A two-stage canal water distribution scheduling method for irrigation districts and related products
By setting a unique shared boundary flow and deviation attribution in the channel system, and using channel segment water volume-prefix aggregation mapping to form complete trajectory samples, the problems of channel segment water volume conservation and error statistics inconsistency are solved, and the accuracy and security of channel scheduling are verified.
Patent Information
- Application Number
- CN202611379211.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-07
- Publication Date
- 2026-11-17
AI Technical Summary
Existing channel scheduling methods cannot maintain water volume conservation and error statistics consistency in multi-channel systems, leading to artificial amplification of error variance and risk. Furthermore, the verification benchmarks for planned flow and achievable flow are inconsistent and cannot be directly compared.
By setting a unique shared boundary flow and deviation attribution, a complete trajectory sample is formed using channel water volume-prefix aggregation mapping, the distribution support set is determined, and the planned flow and achievable flow are checked in pairs under the same verification benchmark to ensure the conservation of channel water volume and consistency of error statistics.
This achieves consistency between water volume conservation and error statistics in the canal section, avoids the repeated accumulation of error variance and risk, ensures the accuracy and safety of the verification results, and maintains a consistent safety assessment throughout the implementation of the optimization scheme.
Smart Images

Figure CN123117686A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of water conservancy project scheduling and water resource optimization allocation technology, specifically to a two-level canal water distribution scheduling method for irrigation districts and related products. Background Technology
[0002] Irrigation districts typically consist of two-tiered canals, including a main canal and multiple branch canals along it. The dispatch center needs to determine the headwater discharge, branch canal intake, and gate control schemes based on available water supply, water demand of each branch canal, allowable water distribution time windows, canal flow capacity, and allowable operating water levels.
[0003] Existing channel scheduling methods can describe the water supply process using time-of-use flow, cumulative water distribution, channel pool water level, or channel pool storage. However, for multiple branch canals arranged sequentially along the main canal, if the upstream channel status is represented by only the cumulative water discharge from the canal head minus the cumulative water intake and losses of the preceding branch canals, without simultaneously taking into account the channel storage at the rolling start point and the outflow at the corresponding control section, the resulting balance cannot directly represent the actual water storage volume between the canal head and the control section, nor can it be directly compared with the upper and lower limits of storage corresponding to the allowable operating water level.
[0004] The shared internal control section between adjacent canal sections serves as both the outflow boundary of the upstream section and the inflow boundary of the downstream section. Setting independent internal boundary flow rates or errors for adjacent sections would disrupt the water conservation between them. Deviations in inflow, demand, gate operation, water conveyance models, and canal conveyance losses may also originate from changes in water volume at the same location. If the same deviation is repeatedly included in multiple error fields, it can artificially amplify error variance, correlation, and tail risks.
[0005] Furthermore, the actual achievable flow rate may change after the optimized planned flow rate is constrained by the gate outflow status, opening range, and maximum allowable opening variation in adjacent time periods. If the planned flow rate and achievable flow rate are evaluated using different models, reference distributions, normalization scales, or allowable risk thresholds, their safety results cannot be directly compared. Maintaining water conservation and statistical consistency in the canal section, and ensuring that the planned flow rate and achievable flow rate are verified using the same benchmark, are technical problems that need to be solved in this field. Summary of the Invention
[0006] To address the aforementioned technical problems, this invention provides a two-level canal water distribution scheduling method and related products for irrigation districts. This method maintains the water volume conservation and error statistics of canal sections with a unique shared boundary flow and a unique attribution of deviations, and ensures that planned flow and achievable flow are checked in pairs under the same verification benchmark.
[0007] This invention is achieved through the following technical solution:
[0008] A two-tiered canal water distribution scheduling method for an irrigation district includes:
[0009] Acquire data on canal system status, water demand, and operational deviations; divide canal control bodies along the main canal direction according to control sections; and determine the canal storage capacity and upper and lower limits of each canal control body.
[0010] A unique shared boundary flow is set for the internal control section shared by adjacent channel sections, and the channel storage capacity and storage upper and lower limits between the channel head and each downstream boundary control section are aggregated into prefix storage capacity and prefix storage capacity corridor, respectively.
[0011] Each operational deviation is assigned to a unique corresponding error field according to its physical source. The nominal quantity and the disturbed quantity of the same operational event or the same set of forecast members are input into the same channel section water volume-prefix aggregation mapping to obtain a complete trajectory sample.
[0012] The physical reachable trajectory set is formed by the physical disturbance domain through the water volume-prefix aggregation mapping of the channel section, and the distribution support set is determined accordingly. Based on the complete trajectory sample and the distribution support set, the risk of the prefix storage exceeding the prefix storage corridor is subject to the risk constraint of the sub-Bruker risk, and the time-sharing flow plan is obtained by solving.
[0013] The planned flow sequence corresponding to the time-sharing flow plan is converted into a gate-constrained realizable flow sequence, and the two are checked in pairs under the same verification criterion.
[0014] Only gate opening control commands that have both passed verification are issued, and the commands are updated on a rolling basis based on the actual execution results.
[0015] Optionally, the storage capacity of each channel section control body is updated according to the upstream boundary inflow, downstream boundary outflow, water intake flow of the branch canals within the channel section, and equivalent channel section loss flow.
[0016] The same shared boundary flow enters the water balance of two adjacent channel control bodies with opposite signs, as downstream boundary outflow in the upstream channel and as upstream boundary inflow in the downstream channel;
[0017] When accumulating aggregated prefixes, the shared boundary traffic within the prefix cancels out.
[0018] The upper and lower limits of storage capacity for each channel segment are retained as hard constraints for each channel segment. The prefix storage corridor is used for risk assessment and is not used as a repetitive deterministic constraint other than the hard constraints for each channel segment.
[0019] Optionally, the control sections set along the main canal from upstream to downstream are sequentially denoted as follows: , No. The control body for each canal section is located at the control section. and between;
[0020] The initial channel storage capacity, lower storage limit, and upper storage limit are determined by integrating the cross-sectional area of the water passage corresponding to the current water level, the minimum allowable operating water level, and the maximum allowable operating water level along the channel section, respectively.
[0021] No. The canal section's storage capacity meets the following requirements for each time period: In the formula, For the first The control body of the first canal section is in the Storage capacity of the canal section in a given time period , and These are the inflow at the upstream boundary, the outflow at the downstream boundary, and the water intake flow of the branch canals within the canal section, respectively. For the equivalent channel section loss flow, The duration is the length of the time period. and These are the lower limit and upper limit of storage capacity, respectively. For internal control section Shared boundary traffic at the location, ;
[0022] No. The prefix storage capacity is the sum of the storage capacity of all canal segments from the head control section to the downstream boundary control section of the prefix. The lower reference boundary and upper reference boundary of the prefix storage capacity corridor are the sum of the lower limit and the upper limit of the storage capacity of the corresponding canal segments, respectively.
