding determination for the alkaline water electrolyzer cluster under high - fluctuation wind power.

[0032] S2. Match the morphology of the climbing segment with the historical termination segments at the same climbing level to determine the remaining - life label of the climbing segment. Here, it should be noted in this step: S2.1. Select historical termination segments with the same climbing level from the historical climbing segment library to obtain the set of historical termination segments at the same level.

[0033] Specifically, the historical climbing segment library is formed during the historical operation of the alkaline water electrolyzer cluster supporting the wind power base; each historical climbing segment in the historical climbing segment library includes segment number, segment start sampling position, segment end sampling position, climbing level, decline duration, total segment decline, maximum single - step decline, single - step decline ratio, number of rising - alternating positions, decline - ratio sequence of each sampling position within the segment, and termination - state flag; among them, the termination - state flag includes natural - return - to - stability termination, grid - connection - limit - release termination, and hydrogen - production - load - shedding termination; to ensure that the historical termination segment and the current climbing segment have the same physical meaning, in this embodiment, historical climbing segments with termination - state flags of natural - return - to - stability termination or grid - connection - limit - release termination are selected as historical termination segments, and historical climbing segments with hydrogen - production - load - shedding termination are not selected as historical termination segments.

[0034] When selecting a set of historical termination segments at the same level, historical termination segments with different climbing levels are first filtered out from the historical climbing segment library. Then, historical termination segments with the same climbing level as the climbing segment are retained. Specifically, if a climbing segment is classified as a steep descent climbing layer, only historical termination segments below the steep descent climbing layer are selected; if a climbing segment is classified as an oscillating fall layer, only historical termination segments below the oscillating fall layer are selected. The set of historical termination segments at the same level must include at least 20 historical termination segments. If there are fewer than 20, historical termination segments of the same climbing level are continuously added to adjacent wind power operation days until the number reaches 20, or reaches the total number of that climbing level in the historical climbing segment library.

[0035] S2.2. Compare the duration of descent, total descent, maximum descent per step, percentage of descent per step, and number of alternating descent positions of the climbing segment with the historical terminating segments in the same level set to obtain a set of segments with similar morphology.

[0036] Specifically, when comparing the climbing segment with historical termination segments in the same level set, the selection is based on the following criteria: descent duration, total descent amount, maximum single-step descent amount, single-step descent percentage, and number of alternating rise positions. The difference in descent duration should not exceed 50% of the sampling length required for a single load reduction response of the hydrogen production unit; the difference in total descent amount should not exceed 25% of the total descent amount of the climbing segment; the difference in maximum single-step descent amount should not exceed 30% of the maximum single-step descent amount of the climbing segment; the difference in single-step descent percentage should not exceed 0.15; and the difference in the number of alternating rise positions should not exceed one. Historical termination segments that meet all the above conditions are included in the set of morphologically similar segments.

[0037] It should be noted that the set of morphologically similar segments adopts the form of segment entries arranged by segment number. Each segment entry includes the historical termination segment number, the starting sampling position of the historical termination segment, the ending sampling position of the historical termination segment, the duration of descent of the historical termination segment, the total descent amount of the historical termination segment, the maximum single-step descent amount of the historical termination segment, the percentage of single-step descent of the historical termination segment, the number of alternating rise positions of the historical termination segment, and the descent ratio sequence of the historical termination segment. For example, the set of morphologically similar segments includes historical termination segments A17, A23, A41, and A56. Among them, the duration of descent of historical termination segment A17 is 18 sampling positions, the total descent amount is 2400kW, the maximum single-step descent amount is 410kW, the percentage of single-step descent is 0.17, and the number of alternating rise positions is 1. The duration of descent of historical termination segment A23 is 20 sampling positions, the total descent amount is 2520kW, the maximum single-step descent amount is 460kW, the percentage of single-step descent is 0.18, and the number of alternating rise positions is 1.

[0038] S2.3. The ratio of the descent amount of the current termination position of the climbing segment relative to the starting position of the climbing segment to the total descent amount of the segment is determined as the current descent ratio. In the set of morphologically similar segments, the positions in each historical termination segment that have the same descent ratio are determined as matching positions. The sampling length between the matching position and the termination position of the corresponding historical termination segment is calculated to obtain the set of remaining sampling lengths.

[0039] Specifically, when calculating the sampling length between the matching position and the corresponding historical termination segment termination position, first calculate the descent amount of the current termination position of the climbing segment relative to the starting position of the climbing segment, and then calculate the current descent ratio; then find the position closest to the current descent ratio in the descent ratio sequence of each historical termination segment; if there are multiple positions with the same difference from the current descent ratio, then take the position with the later sampling order as the matching position to avoid underestimating the remaining descent period.