[0023] Optionally, methods for assigning each operational deviation to a unique corresponding error field according to its physical source include:
[0024] For each deviation in a running event or ensemble forecast member, a deviation source and attribution record is established. The deviation source and attribution record includes the running event or ensemble forecast member identifier, physical location identifier, data collection period, data value, calculation benchmark, benchmark version, corresponding error field identifier, and attribution status.
[0025] The deviation sources and attribution records are processed in the following order: water demand target correction, gate execution residual, water conveyance model residual, and loss residual. Specifically, the water demand target correction is converted into the time-sharing branch canal water intake field after scheduling response relationship transformation. The gate execution residual is calculated based on the target gate flow rate including the time-sharing branch canal water intake correction. The water conveyance model residual of the internal control section is only entered into the shared boundary flow field of the internal control section. The loss residual is the remaining water balance deviation after deducting the attributed deviation.
[0026] Each of the aforementioned sources and attribution records is marked as assigned after the data value of the corresponding error field is formed, and is no longer used to form the data value of other error fields.
[0027] Optionally, all deviation sources and attribution records corresponding to the same complete trajectory sample adopt the same running event or the same set of forecast member identifiers, unified time period grids, unified canal system topology versions, and unified reference versions;
[0028] The canal segment water volume-prefix aggregation mapping, which consists of canal segment storage update, adjacent boundary connection, causal water conveyance, and prefix aggregation, is denoted as: The complete trajectory sample is: In the formula, This is the nominal initial state. As of the date Nominal input for each time period, and The first The first data source The initial state perturbation and input perturbation corresponding to each running event or set of forecast members. The total number of time periods included in the scheduling period. This means concatenating the output components into a complete trajectory vector according to the time period order and the preset prefix order;
[0029] Each complete trajectory sample contains the output difference of the same running event or the same set of forecast members across all prefixes and all time periods, without treating any single prefix or single time period component as an independent sample.
[0030] Optionally, the method for determining the distributed support set based on the set of physically reachable trajectories formed by the water volume-prefix aggregation mapping of the physical disturbance domain includes:
[0031] The physical disturbance domain is established based on the water level measurement error range, the boundary water inflow range, the water demand target correction range, the gate execution capacity, the applicable range of the water conveyance model, and the range of water conveyance loss parameters of the canal section. The physical disturbance domain retains the coupling relationship between each disturbance component caused by equipment capacity, water conservation, and time continuity.
[0032] When the water volume-prefix aggregation mapping of the canal section is a linear or affine mapping, the set of physically reachable trajectories is used as the distributed support set;
[0033] When the water volume-prefix aggregation mapping of the channel section is a nonlinear mapping, the conservative external approximation set that has been verified by coverage and covers the set of physically reachable trajectories will be used as the distribution support set.
[0034] Optionally, the normalization scale is determined by the corridor half-width of the prefix storage corridor, the normalization scale is used to make the complete trajectory sample and the distribution support set dimensionless, and the trajectory base distance is defined by the normalization scale.
[0035] A reference distribution is formed based on the dimensionless complete trajectory samples. Candidate probability distributions whose first-order Wasserstein distance from the reference distribution does not exceed a preset radius and whose entire probability mass is located in the dimensionless distribution support set are formed into a distribution uncertainty set. The first-order Wasserstein distance is constructed based on the trajectory basic distance.
[0036] Establish a maximum normalized two-sided superboundary function covering all prefixes, all time periods, and both sides of the corridor: ;
[0037] The subbulb risk constraint is to impose an allowable risk threshold on the subbulb conditional risk value of the maximally normalized two-sided superboundary function: ;
[0038] In the formula, For time-sharing traffic decisions, The complete trajectory after dimensionless transformation. For the first The prefix in the th ... The weight of each time period, For the first The prefix in the th ... The nominal normalized state relative to the center of the corridor for each time period Half the width of the corridor For normalization, It is a directional symbol for both sides. For the set of uncertain distributions, For candidate probability distribution, For confidence level, As an auxiliary threshold variable, It is a positive part function. To set the permissible risk threshold.
[0039] Optionally, the method for converting the planned flow sequence corresponding to the time-sharing flow plan into a gate-constrained realizable flow sequence includes:
[0040] Based on the planned flow rate of each branch canal, the predicted upstream water level, the predicted downstream water level, the gate outflow status, and the calibrated gate flow relationship, the executable opening of each branch canal in each unexecuted period is determined under the constraints of the upper and lower limits of the opening and the maximum allowable change in opening between adjacent periods, and the corresponding achievable flow rate is determined based on the executable opening.
[0041] The planned flow of all branch canals during all non-execution periods is composed of the planned flow sequence, and the corresponding achievable flow is composed of the achievable flow sequence.
[0042] Pairwise verification methods include:
[0043] Replace the corresponding branch canal water intake boundary flow in the planned flow sequence with the realized flow sequence as a whole, and re-input the canal segment water volume-prefix aggregation mapping;
[0044] The verification criteria are fixed before the start of this pairwise verification, and none of the verification criteria are changed before the completion of this pairwise verification.
[0045] The verification criteria include the canal section water volume-prefix aggregation mapping, prefix storage corridor, normalized scale, reference distribution, distribution support set and its coverage conditions, confidence level and allowable risk threshold.
[0046] Optionally, the pairwise verification is performed in the following manner:
[0047] First, verify the channel-by-channel storage capacity and the risk of the distributed bulus bar corresponding to the planned flow sequence. If the plan fails the verification, return to the time-of-use flow plan solution.
[0048] After the planning side passes the verification, the realizable flow sequence is verified using the same verification benchmark. If the realizable side fails the verification, the corresponding realizable flow range is written back to the scheduling constraints and solved again.
[0049] When the coverage conditions of the distributed support set are no longer met, or when the water level, flow rate, gate opening or outflow status corresponding to any sequence exceeds the applicable range of the distributed support set record, the verification results of the planned flow rate sequence and the realized flow rate sequence under the original verification benchmark become invalid, and a pairwise verification is performed again under the updated verification benchmark.
[0050] If either side fails the verification, the gate opening control command will not be issued.
[0051] Optionally, the method of rolling updates based on actual execution results includes:
[0052] The measured or state-assimilated storage capacity of the channel segment is used as the initial storage capacity of the channel segment for the next rolling cycle.
[0053] Update the remaining water demand of each branch canal based on the actual flow rate.
[0054] Lock the flow rate, opening degree, and water distribution for the executed period, and retain the historical boundary flow rate and model state required by the causal water conveyance model;
[0055] The newly added running deviation forms a new complete trajectory sample according to the allocation method of the error field, and the reference distribution and the distribution support set are only updated between adjacent rolling cycles.
[0056] An electronic device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the two-level canal water distribution scheduling method for irrigation districts as described above.
[0057] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the two-level canal water distribution scheduling method for irrigation districts as described above.
[0058] Compared with the prior art, the present invention has the following features and beneficial effects:
[0059] Based on the water balance of the canal section, the unique shared boundary flow and the unique attribution of deviation at the internal control section, the risk sample is consistent with the actual water conservation relationship of the canal. The internal boundary flow or the same field deviation will not be repeatedly set or injected, thus avoiding the artificial amplification of error variance, correlation and tail risk.
[0060] The initial state disturbances and process disturbances of the same operational event or the same set of forecast members are mapped to form complete trajectory samples through the water volume-prefix aggregation mapping of the same channel segment. The physical disturbance domain is mapped to a set of physically reachable trajectories and the distribution support set is determined accordingly. The risk assessment is limited to the physically reachable trajectory range. The corridor half-width of the prefix storage corridor is used to unify the normalized scale and the trajectory base distance to avoid highly overlapping prefix risks being repeatedly accumulated item by item.
[0061] The planned flow sequence and the achievable flow sequence constrained by the gate are checked in pairs using the same verification benchmark. When the verification benchmark is updated, the existing results of both sides become invalid at the same time. If either side fails, no gate opening control command will be issued, so that the final execution plan and the optimization stage are under the same safety evaluation caliber. The rolling update is based on the measured or state-assimilated channel section storage, actual execution flow, and remaining water demand, and retains the causal water transfer history. Attached Figure Description
[0062] The accompanying drawings illustrate exemplary embodiments of the present invention and, together with the description thereof, serve to explain the principles of the invention. These drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, but do not constitute a limitation on the embodiments of the present invention.
[0063] Figure 1 This is a schematic diagram of the overall process of the two-level canal water distribution and scheduling method in the irrigation area described in this invention. Detailed Implementation
[0064] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.
[0065] It should also be noted that, for ease of description, only the parts relevant to the present invention are shown in the accompanying drawings.
[0066] Where there is no conflict, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.
[0067] Example 1
[0068] This embodiment describes the overall process of the two-level canal water allocation and scheduling method in an irrigation district. For example... Figure 1 As shown, the method includes steps S1 to S6:
[0069] Step S1: Acquire data and establish a channel control system;
[0070] The scheduling objects in this embodiment include a main canal and multiple branch canals running along the main canal from upstream to downstream. The head of the main canal is designated as the control section. Other control sections shall be set up downstream of the intake of the branch canal and at the locations of sluice gates, drainage gates, or important water level monitoring points that need to be included in the water balance, and shall be sequentially recorded from upstream to downstream along the main canal as follows: .
[0071] control section and The section of the main canal between them is defined as the first Each canal section control body, among which The intake of the branch canal serves as the lateral boundary of its respective canal segment; unspecified water volume changes due to seepage, evaporation, and other factors causing a reduction in water volume within the canal segment are converted into equivalent canal segment loss flow. For the first... A prefix is used to number its downstream section. The headworks section To downstream boundary control section The set consists of all the canal segments between them. The sum of the storage capacity of each canal segment within the set is called the first segment. The prefix storage capacity of a prefix. This definition applies to situations where there are additional control gates or monitoring sections between tributaries, and does not require the number of canal segments to be the same as the number of tributaries.
[0072] The scheduling cycle is divided into There are 1 time periods, each with a length of 1 / 2. .
[0073] The acquired data includes: canal system connection and cross-sectional data; current and permissible operating water levels; flow rates at the canal head, control sections, and branch canals; gate opening, upstream and downstream water levels, and flow relationships; branch canal water demand, permissible water distribution time windows, and available water volume at the canal head; canal section water conveyance loss parameters and water conveyance model parameters; and operational deviation data from historical operation, recent rolling forecasts, and ensemble forecasts.
[0074] The channel storage capacity referred to in this embodiment is the actual water storage volume of the channel section obtained based on the relationship between water level and cross-section, rather than the difference between the cumulative water release and the cumulative water intake. The data values are measured values, state assimilation values, or ensemble prediction values.
[0075] Step S2: Configure shared boundary traffic and aggregate prefix storage;
[0076] For each internal control section shared by adjacent canal sections, only one shared boundary flow is set. The shared boundary flow enters the water balance of the control bodies of the two adjacent canal sections with opposite signs. According to the canal system topology, the canal section storage and the upper and lower limits of storage between the canal head and each downstream boundary control section are aggregated into prefix storage and prefix storage corridors, respectively.
[0077] Accordingly, the status of the upstream channel is represented by the actual water storage volume, which includes the rolling starting point storage volume and the boundary outflow, and can be directly compared with the upper and lower limits of the storage volume corresponding to the allowable operating water level.
[0078] Step S3: Generate complete trajectory samples;
[0079] Each operational deviation is assigned to a unique corresponding error field according to its physical source. The nominal quantity and the disturbed quantity of the same operational event or the same set of forecast members are input into the same channel segment water volume-prefix aggregation mapping to obtain a complete trajectory sample.
[0080] The nominal quantity includes the nominal initial state and the nominal input, and the disturbed quantity includes the disturbed initial state and the disturbed input. The unique attribution of the deviation ensures that the same on-site water volume change will not be repeatedly injected into multiple error fields.
[0081] Step S4: Construct the risk constraints of the sub-Blu-ray bar and solve the time-sharing flow plan;
[0082] The physical reachable trajectory set is formed by the physical disturbance domain through the water volume-prefix aggregation mapping of the channel section, and the distribution support set is determined accordingly. Based on the complete trajectory sample and the distribution support set, the risk of the prefix storage exceeding the prefix storage corridor is subject to the risk constraint of the sub-Blu-ray risk constraint. Under the constraints of water demand, allowable water distribution time window, flow capacity, upper and lower limits of channel storage and gate capacity, the time-sharing flow plan is obtained.
[0083] Step S5: Convert the planned flow sequence into an achievable flow sequence and verify them in pairs using the same verification benchmark;
[0084] The planned flow sequence corresponding to the time-sharing flow plan is converted into a gate-constrained realizable flow sequence, and the two are checked in pairs under the same verification criterion.
[0085] Step S6: Issue gate opening control commands and update them on a rolling basis.
[0086] Only gate opening control commands that have both passed verification are issued, and the next cycle is updated based on the actual execution results after execution.
[0087] Example 2
[0088] Based on Example 1, this embodiment illustrates the method for determining the storage capacity of a channel segment, the water balance of the channel segment, the unique shared boundary flow, and the aggregation relationship of the prefix storage capacity.