[0040] As an example, the remaining sampling length is calculated using the following formula: in, For the first The remaining sampling length corresponding to each historical termination segment; For the first The termination position number of each historical termination segment; For the first The matching position number in each historical termination segment that corresponds to the current descent ratio; This refers to the sequence number of the historical termination segment in a set of morphologically similar segments.

[0041] For example, if the total descent of a current uphill segment is 2400kW, and the descent from the current termination position to the starting position is 1440kW, then the current descent ratio is 0.60. The termination position of historical termination segment A17 in the set of similar segments is numbered 136, and the position closest to 0.60 in the descent ratio sequence is numbered 124. Therefore, the remaining sampling length corresponding to historical termination segment A17 is 12 sampling positions. The termination position of historical termination segment A23 is numbered 212, and the position closest to 0.60 in the descent ratio sequence is numbered 198. Therefore, the remaining sampling length corresponding to historical termination segment A23 is 14 sampling positions. After processing all similar segments in sequence, the set of remaining sampling lengths is obtained.

[0042] S2.4 Arrange the sampling lengths in the remaining sampling length set in ascending order of value, and take the sampling length in the middle position as the representative length of the remaining lifetime; when the representative length of the remaining lifetime is less than or equal to the sampling length required for one load reduction response of the hydrogen production unit, the ramp segment is divided into short remaining lifetime labels.

[0043] In this embodiment, the sampling length required for a single load reduction response of the hydrogen production unit comes from the load reduction response test of the atmospheric pressure alkaline electrolyzer hydrogen production unit. The load reduction response test is conducted when the hydrogen production unit is at 70% to 90% of its rated load. A load reduction command of 10% of the rated load is issued to the hydrogen production unit, and the sampling length between the sampling position where the load reduction command is issued and the sampling position where the current load of the hydrogen production unit enters the target load ± 2% of the rated load is recorded. Ten consecutive load reduction response tests are conducted on the same hydrogen production unit, and the maximum value of the sampling lengths obtained from the ten tests is taken as the sampling length required for a single load reduction response of the hydrogen production unit. For an atmospheric pressure alkaline electrolyzer cluster composed of multiple hydrogen production units, the maximum value of the sampling lengths required for a single load reduction response of each hydrogen production unit is taken as the sampling length required for a single load reduction response of the hydrogen production unit at the cluster level.

[0044] As an example, with a sampling interval of 1 second, a certain atmospheric pressure alkaline electrolyzer cluster includes hydrogen production units E1 to E6. The maximum sampling lengths obtained from a single load reduction response test of each hydrogen production unit are 10, 11, 12, 10, 11, and 12 sampling positions, respectively. Therefore, the sampling length required for a single load reduction response of the hydrogen production units at the cluster level is 12 sampling positions. If the set of remaining sampling lengths is arranged in ascending order of value as 8, 10, 12, 14, and 16, then the sampling length in the middle position is 12 sampling positions, and the remaining lifetime representative length is 12 sampling positions. Since the remaining lifetime representative length is less than or equal to 12 sampling positions, the ramp segment is classified as a short remaining lifetime label.

[0045] S2.5 When the remaining lifetime representative length is greater than the sampling length required for one load reduction response of the hydrogen production unit, and less than or equal to the sampling length required for two consecutive load reduction responses of the hydrogen production unit, the ramp segment is classified as a medium remaining lifetime label.

[0046] In this embodiment, the sampling length required for two consecutive load reduction responses of the hydrogen production unit comes from the continuous load reduction response test of the same hydrogen production unit. The continuous load reduction response test is conducted when the hydrogen production unit is between 80% and 90% of its rated load. First, a load reduction command of 10% of the rated load is issued. After the current load of the hydrogen production unit enters the range of the first target load ± 2% of the rated load, a second load reduction command of 10% of the rated load is immediately issued, and the sampling length from the sampling position of the first load reduction command to the current load of the hydrogen production unit entering the range of the second target load ± 2% of the rated load is recorded. Ten consecutive load reduction response tests are conducted on the same hydrogen production unit, and the maximum value of the sampling length obtained from the ten tests is taken as the sampling length required for two consecutive load reduction responses of the hydrogen production unit. For atmospheric pressure alkaline electrolyzer clusters, the maximum value of the sampling length required for two consecutive load reduction responses of each hydrogen production unit is taken as the sampling length required for two consecutive load reduction responses of the hydrogen production units at the cluster level.