[0089] Let the first The upstream and downstream coordinates of each canal section are as follows: and The canal section is at coordinates Location, water level elevation is The cross-sectional area of the water flow at that time is The measured or state-assimilated water level elevation at the starting point of the rolling process is recorded as follows: The minimum and maximum permissible operating water level elevations are respectively denoted as and ;
[0090] The cross-sectional areas of the three corresponding to the water passage are along to Integrating the same channel segment range, the storage capacity of the starting channel segment is obtained respectively. Lower limit of storage and storage limit .
[0091] Regular cross sections can use analytical area relationships, while complex cross sections can establish segmented area relationships based on measured cross section points; the three types of storage capacity use the same channel range and water level elevation benchmark.
[0092] No. The first section of the canal The water balance for each time period, the unique shared boundary relationship between adjacent canal sections, and the hard constraint on storage capacity for each canal section are as follows: In the formula, and These are the upstream boundary inflow and the downstream boundary outflow, respectively. The water intake flow rate of the branch canals within the canal section. For the equivalent channel section loss flow; , For internal control section The unique shared boundary flow is shared by adjacent canal sections.
[0093] The same It is treated as an outflow in the upstream water balance and as an inflow in the downstream water balance; only this one boundary coordinate is set for both nominal calculation and disturbed calculation.
[0094] The boundary flow of the canal section can be determined by a hydrodynamic model based on the one-dimensional Saint-Venant equations, an integral time delay model, a state-space model, or other causal water conveyance models calibrated by actual measurements.
[0095] No. The output for each time period only uses the boundary inflow, branch canal water intake, canal storage status, and water conveyance model parameters of the current and historical time periods, and does not use information from future time periods.
[0096] The prefix storage and the reference boundary of the prefix storage corridor are determined by the set The storage capacity of each canal segment within the set and its upper and lower limits are determined by summing them separately; the water balance of each canal segment within the set is summed, and the shared boundary flows within the prefix cancel each other out. The above relationships are uniformly expressed as: ,in, For the first The prefix in the th ... Prefix accumulation for each time period and These are the lower and upper reference boundaries of the prefix storage corridor, respectively, used to determine the normalized scale and the two-sided over-limit threshold, and are not repeated as deterministic hard constraints in addition to the upper and lower limits of storage for each channel segment. For the headwaters to flow in at the boundary of the canal, For the first The prefix is at the downstream boundary control section. Outflow at the boundary and satisfying ;when At that time, the downstream boundary control section is an internal control section and satisfies the following conditions: .
[0097] The above relationship indicates that the prefix storage includes the rolling start state and the outflow from the downstream boundary control section, and is not the balance obtained by deducting the cumulative water intake and losses from the cumulative water release at the head of the canal.
[0098] To further illustrate the computational relationships between unique shared boundary flow, channel water conservation, prefix aggregation, and error mapping from the same source, a specific example is given below.
[0099] Take two adjacent canal sections, time period length for .
[0100] The initial storage capacity of the first canal section is The upstream inflow, downstream outflow, branch canal intake flow, and equivalent canal section loss flow are respectively , , and ;
[0101] The initial storage capacity of the second canal section is Its upstream inflow and downstream outflow of the first channel section share the same... The downstream outflow, the branch canal intake flow, and the equivalent canal section loss flow are respectively , and .
[0102] Based on the aforementioned water balance calculation for each canal segment, the storage capacity of the first and second canal segments at the end of the first time period is... and They are respectively and corresponding prefix storage for .
[0103] In the above calculations, the internal boundary flow It is set only once, as an outflow in the first channel segment and an inflow in the second channel segment; the summation of the two channel segments cancels each other out. If the lower and upper storage limits of the first channel segment are respectively... and The lower and upper limits of the storage capacity for the second canal section are respectively and The corresponding prefix reference interval is obtained by adding the two sets of limits and is used for normalizing the scale and the two-sided over-limit threshold, and is not used as a separate deterministic constraint.
[0104] Further, let the error of the internal boundary shared flow be... If the signs of the storage error in adjacent canal sections are opposite, then: In the formula, and These are the storage errors in the first and second channel segments caused by the shared boundary flow error, respectively. This is the second prefix storage error caused by the shared boundary flow error.
[0105] Therefore, the same internal boundary flow error is generated only once and cancels out within the prefix; only the boundary flow error that crosses the downstream boundary control section of the prefix is retained.
[0106] Example 3
[0107] This embodiment illustrates the unique attribution rules for operational deviations and the process of generating complete trajectory samples.
[0108] The nominal initial state is formed by the nominal storage capacity of each channel segment at the starting point of the rolling process, and the nominal input is formed by the shared boundary flow of the control section at each time period, the water intake flow of the branch canals within the channel segment, and the equivalent channel segment loss flow.
[0109] The flow rate at the internal control section occupies only one coordinate in the input; the internal boundary flow rate generated by the canal section water conveyance model is no longer repeatedly sampled as an independent exogenous disturbance.
[0110] For each deviation in each running event or ensemble forecast member, a deviation source and attribution record is established. The deviation source and attribution record includes the running event or ensemble forecast member identifier, physical location identifier, data acquisition period, data value, calculation baseline, baseline version, corresponding error field identifier, and attribution status.
[0111] Deviations are processed in the following order: water demand target correction, gate execution residual, water conveyance model residual, and loss residual. Water demand target correction first changes the remaining water demand or the cumulative final value of water allocation, and only after conversion by the preset scheduling response relationship is it entered into the time-sharing branch canal water intake field; gate execution residual is calculated based on the target gate flow rate that already includes the time-sharing branch canal water intake correction; the water conveyance model residual of the internal control section is only entered into the shared boundary flow field of the internal control section, and the difference when the simplified water conveyance model is inconsistent with the water balance of the canal section is also included in this field; loss residual is only the remaining water balance deviation after deducting the first three types of deviations.
[0112] Once a deviation source and attribution record generates the corresponding error field data value, it is marked as assigned and will not be used to generate other error fields. Records corresponding to the same complete trajectory sample use the same running event or the same ensemble forecast member identifier, the same time period grid, the same canal system topology version, and the same reference version. Scalar deviations from different running events, different ensemble forecast members, or different reference versions are not concatenated into the same complete trajectory sample.
[0113] When a missing measurement occurs, only records that have been restored to a uniform dimension and marked with data state after state assimilation can be included in the corresponding complete trajectory sample; records that cannot be restored are excluded from the cause and are not treated as independent scalar samples.
[0114] This embodiment and Figure 1 The complete prefix-time period trajectory from the same source, as mentioned above, refers to the complete trajectory sample, which is a trajectory formed by the same running event or the same ensemble forecast member under a unified time period grid, a unified canal system topology version, and a unified baseline version, covering all prefixes and all time periods, and not just trajectories from the same data source. It will be from the rolling start point to the... The mapping of sequentially executing channel segment storage update, adjacent boundary connection, causal water conveyance, and prefix aggregation in each time period is denoted as follows: This is called the canal section water volume-prefix aggregation mapping.