[0047] For example, with a sampling interval of 1 second, the maximum sampling lengths obtained from two consecutive load reduction response tests of hydrogen production units E1 to E6 are 22, 23, 24, 22, 23, and 24 sampling positions, respectively. Therefore, the sampling length required for two consecutive load reduction responses of hydrogen production units at the cluster level is 24 sampling positions. When the remaining lifetime representative length is greater than 12 sampling positions and less than or equal to 24 sampling positions, the ramp segment is classified as a medium remaining lifetime label. When the remaining lifetime representative length is greater than 24 sampling positions, the ramp segment is classified as a long remaining lifetime label.

[0048] S2.6 When the remaining lifetime representative length is greater than the sampling length required for two consecutive load reduction responses of the hydrogen production unit, the ramp segment is classified as a long remaining lifetime tag.

[0049] Preferably, step S2 allows for the estimation of the remaining descent period of the current climbing segment by utilizing historically terminated segments already completed at the same climbing level. This addresses the problem of difficulty in handling oscillating descents, step falls, and rapid climbing differences when extrapolating solely based on the current slope. This step correlates the remaining lifetime tag with the load reduction response capability of the atmospheric pressure alkaline electrolyzer, ensuring that the subsequent load shedding advance is no longer determined solely by the instantaneous power gap, thereby reducing the risk of increased DC bus power deviation caused by the lag in the hydrogen production unit's load reduction.

[0050] S3. Based on the remaining descent period defined by the remaining lifetime tag, sort and screen the hydrogen production unit's load reduction lag state and stable load boundary to obtain candidate load shedding unit groups and load shedding lead times. Note that the following should be noted in this step: S3.1 Determine the remaining descent period according to the remaining lifetime label; whereby a short remaining lifetime label corresponds to a remaining descent period less than or equal to the sampling length of the descent response of one hydrogen production unit, a medium remaining lifetime label corresponds to a remaining descent period greater than one and less than or equal to the sampling length of the descent response of two hydrogen production units, and a long remaining lifetime label corresponds to a remaining descent period greater than the sampling length of the descent response of two hydrogen production units.

[0051] S3.2. Compare the current load of the hydrogen production unit with the stable load boundary, retain the hydrogen production units whose current load is higher than the stable load boundary, and obtain the load reduction margin from the difference between the current load and the stable load boundary.

[0052] It should be noted that the current load of the hydrogen production unit in this embodiment is taken from the active power measurement value of the DC output side of the rectifier power supply of each hydrogen production unit. The active power measurement value is obtained at a 1-second sampling interval and uses the same sampling position number as the wind power sampling value in S1. The stable load boundary is taken from the lowest load boundary that the atmospheric pressure alkaline electrolyzer can operate for a long time under stable hydrogen production conditions. The stable load boundary is jointly determined by the rated power of the electrolyzer, the alkaline solution circulation state, the cell temperature range, and the gas-liquid separation stability.

[0053] In this embodiment, the stable load boundary is taken as 35% of the rated power of the hydrogen production unit. When the rated power of the hydrogen production unit is 2000kW, the stable load boundary is 700kW; when the rated power of the hydrogen production unit is 2500kW, the stable load boundary is 875kW.

[0054] When comparing the current load of a hydrogen production unit with the stable load boundary, if the current load of the hydrogen production unit is greater than the stable load boundary, the hydrogen production unit is retained as a hydrogen production unit that can participate in load reduction, and the difference between the current load of the hydrogen production unit and the stable load boundary is determined as the load reduction margin; if the current load of the hydrogen production unit is less than or equal to the stable load boundary, the hydrogen production unit does not enter the candidate load reduction unit group; to avoid the hydrogen production unit repeatedly entering and leaving the candidate load reduction unit group near the stable load boundary, this embodiment introduces a boundary hysteresis of 2% of the rated power during the comparison; when the current load of the hydrogen production unit is higher than the stable load boundary plus 2% of the rated power, it enters the candidate screening; when the current load of the hydrogen production unit is lower than or equal to the stable load boundary, it leaves the candidate screening.

[0055] For example, the rated power of hydrogen production units E1 to E6 is 2000kW, and the stable load boundary is 700kW. At a certain sampling location, the current load of hydrogen production unit E1 is 1700kW, the current load of hydrogen production unit E2 is 1500kW, the current load of hydrogen production unit E3 is 720kW, the current load of hydrogen production unit E4 is 1300kW, the current load of hydrogen production unit E5 is 690kW, and the current load of hydrogen production unit E6 is 1600kW. Since the current loads of E1, E2, E4, and E6 are higher than the stable load boundary plus the entry boundary of 40kW, E1, E2, E4, and E6 are included in the candidate screening. E3 is not higher than the entry boundary, and E5 is lower than the stable load boundary, so E3 and E5 are not included in the candidate screening. The corresponding load reduction margins are 1000kW, 800kW, 600kW, and 900kW, respectively.

[0056] S3.3 Compare the hysteresis sampling length in the hysteresis state of the hydrogen production unit with the remaining hysteresis period, and retain the hydrogen production units whose hysteresis sampling length is less than or equal to the remaining hysteresis period.

[0057] In this embodiment, the load reduction lag state of the hydrogen production unit is obtained by the sampling length between the historical load reduction command and the actual load reduction response of each hydrogen production unit. Specifically, after each hydrogen production unit receives a load reduction command, it records the sampling position where the load reduction command was issued and continuously obtains the current load of the hydrogen production unit. When the current load of the hydrogen production unit first enters the load reduction target load ±2% of the rated power range, the sampling position is taken as the load reduction arrival sampling position. The number of sampling positions between the load reduction arrival sampling position and the sampling position where the load reduction command was issued is the lag sampling length. If the hydrogen production unit enters the target load range within the sampling length required for one load reduction response of the hydrogen production unit, then the hydrogen production unit is in a normal load reduction state. If the hydrogen production unit enters the target load range after exceeding the sampling length required for one load reduction response of the hydrogen production unit but not exceeding the sampling length required for two consecutive load reduction responses of the hydrogen production unit, then the hydrogen production unit is in a delayed load reduction state. If the hydrogen production unit still does not enter the target load range after exceeding the sampling length required for two consecutive load reduction responses of the hydrogen production unit, then the hydrogen production unit is in a non-priority load reduction state.

[0058] When comparing the hysteresis sampling length with the remaining descent period, the remaining descent period corresponding to the short remaining lifetime tag, medium remaining lifetime tag, or long remaining lifetime tag determined by S2 is used as the comparison boundary. If the hysteresis sampling length of the hydrogen production unit is less than or equal to the remaining descent period, the hydrogen production unit is retained. If the hysteresis sampling length of the hydrogen production unit is greater than the remaining descent period, the hydrogen production unit is not included in the candidate load shedding unit group for this round. For the short remaining lifetime tag, the remaining descent period is less than or equal to 12 sampling positions, so only hydrogen production units with a hysteresis sampling length of no more than 12 sampling positions are retained. For the medium remaining lifetime tag, the remaining descent period is greater than 12 sampling positions and no more than 24 sampling positions, so hydrogen production units with a hysteresis sampling length of no more than the corresponding remaining descent period are retained. For the long remaining lifetime tag, the remaining descent period is greater than 24 sampling positions, and hydrogen production units with a longer hysteresis sampling length but still having load shedding feedback can be accepted.

[0059] S3.4. Sort the retained hydrogen production units in descending order of available load capacity and ascending order of hysteresis sampling length to obtain candidate load-cutting unit groups.

[0060] For example, the current ramp segment is divided into a medium remaining life tag, and 20 sampling positions are taken for the remaining descent period. The lag sampling lengths of hydrogen production units E1, E2, E4, and E6 are 9, 14, 21, and 11 sampling positions, respectively. Therefore, E1, E2, and E6 are retained, while E4 is not included in the candidate load shedding unit group in this round because its lag sampling length exceeds 20 sampling positions. Subsequently, the candidate load shedding unit groups are obtained by arranging them in descending order of available load capacity and ascending order of lag sampling length. For example, E1 has an available load capacity of 1000kW and a lag sampling length of 9 sampling positions; E6 has an available load capacity of 900kW and a lag sampling length of 11 sampling positions; and E2 has an available load capacity of 800kW and a lag sampling length of 14 sampling positions. Therefore, the order of the candidate load shedding unit groups is E1, E6, and E2.

[0061] S3.5. Based on the start and end positions of the remaining descent period, calculate the difference between the wind power value corresponding to the start position and the wind power value corresponding to the end position to obtain the remaining descent power.

[0062] S3.6 When the DC bus power deviation is greater than zero, the remaining power reduction amount is added to the DC bus power deviation to obtain the load amount to be cut off in advance; when the DC bus power deviation is less than or equal to zero, the remaining power reduction amount is used as the load amount to be cut off in advance.