[0115] For the The first data source For each running event or ensemble forecast member, its corresponding set of deviation sources and attribution records, forming the nominal initial state and input, and the disturbed initial state and input, are respectively fed into the same mapping to obtain: In the formula, This represents the nominal initial storage vector for the channel section. As of the date Nominal input for each time period, and These are the initial state perturbation and input perturbation for the corresponding running event or ensemble forecast member, respectively. The total number of time periods included in the scheduling period. Indicates according to The time periods are ordered, and the output components are concatenated into a complete trajectory vector according to the preset prefix order.
[0116] Each All prefix and time-period error components containing the same running event or the same ensemble of forecast members are included as a single sample unit in the reference distribution construction, without treating individual prefixes, individual time periods, or scalar errors as independent samples. For nonlinear models, the trajectory component is formed by the difference between the two model outputs of the perturbed input and the nominal input.
[0117] Example 4
[0118] This embodiment illustrates the construction of the physical disturbance domain and distributed support set, the normalization process, and the form of the distributed bar risk constraint, and provides a solution method for time-sharing flow planning.
[0119] The initial accumulation disturbance and the input disturbances at each time period are combined to form a physical disturbance vector. The range of water level measurement or state assimilation error, the range of inflow at the head of the canal, the range of correction for the water demand target of the branch canals, the gate execution capacity, the applicable range of the water conveyance model, and the range of water conveyance loss parameters in the canal section jointly define the component range and coupling relationship of the physical disturbance vector; all physical disturbance vectors that satisfy the component range and coupling relationship, the uniqueness of the shared boundary flow, the unique attribution of the deviation, and the cross-time period physical constraints constitute the physical disturbance domain, denoted as . .
[0120] The initial water level error is converted into storage disturbance using the same water level-cross-sectional area relationship as the initial storage; the remaining errors in the water balance entering the canal section are expressed in flow rate units and multiplied by the time period length when the storage is updated. The physical disturbance domain simultaneously preserves the coupling relationship between equipment capacity, water conservation, and time continuity, and does not arbitrarily construct independent box supports based on the sample extreme values of each trajectory component.
[0121] Physical disturbance The complete trajectory formed by the water volume-prefix aggregation mapping of the same canal segment is denoted as The physical perturbation domain is mapped as described above to obtain the set of physically reachable trajectories, which is then used as the distribution support set. .
[0122] When the channel segment model is a linear or affine model, the set of physically reachable trajectories or its convex representation is used as the distributed support set;
[0123] When the channel segment model is a nonlinear model, the conservative external approximation set that has been verified by coverage and covers the set of physically reachable trajectories will be used as the distributed support set.
[0124] The final distributed support set adopted is denoted as .
[0125] The conservative external approximation set records the applicable water level, flow rate, gate opening, and outflow state range and error margin, and remains fixed during a pairwise verification; when the planned flow rate or achievable flow rate exceeds the applicable range, it is processed according to the verification benchmark update rule described in Example 5.
[0126] After obtaining the complete trajectory samples and distribution support set, the normalized scale is determined based on the corridor half-width of the prefix accumulation corridor. The corridor half-width reflects the allowable accumulation range of different prefixes in the corresponding time period.
[0127] Let the half-width of the corridor be denoted as The normalized scale that is not less than half the width of the corridor is denoted as . The complete trajectory samples and distribution support sets are transformed to dimensionless coordinates using the normalized scale, and the trajectory base distance is defined using this scale: In the formula, For the first The prefix in the th ... The corridor is half the width during each time period. The smallest positive scale. For normalization, This corresponds to the scaling factor. It is a diagonal matrix composed of the prefix-time scale coefficients. and Support set for unnormalized distribution Any two complete prefixes within the time period trajectory, and For the same After transformation, it is located in the normalized distribution support set. The complete trajectory within, This represents the trajectory-based distance between two normalized complete trajectories. express Norm, The norm order is 1 or 2. The same volumetric error has a larger normalization effect in prefixes with narrower storage corridors; the ratio of corridor half-width to normalized scale. Used to determine the upper and lower boundary thresholds.
[0128] Reference distribution Formed from the empirical distribution of normalized complete trajectory samples, each complete trajectory is considered a sample unit, without splitting its prefix and time components. The first-order Wasserstein distance constructed based on the aforementioned trajectory fundamental distance is denoted as... The radius of the Wasserstein sphere is determined based on either independently reserved trajectories or block resampling that maintains time correlation; when resampling is used, the complete trajectory is taken as the smallest resampling unit.
[0129] As mentioned above As a normalized distribution support set, and within the normalized distribution support set, a distribution uncertainty set is established: In the formula, For a set with uncertain distribution, To support Candidate probability distribution on Let be the radius of the Wasserstein sphere. This definition requires that all probability masses of the candidate distribution lie within the normalized distribution support set, excluding trajectories formed by arbitrarily splicing together error components outside the distribution support set.
[0130] Let the time-sharing traffic decision be denoted as The decision corresponding to the first The prefix in the th ... The prefix accumulation for each time period is denoted as The corresponding corridor center and nominal normalized state are respectively denoted as and Based on this, a maximum normalized two-sided superboundary function covering the entire prefix, all time periods, and both sides of the corridor is established: , In the formula, To normalize the complete trajectory The Middle The prefix in the th ... The weight of each time period, These are bilateral directional symbols, among which Corresponding to the upper boundary of the corridor, Corresponding to the lower boundary of the corridor, It is the maximum normalized two-sided superboundary function.
[0131] when At that time, all prefixes did not exceed the corresponding corridor in any time period. Using a single maximum normalized two-sided superboundary function instead of summing each overlapping prefix can avoid the repeated accumulation of deviations in the same upstream channel segment.
[0132] At confidence level Next, apply the Distributed Robust Conditional Value-at-Risk (DR-CVaR) constraint to the maximum normalized two-sided superboundary function: In the formula, To divide the risk value of the blob, For confidence level, For the set of real numbers, Represents the auxiliary threshold variable for real numbers Remove the boundary. Represents a set with an uncertain distribution The candidate probability distribution takes the worst-case supremum. Represents the mathematical expectation. This indicates that the normalized complete trajectory follows a candidate probability distribution. , Denotes the positive part function, To set the permissible risk threshold.
[0133] The sub-Brook bar risk constraint referred to in this embodiment refers to the constraint that evaluates the risk of the prefix storage exceeding the prefix storage corridor for the worst-case distribution within the distribution uncertainty set and applies a threshold limit, including but not limited to the sub-Brook bar conditional risk value constraint applied to the maximum normalized two-sided superboundary function; in other embodiments, the sub-Brook bar chance constraint or worst-case expectation constraint may also be used within the same distribution uncertainty set.