[0063] S3.7. According to the sorting order of the candidate load shedding unit groups, compare the load to be shedding in advance with the available load margin of each hydrogen production unit in turn; where: When the amount of load to be cut off in advance is greater than the available load margin of the current hydrogen production unit, the available load margin of the current hydrogen production unit is used as the unit cut-off advance amount of the current hydrogen production unit, and the amount of load to be cut off in advance is reduced by the unit cut-off advance amount and then allocated to the next hydrogen production unit. When the amount of load to be cut off in advance is less than or equal to the available load margin of the current hydrogen production unit, the amount of load to be cut off in advance will be used as the unit cut-off advance amount of the current hydrogen production unit, and the allocation to subsequent hydrogen production units will be stopped. The hydrogen production units with allocated unit load shedding advances and their unit load shedding advances are arranged to obtain the load shedding advance.

[0064] In a preferred embodiment, after obtaining the candidate load-cutting unit group, the remaining power reduction amount is calculated based on the start and end positions of the remaining reduction period, and the amount of load to be cut off in advance is determined based on the DC bus power deviation. The DC bus power deviation is the difference between the current absorbed power of the hydrogen production unit and the actual input power of the DC bus. If the DC bus power deviation is greater than 0, it indicates that the current absorbed power of the atmospheric pressure alkaline electrolyzer cluster is higher than the actual input power of the DC bus, and this deviation needs to be added to the remaining power reduction amount. If the DC bus power deviation is less than or equal to 0, it indicates that the actual input power of the DC bus can cover the hydrogen production unit. The current absorbed power is used to determine the amount of load to be cut off in advance, which is the remaining power reduction amount. Then, the amount of load to be cut off in advance is allocated according to the candidate load shedding unit group. Taking the aforementioned candidate load shedding unit groups E1, E6, and E2 as an example, if the remaining power reduction amount is 1800kW and the DC bus power deviation is 300kW, then the amount of load to be cut off in advance is 2100kW. E1 is allocated 1000kW, leaving 1100kW; E6 is allocated 900kW, leaving 200kW; E2 is allocated 200kW and then the allocation stops. Thus, the load shedding advance amount is obtained as follows: E1 is 1000kW, E6 is 900kW, and E2 is 200kW.

[0065] It should be noted that step S3 allows priority to be given to hydrogen production units with current loads higher than the stable load boundary, larger load reduction margins, and faster load reduction responses during the remaining load reduction period. This solves the problem of inaccurate load shedding selection in atmospheric pressure alkaline electrolyzer clusters caused by the superposition of low unit load boundaries, response lag, and DC bus power deviations when wind power continues to decline. This step ensures that the load shedding lead time is constrained by both the remaining wind power reduction and the stable operating boundary of the electrolyzer, which can reduce the risks of low load fluctuations in the electrolyzer, insufficient power of the DC bus, and frequent and significant load reductions in hydrogen production units.

[0066] S4. Adjust the load shedding advance based on the load reduction feedback status of the candidate load shedding unit group and the DC bus power deviation to generate the hydrogen production unit load shedding sequence. Note that the following should be noted in this step: S4.1. Issue load reduction instructions to the candidate load reduction unit group according to the load reduction advance amount, and obtain the load reduction feedback status of each hydrogen production unit in the candidate load reduction unit group; wherein, the load reduction feedback status includes the load reduction status, the delayed load reduction status, and the no load reduction status.

[0067] Specifically, when issuing load reduction instructions to the candidate load shedding unit group according to the load shedding advance, the hydrogen production energy management controller converts the unit load shedding advance of each hydrogen production unit into the target active power of the rectifier power supply according to the order of the candidate load shedding unit group obtained in S3. The target active power is the power value after subtracting the corresponding unit load shedding advance from the current load of the hydrogen production unit, and the target active power shall not be lower than the stable load boundary. If the target active power calculated according to the unit load shedding advance is lower than the stable load boundary, the target active power is taken as the stable load boundary, and the excess load shedding advance is distributed to the next hydrogen production unit in the candidate load shedding unit group.

[0068] Furthermore, the load reduction command includes the hydrogen production unit number, target active power, load reduction start sampling position, allowable sampling position, and the unit's load shedding advance. The allowable sampling position is determined based on the sampling length required for one load reduction response of the hydrogen production unit. For example, if the sampling length required for one load reduction response of hydrogen production unit E1 is 12 sampling positions, and the load reduction start sampling position is the 500th sampling position, then the allowable sampling position is the 512th sampling position. After the load reduction command is issued to the hydrogen production unit controller, the hydrogen production unit controller adjusts the DC output power of the rectifier power supply and feeds back the current load, target active power, load reduction execution status, and fault alarm status.

[0069] Upon obtaining the load reduction feedback status, the hydrogen production energy management controller obtains the current load of each hydrogen production unit in the candidate load reduction unit group at a 1-second sampling interval after the load reduction command is issued. If the current load of the hydrogen production unit enters the target active power ±2% rated power range before reaching the sampling position, and there are no rectifier power supply fault alarms, alkali circulation fault alarms, or tank temperature over-limit alarms, then the hydrogen production unit is marked as having a reduced load status. If the current load of the hydrogen production unit does not enter the target active power ±2% rated power range before reaching the sampling position, but the load reduction direction is consistent with the load reduction command, and there are no fault alarms, then the hydrogen production unit is marked as having a delayed load reduction status. If the current load of the hydrogen production unit does not show a reduction change consistent with the load reduction command, or if any of the rectifier power supply fault alarms, alkali circulation fault alarms, or tank temperature over-limit alarms occur, then the hydrogen production unit is marked as not having a reduced load status.

[0070] S4.2. Keep the hydrogen production units with the load reduction feedback status in the reduced load state in the candidate load shedding unit group, move the hydrogen production units with the load reduction feedback status in the delayed load reduction state to the next group, and remove the hydrogen production units with the load reduction feedback status in the unreduced load state.

[0071] S4.3 When the DC bus power deviation is greater than zero, the load shedding advance is increased by the DC bus power deviation to obtain the adjusted load shedding advance; when the DC bus power deviation is less than or equal to zero, the adjusted load shedding advance is taken as the load shedding advance.

[0072] S4.4 According to the adjusted load shedding advance, the hydrogen production units that are retained and moved later are arranged from largest to smallest according to the load reduction margin, so as to obtain the load shedding sequence of the hydrogen production units.

[0073] Specifically, based on the adjusted load shedding advance, when arranging the retained and relocated hydrogen production units according to their available load reduction margin from largest to smallest, the hydrogen production units already in a reduced load state are first prioritized, and the hydrogen production units in a lagging load reduction state are second prioritized; hydrogen production units not in a reduced load state are not included in the sorting; within the first priority group, they are arranged from largest to smallest available load reduction margin; when available load reduction margins are the same, they are arranged from smallest to largest lag sampling length; when lag sampling lengths are still the same, they are arranged in ascending order of hydrogen production unit number; the second priority group is also arranged according to the same rules, but the second priority group as a whole is placed after the first priority group; after sorting, the adjusted load shedding advance is allocated to the arranged hydrogen production units in sequence to obtain the hydrogen production unit load shedding order.

[0074] Preferably, the hydrogen production unit load shedding sequence includes the hydrogen production unit number, target active power, unit load shedding advance amount, and arrangement order.

[0075] In a preferred example, the wind power base is equipped with six atmospheric pressure alkaline electrolyzer hydrogen production units. The rated power of each hydrogen production unit, E1 to E6, is 2000kW, and the stable load boundary is 700kW. The candidate load shedding unit group obtained through step S3 is E1, E6, and E2, with shedding lead times of 1000kW for E1, 900kW for E6, and 200kW for E2, respectively. After the load reduction command is issued, the current load of E1 decreases from 1700kW to 710kW, falling within the target active power range of ±40kW. The load on E6 has decreased from 1600kW to 1180kW, which is not yet within the target active power range of 700kW±40kW, but the load decrease direction is consistent and there is no fault alarm, so it is marked as a delayed load reduction state; the load on E2 has decreased from 1500kW to 1300kW, which is within the target active power range of 1300kW±40kW, so it is marked as a load reduction state; at this time, the DC bus power deviation is 200kW, and the adjusted load shedding advance needs to be increased by 200kW.

[0076] Based on the load reduction feedback status, E1 and E2 enter the first priority group, and E6 enters the second priority group. E1 has a remaining load reduction capacity of 10kW, E2 has a remaining load reduction capacity of 600kW, and E6 has a remaining load reduction capacity of 480kW. The first priority group is arranged in descending order of load reduction capacity as E2 and E1, and the second priority group is E6. Since the additional load reduction advance is 200kW, it is first allocated to E2, and the target active power of E2 is adjusted from 1300kW to 1100kW. E1 and E6 will not be allocated any more. The final hydrogen production unit load reduction order is as follows: the first order is E2, with a unit load reduction advance of 200kW and a target active power of 1100kW; the second order is E1, maintaining the target active power of 710kW after load reduction; the third order is E6, which is retained in the backward state and will participate in the sorting based on its load reduction feedback status at the next sampling position.