[0134] When the trajectory base distance adopts the first norm and the distribution support set is represented by a bounded polyhedron, the risk value constraint of the sub-Bruker condition can be transformed into a finite linear constraint through the duality of standard linear programming; when the second norm is adopted, it can be transformed into a second-order cone constraint.
[0135] Under the premise of satisfying the water demand of branch canals, the allowable water distribution time window, the flow capacity of the head and branch canals, the water balance of the canal section, the upper and lower limits of the storage capacity of each canal section, the gate capacity, and the above-mentioned allowable risk thresholds, the time-sharing flow plan is obtained by taking at least one of the following as the scheduling objectives: undersupply or oversupply, water transmission loss, and flow change.
[0136] Example 5
[0137] This embodiment illustrates the determination of achievable flow rates and the pairwise verification of planned flow rate sequences and achievable flow rate sequences.
[0138] After completing the time-sharing flow plan solution, the planned flow of all branch canals during all unexecuted periods is composed into a planned flow sequence.
[0139] Let the first The branch canal in the first The planned traffic for each time period is The predicted water levels upstream and downstream of the sluice gate are respectively and The current relationship of the gate, as determined by calibration, is .
[0140] Determine the feasible flow opening range based on free outflow, submerged outflow, or other actual outflow conditions. Then, select the feasible flow opening that minimizes the flow deviation, constrained by the upper and lower limits of the flow opening and the maximum allowable change in flow opening between adjacent time periods. Calculate the achievable flow rate. In the formula, For the first The branch canal in the first The set of feasible openness for each time period The candidate apertures in the set of feasible apertures, For executable opening, and These are the lower and upper limits of the opening, respectively. This represents the maximum permissible change in opening degree between adjacent time periods. This represents the executable opening level of the previous time period; for the first unexecuted time period, Take the actual opening degree of the previous executed time period.
[0141] When multiple feasible gate openings exist that achieve the same flow deviation, the opening with the smallest change relative to the previous control period is preferred; this solution does not require the gate flow relationship to have a globally unique inverse function. All branch canals are considered during all unexecuted periods. This can form a flow sequence.
[0142] A preset flow execution tolerance is set; when the absolute value of the difference between the achievable flow and the planned flow in the corresponding branch channel during the corresponding time period does not exceed this tolerance, the flow execution tolerance is deemed to be met for the corresponding time period.
[0143] Before each pairwise verification of the planned flow sequence and the achievable flow sequence begins, the following parameters are fixed for this pairwise verification: channel water volume-prefix aggregation mapping, prefix storage corridor, corridor half-width, normalized scale and scale matrix, reference distribution, Wasserstein sphere radius, set of physically reachable trajectories, distribution support set and its coverage conditions, confidence level and allowable risk threshold. These parameters are used as the verification benchmark for this pairwise verification.
[0144] Before this pairwise verification is completed, none of the verification criteria will be changed individually based on whether one side passes or fails. The flow execution tolerance is only used as a verification criterion for the feasible side and remains unchanged during this pairwise verification.
[0145] Paired checks are performed in the following order:
[0146] 1. Use the aforementioned verification benchmark to verify the channel-by-channel storage capacity and sub-branch risk corresponding to the planned flow sequence; if the plan fails, return to the time-of-use flow plan solution.
[0147] 2. After the planned flow is approved, the corresponding branch canal water intake boundary flow in the planned flow sequence is replaced with the realized flow sequence as a whole. The same canal segment water volume-prefix aggregation mapping is re-entered, and the realized flow is verified using the same verification benchmark.
[0148] 3. If the achievable flow range does not meet the upper and lower limits of storage capacity per channel segment, the allowable risk threshold, or the flow execution tolerance, the corresponding achievable flow range will be written back to the scheduling constraints and solved again.
[0149] 4. When the coverage conditions of the distributed support set are no longer met, or the water level, flow rate, gate opening or outflow status corresponding to the planned flow sequence or the achievable flow sequence exceeds the applicable range of the distributed support set record, the verification results on both sides under the original verification benchmark become invalid at the same time, and are re-verified in the same order under the updated verification benchmark.
[0150] Do not compare the results from both sides under different verification criteria.
[0151] 5. If either side fails to pass the check, no gate opening control command will be issued; only the executable opening that passes the check on both sides will be used as the gate opening control command for this rolling cycle.
[0152] If there is no feasible opening degree that meets the equipment capacity and gate outflow conditions, the output state is infeasible.
[0153] Example 6
[0154] This embodiment illustrates the rolling update after the gate opening control command is executed.
[0155] After the gate opening control command is executed, the branch canal water intake flow rate confirmed based on the actual flow measurement data or state assimilation results will be recorded as the actual executed flow rate. .
[0156] The actual flow rate is used to calculate the actual water distribution volume, update the remaining water demand, and generate new operational deviation records. It does not include the achievable flow rate calculated based on the predicted water level and the executable opening degree. This is directly considered as actual traffic that has already been executed.
[0157] In the The actual time when the next rolling update occurs The canal storage capacity is updated based on the measured water level, boundary flow, and state assimilation results. Measured or state-assimilated storage capacity of each canal section As the first Initial channel storage capacity for each rolling cycle .
[0158] For the A branch canal, based on the target water demand. Deducting up to the actual time According to the set of executed time periods Actual execution traffic and duration The difference between the accumulated actual water distribution and the actual water volume is calculated as zero if the difference is less than zero, thus yielding the remaining water demand. .
[0159] The flow rate, opening degree, and water distribution volume of the executed period remain locked and will not re-enter the decision sequence of the unexecuted period.
[0160] For water conveyance models of canal sections with memory characteristics, the actual time is retained. The previously preset boundary flow, lateral water intake, equivalent channel loss flow, gate opening and operating status, and internal model status are used, and the actual opening and actual flow of the previous executed period are used as the benchmark values for the first control variable change constraint of the next rolling cycle.
[0161] New operational deviations are assigned to an error field according to the aforementioned deviation sources and attribution rules, and then entered into the sample cache according to the generation rules for complete trajectory samples. Only after a complete prefix-time period trajectory with a unified dimension is formed is it used as a new sample to participate in the reference distribution update; single time periods or incomplete trajectories are not treated as independent samples.
[0162] The reference distribution, Wasserstein sphere radius, and distribution support set are updated only between adjacent rolling cycles; the calibration benchmark is not updated until a pairwise check has been completed.
[0163] Example 7
[0164] This embodiment provides a two-level canal water distribution and scheduling system for an irrigation district, including a data acquisition module, a canal segment modeling module, a trajectory and risk module, a scheduling verification module, and a rolling update module.