[0077] It should be noted that, through step S4, this embodiment can continue to make secondary adjustments based on the actual load reduction feedback status of the hydrogen production unit and the DC bus power deviation after the load reduction advance amount has been allocated. This solves the problem that the load reduction sequence is still executed according to the initial order when there are response differences, local lags, or no response from the hydrogen production unit after the load reduction command is issued. This step enables the atmospheric pressure alkaline electrolyzer cluster to rearrange the load reduction targets according to the actual load reduction feedback during the period when the wind power continues to decline, reducing the probability of unresponsive hydrogen production units occupying the load reduction quota and improving the stability of DC bus power balance.

[0078] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for load shedding in hydrogen production units based on wind power ramp-up stratification, characterized in that, include: The climbing segments are divided according to the continuous decreasing intervals, slope turning points, and alternating rising points during the downward process of wind power, and the climbing levels are divided according to the descent pattern of the climbing segments. The segmentation of the climbing segment includes: calculating the difference between adjacent power sampling values ​​during the downward wind power process according to the sampling order to obtain a power change sequence; grouping the sampling positions in the power change sequence that are continuously decreasing in direction into a continuous decreasing interval; within the continuous decreasing interval, determining the position where the direction of the difference between adjacent power change values ​​changes as the slope inflection position, and segmenting the continuous decreasing interval into multiple candidate climbing segments according to the slope inflection position; identifying sampling position combinations between adjacent candidate climbing segments that change from a decreasing direction to an increasing direction and then from an increasing direction to a decreasing direction, and determining the sampling position combinations as increasing alternating positions; and performing boundary correction on the candidate climbing segments based on the slope inflection position and the increasing alternating position to obtain the climbing segment. The remaining lifetime label of the climbing segment is determined by performing morphological matching with historical termination segments under the same climbing level. Based on the remaining descent period defined by the remaining lifetime tag, the lag state of the hydrogen production unit and the stable load boundary are sorted and screened to obtain candidate load shedding unit groups and load shedding advance amounts. Obtaining the candidate load shedding unit group includes: determining the remaining load reduction period according to the remaining lifetime tags; wherein, the short remaining lifetime tag corresponds to a remaining load reduction period less than or equal to the sampling length of the load reduction response of one hydrogen production unit, the medium remaining lifetime tag corresponds to a remaining load reduction period greater than one and less than or equal to the sampling length of the load reduction response of two hydrogen production units, and the long remaining lifetime tag corresponds to a remaining load reduction period greater than the sampling length of the load reduction response of two hydrogen production units; comparing the current load of the hydrogen production unit with the stable load boundary, retaining hydrogen production units whose current load is higher than the stable load boundary, and obtaining the load reduction margin from the difference between the current load and the stable load boundary; comparing the hysteresis sampling length in the load reduction hysteresis state of the hydrogen production unit with the remaining load reduction period, retaining hydrogen production units whose hysteresis sampling length is less than or equal to the remaining load reduction period; sorting the retained hydrogen production units in descending order of the load reduction margin and ascending order of the hysteresis sampling length to obtain the candidate load shedding unit group; The load shedding advance amount is adjusted based on the load reduction feedback status of the candidate load shedding unit group and the DC bus power deviation to generate the hydrogen production unit load shedding sequence. This includes: issuing a load reduction command to the candidate load shedding unit group according to the load shedding advance amount, and obtaining the load reduction feedback status of each hydrogen production unit in the candidate load shedding unit group; retaining hydrogen production units with a load reduction feedback status of "load reduced" in the candidate load shedding unit group, moving hydrogen production units with a delayed load reduction feedback status to the next position, and removing hydrogen production units with a non-load reduction feedback status; when the DC bus power deviation is greater than zero, increasing the load shedding advance amount by the DC bus power deviation to obtain the adjusted load shedding advance amount; when the DC bus power deviation is less than or equal to zero, taking the adjusted load shedding advance amount as the load shedding advance amount; and arranging the retained and moved hydrogen production units according to their available load reduction margin from largest to smallest according to the adjusted load shedding advance amount to obtain the hydrogen production unit load shedding sequence.