[0165] The data acquisition module is used to acquire data on canal system status, water demand, and operational deviations, including canal system connection and cross-sectional data, water level and flow monitoring data, gate opening and operating status, branch canal water demand, allowable water distribution time windows, and operational deviation data from historical operation, recent rolling forecasts, and ensemble forecasts.
[0166] The canal segment modeling module is connected to the data acquisition module and is used to divide the canal segment control body along the main canal, determine the canal segment storage capacity and its upper and lower limits, set a unique shared boundary flow for the internal control section shared by adjacent canal segment control bodies, and form a prefix storage capacity and a prefix storage capacity corridor.
[0167] The trajectory and risk module is used to assign each operational deviation to a unique corresponding error field, generate complete trajectory samples and a set of physically reachable trajectories, and determine the distribution support set.
[0168] The scheduling verification module is used to solve the time-sharing flow plan under the risk constraint of the distributed bloc, convert the planned flow sequence into a realizable flow sequence subject to gate constraints, perform pairwise verification on the two under the same verification benchmark, and output the gate opening control command that passes the verification.
[0169] The rolling update module is used to update the next rolling cycle based on the measured or state-assimilated channel storage, actual flow rate, and remaining water demand.
[0170] The specific processing procedures executed by each module are implemented in accordance with the methods described in Examples 1 to 6. Each module can be implemented by a software program running on a scheduling server or industrial control computer, or it can be implemented by combining software with field acquisition and control hardware.
[0171] Example 8
[0172] This embodiment provides an electronic device, including a processor, a memory, a communication interface, and a field control interface. The memory stores a computer program and data required for implementing the aforementioned two-level irrigation canal water allocation and scheduling method; when the processor executes the computer program, the electronic device performs the aforementioned two-level irrigation canal water allocation and scheduling method.
[0173] Memory is used to store software programs and modules. The processor executes various terminal functions and data processing by running the software programs and modules stored in memory. Memory can mainly consist of a program storage area and a data storage area. The program storage area can store the operating system, at least one executable program required for a given function, etc.
[0174] The storage data area can store data created based on the use of the terminal. Furthermore, the memory can include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory, or other volatile solid-state storage devices.
[0175] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, it causes the processor to perform the above-described two-level canal water distribution scheduling method for irrigation districts.
[0176] Without loss of generality, computer-readable media can include computer storage media and communication media. Computer storage media includes volatile and non-volatile, removable and non-removable media implemented using any method or technology for storing information such as computer-readable instruction data structures, program modules, or other data. Computer storage media includes RAM, ROM, EPROM, EEPROM, flash memory or other solid-state storage technologies, CD-ROM, DVD or other optical storage, magnetic tape cassettes, magnetic tape, disk storage, or other magnetic storage devices. Of course, those skilled in the art will recognize that computer storage media are not limited to the above-mentioned types. The aforementioned system memories and mass storage devices can be collectively referred to as memory.
[0177] Those skilled in the art should understand that the above embodiments are merely for illustrating the present invention and are not intended to limit the scope of the invention. Those skilled in the art can make other changes or modifications based on the above invention, and these changes or modifications still fall within the scope of the present invention.
Claims
1. A two-stage canal water distribution and scheduling method for an irrigation district, characterized in that, include: Acquire data on canal system status, water demand, and operational deviations; divide canal control bodies along the main canal direction according to control sections; and determine the canal storage capacity and upper and lower limits of each canal control body. A unique shared boundary flow is set for the internal control section shared by adjacent channel sections, and the channel storage capacity and storage upper and lower limits between the channel head and each downstream boundary control section are aggregated into prefix storage capacity and prefix storage capacity corridor, respectively. Each operational deviation is assigned to a unique corresponding error field according to its physical source. The nominal quantity and the disturbed quantity of the same operational event or the same set of forecast members are input into the same channel section water volume-prefix aggregation mapping to obtain a complete trajectory sample. The physical reachable trajectory set is formed by the physical disturbance domain through the water volume-prefix aggregation mapping of the channel section, and the distribution support set is determined accordingly. Based on the complete trajectory sample and the distribution support set, the risk of the prefix storage exceeding the prefix storage corridor is subject to the risk constraint of the sub-Bruker risk, and the time-sharing flow plan is obtained by solving. The planned flow sequence corresponding to the time-sharing flow plan is converted into a gate-constrained realizable flow sequence, and the two are checked in pairs under the same verification criterion. Only gate opening control commands that have both passed verification are issued, and the commands are updated on a rolling basis based on the actual execution results.
2. The two-level canal water distribution and scheduling method for irrigation districts according to claim 1, characterized in that, Each control body of the canal section updates the canal section storage capacity according to the inflow at the upstream boundary, the outflow at the downstream boundary, the water intake flow of the branch canals within the canal section, and the equivalent canal section loss flow. The same shared boundary flow enters the water balance of two adjacent channel control bodies with opposite signs, as downstream boundary outflow in the upstream channel and as upstream boundary inflow in the downstream channel; When accumulating aggregated prefixes, the shared boundary traffic within the prefix cancels out. The upper and lower limits of storage capacity for each channel segment are retained as hard constraints for each channel segment. The prefix storage corridor is used for risk assessment and is not used as a repetitive deterministic constraint other than the hard constraints for each channel segment.
3. The two-level canal water distribution and scheduling method for irrigation districts according to claim 2, characterized in that, The control sections set along the main canal from upstream to downstream are sequentially denoted as follows: , No. The control body for each canal section is located at the control section. and between; The initial channel storage capacity, lower storage limit, and upper storage limit are determined by integrating the cross-sectional area of the water passage corresponding to the current water level, the minimum allowable operating water level, and the maximum allowable operating water level along the channel section, respectively. No. The canal section's storage capacity meets the following requirements for each time period: In the formula, For the first The control body of the first canal section is in the Storage capacity of the canal section in a given time period , and These are the inflow at the upstream boundary, the outflow at the downstream boundary, and the water intake flow of the branch canals within the canal section, respectively. For the equivalent channel section loss flow, The duration is the length of the time period. and These are the lower limit and upper limit of storage capacity, respectively. For internal control section Shared boundary traffic at the location, ; No. The prefix storage capacity is the sum of the storage capacity of all canal segments from the head control section to the downstream boundary control section of the prefix. The lower reference boundary and upper reference boundary of the prefix storage capacity corridor are the sum of the lower limit and the upper limit of the storage capacity of the corresponding canal segments, respectively.