2. The hydrogen production unit load shedding method based on wind power ramping and stratification according to claim 1, characterized in that, The climbing levels are divided according to the descent pattern of the climbing segment, including: The number of consecutive descent sampling positions within the climbing segment is counted to obtain the descent duration; the power difference between the start and end positions of the climbing segment is determined as the total descent of the segment, and the maximum power decrease between adjacent sampling positions within the climbing segment is determined as the maximum single-step descent. Calculate the ratio of the maximum single-step descent to the total descent of the segment to obtain the single-step descent percentage; when the single-step descent percentage is greater than or equal to two-thirds, and there are descent sampling positions before and after the sampling position where the maximum single-step descent is located, the climbing segment is divided into a step drop layer. When the climbing segment is not divided into a step drop layer, the power drop of two adjacent segments is compared in the sampling order; when the ratio of the number of times the power drop of the latter is greater than the power drop of the former to the total number of adjacent comparisons is greater than or equal to two-thirds, and the duration of the drop is greater than or equal to the sampling length required for one load reduction response of the hydrogen production unit, the climbing segment is divided into a rapid descent climbing layer. When the climbing segment is not divided into a step drop layer or a steep descent climbing layer, the number of rising alternating positions within the climbing segment is counted; when the number of rising alternating positions is greater than or equal to 2, and the termination power value between any two adjacent rising alternating positions is lower than the starting power value, the climbing segment is divided into an oscillating fall layer. When the climbing segment is not classified into a step drop layer, a rapid descent climbing layer, or an oscillating fall layer, the climbing segment is classified into a gentle descent climbing layer.

3. The hydrogen production unit load shedding method based on wind power ramping and stratification according to claim 2, characterized in that, The climbing segment is morphologically matched with historical termination segments at the same climbing level, including: Select historical termination segments of the same climbing level from the historical climbing segment library to obtain a set of historical termination segments of the same level. The descending duration, total descent amount, maximum descent amount per step, percentage of descent per step, and number of alternating rebound positions of the climbing segment are compared item by item with the historical termination segments in the same level historical termination segment set to obtain a set of morphologically similar segments. The ratio of the descent amount from the current termination position of the climbing segment to the starting position of the climbing segment to the total descent amount of the segment is determined as the current descent ratio. In the set of similar segments, the positions in each historical termination segment that have the same descent ratio are determined as matching positions. The sampling length between the matching position and the termination position of the corresponding historical termination segment is calculated to obtain the set of remaining sampling lengths.

4. The hydrogen production unit load shedding method based on wind power ramping and stratification according to claim 3, characterized in that, Determining the remaining lifetime label of the climbing segment includes: Arrange the sampling lengths in the set of remaining sampling lengths in ascending order of value, and take the sampling length in the middle position as the representative length of the remaining lifetime; when the representative length of the remaining lifetime is less than or equal to the sampling length required for one load reduction response of the hydrogen production unit, the ramp segment is divided into short remaining lifetime tags. When the remaining lifetime representative length is greater than the sampling length required for one load reduction response of the hydrogen production unit, and less than or equal to the sampling length required for two consecutive load reduction responses of the hydrogen production unit, the ramp segment is classified as a medium remaining lifetime tag. When the remaining lifetime represents a length greater than the sampling length required for two consecutive load reduction responses of the hydrogen production unit, the ramp segment is classified as a long remaining lifetime tag.

5. The hydrogen production unit load shedding method based on wind power ramping and stratification according to claim 1, characterized in that, The method for obtaining the load shedding advance includes: Based on the start and end positions of the remaining descent period, the difference between the wind power value corresponding to the start position and the wind power value corresponding to the end position is calculated to obtain the remaining descent power. When the DC bus power deviation is greater than zero, the remaining power reduction amount is added to the DC bus power deviation to obtain the load amount to be cut off in advance; when the DC bus power deviation is less than or equal to zero, the remaining power reduction amount is used as the load amount to be cut off in advance. According to the sorting order of the candidate load shedding unit groups, the load amount to be shedding in advance is compared with the load margin that can be reduced for each hydrogen production unit in turn. When the amount of load to be cut off in advance is greater than the available load margin of the current hydrogen production unit, the available load margin of the current hydrogen production unit is used as the unit load cut-off advance amount of the current hydrogen production unit, and the amount of load to be cut off in advance is deducted from the unit load cut-off advance amount and then allocated to the next hydrogen production unit. When the amount of load to be cut off in advance is less than or equal to the available load margin of the current hydrogen production unit, the amount of load to be cut off in advance is taken as the unit cut-off advance amount of the current hydrogen production unit, and the allocation to subsequent hydrogen production units is stopped. The hydrogen production units with allocated unit load shedding advances and their unit load shedding advances are arranged to obtain the load shedding advance.

6. The hydrogen production unit load shedding method based on wind power ramping and stratification according to claim 5, characterized in that, The DC bus power deviation is the difference between the current power absorbed by the hydrogen production unit and the actual input power of the DC bus.

7. The hydrogen production unit load shedding method based on wind power ramping and stratification according to claim 1, characterized in that, The load reduction feedback status includes a load reduction status, a delayed load reduction status, and a no-load reduction status.

Images