4. The two-level canal water distribution and scheduling method for irrigation districts according to claim 1, characterized in that, Methods for assigning various operational deviations to a unique corresponding error field based on their physical source include: For each deviation in a running event or ensemble forecast member, a deviation source and attribution record is established. The deviation source and attribution record includes the running event or ensemble forecast member identifier, physical location identifier, data collection period, data value, calculation benchmark, benchmark version, corresponding error field identifier, and attribution status. The deviation sources and attribution records are processed in the following order: water demand target correction, gate execution residual, water conveyance model residual, and loss residual. Specifically, the water demand target correction is converted into the time-sharing branch canal water intake field after scheduling response relationship transformation. The gate execution residual is calculated based on the target gate flow rate including the time-sharing branch canal water intake correction. The water conveyance model residual of the internal control section is only entered into the shared boundary flow field of the internal control section. The loss residual is the remaining water balance deviation after deducting the attributed deviation. Each of the aforementioned sources and attribution records is marked as assigned after the data value of the corresponding error field is formed, and is no longer used to form the data value of other error fields.
5. The two-level canal water distribution and scheduling method for irrigation districts according to claim 4, characterized in that, All deviation sources and attribution records corresponding to the same complete trajectory sample adopt the same operational event or the same set of forecast member identifiers, unified time period grids, unified canal system topology versions, and unified reference versions; The canal segment water volume-prefix aggregation mapping, which consists of canal segment storage update, adjacent boundary connection, causal water conveyance, and prefix aggregation, is denoted as: The complete trajectory sample is: In the formula, This is the nominal initial state. As of the date Nominal input for each time period, and The first The first data source The initial state perturbation and input perturbation corresponding to each running event or set of forecast members. The total number of time periods included in the scheduling period. This means concatenating the output components into a complete trajectory vector according to the time period order and the preset prefix order; Each complete trajectory sample contains the output difference of the same running event or the same set of forecast members across all prefixes and all time periods, without treating any single prefix or single time period component as an independent sample.
6. The two-level canal water distribution and scheduling method for irrigation districts according to claim 1, characterized in that, The method for determining the distributed support set based on the set of physically reachable trajectories formed by the water volume-prefix aggregation mapping of the physical disturbance domain includes: The physical disturbance domain is established based on the water level measurement error range, the boundary water inflow range, the water demand target correction range, the gate execution capacity, the applicable range of the water conveyance model, and the range of water conveyance loss parameters of the canal section. The physical disturbance domain retains the coupling relationship between each disturbance component caused by equipment capacity, water conservation, and time continuity. When the water volume-prefix aggregation mapping of the canal section is a linear or affine mapping, the set of physically reachable trajectories is used as the distributed support set; When the water volume-prefix aggregation mapping of the channel section is a nonlinear mapping, the conservative external approximation set that has been verified by coverage and covers the set of physically reachable trajectories will be used as the distribution support set.
7. The two-level canal water distribution and scheduling method for irrigation districts according to claim 6, characterized in that, The normalized scale is determined by the corridor half-width of the prefix storage corridor, the normalized scale is used to make the complete trajectory sample and the distribution support set dimensionless, and the trajectory base distance is defined by the normalized scale. A reference distribution is formed based on the dimensionless complete trajectory samples. Candidate probability distributions whose first-order Wasserstein distance from the reference distribution does not exceed a preset radius and whose entire probability mass is located in the dimensionless distribution support set are formed into a distribution uncertainty set. The first-order Wasserstein distance is constructed based on the trajectory basic distance. Establish a maximum normalized two-sided superboundary function covering all prefixes, all time periods, and both sides of the corridor: ; The subbulb risk constraint is to impose an allowable risk threshold on the subbulb conditional risk value of the maximally normalized two-sided superboundary function: ; In the formula, For time-sharing traffic decisions, The complete trajectory after dimensionless transformation. For the first The prefix in the th ... The weight of each time period, For the first The prefix in the th ... The nominal normalized state relative to the center of the corridor for each time period Half the width of the corridor For normalization, It is a directional symbol for both sides. For the set of uncertain distributions, For candidate probability distribution, For confidence level, As an auxiliary threshold variable, It is a positive part function. To set the permissible risk threshold.
8. The two-level canal water distribution and scheduling method for irrigation districts according to claim 1, characterized in that, The method for converting the planned flow sequence corresponding to the time-sharing flow plan into a gate-constrained realizable flow sequence includes: Based on the planned flow rate of each branch canal, the predicted upstream water level, the predicted downstream water level, the gate outflow status, and the calibrated gate flow relationship, the executable opening of each branch canal in each unexecuted period is determined under the constraints of the upper and lower limits of the opening and the maximum allowable change in opening between adjacent periods, and the corresponding achievable flow rate is determined based on the executable opening. The planned flow of all branch canals during all non-execution periods is composed of the planned flow sequence, and the corresponding achievable flow is composed of the achievable flow sequence. Pairwise verification methods include: Replace the corresponding branch canal water intake boundary flow in the planned flow sequence with the realized flow sequence as a whole, and re-input the canal segment water volume-prefix aggregation mapping; The verification criteria are fixed before the start of this pairwise verification, and none of the verification criteria are changed before the completion of this pairwise verification. The verification criteria include the canal section water volume-prefix aggregation mapping, prefix storage corridor, normalized scale, reference distribution, distribution support set and its coverage conditions, confidence level and allowable risk threshold.
9. The two-level canal water distribution and scheduling method for irrigation districts according to claim 8, characterized in that, The pairwise verification is performed in the following manner: First, verify the channel-by-channel storage capacity and the risk of the distributed bulus bar corresponding to the planned flow sequence. If the plan fails the verification, return to the time-of-use flow plan solution. After the planning side passes the verification, the realizable flow sequence is verified using the same verification benchmark. If the realizable side fails the verification, the corresponding realizable flow range is written back to the scheduling constraints and solved again. When the coverage conditions of the distributed support set are no longer met, or when the water level, flow rate, gate opening or outflow status corresponding to any sequence exceeds the applicable range of the distributed support set record, the verification results of the planned flow rate sequence and the realized flow rate sequence under the original verification benchmark become invalid, and a pairwise verification is performed again under the updated verification benchmark. If either side fails the verification, the gate opening control command will not be issued.
10. The two-level canal water distribution and scheduling method for irrigation districts according to claim 1, characterized in that, Methods that update on a rolling basis based on actual execution results include: The measured or state-assimilated storage capacity of the channel segment is used as the initial storage capacity of the channel segment for the next rolling cycle. Update the remaining water demand of each branch canal based on the actual flow rate. Lock the flow rate, opening degree, and water distribution for the executed period, and retain the historical boundary flow rate and model state required by the causal water conveyance model; The newly added running deviation forms a new complete trajectory sample according to the allocation method of the error field, and the reference distribution and the distribution support set are only updated between adjacent rolling cycles.
11. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the two-level canal water distribution scheduling method for irrigation districts as described in any one of claims 1 to 10.
12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the two-level canal water distribution scheduling method for irrigation districts as described in any one of claims 1 to 10.
