A green electricity intelligent deployment method and system for a cold chain park and a highway network

Abstract

The application discloses a kind of green electricity intelligent deployment method and system of cold chain park and highway network cooperation, it is related to energy time series data processing and rolling optimization scheduling technical field, including obtaining cold chain park and highway network cooperative operation data, constructs unified time slice scheduling input vector.Based on scheduling input vector, calculate temperature virtual energy storage state quantity, calculate temperature virtual energy storage schedulable boundary.Under schedulable boundary, execute rolling time domain cooperative optimization solution, generate and issue green electricity intelligent deployment instruction.The application constructs computable scheduling input with unified time service and confidence identifier, maps cold chain temperature control margin to temperature virtual energy storage state quantity and forms schedulable boundary, then performs rolling cooperative optimization under the boundary and closed-loop issues instruction, realizes the intelligent deployment of solvable, executable, traceable under green electricity fluctuation, electricity price change and operation event disturbance across park and service area.

Description

Technical Field

[0001] This invention relates to the field of energy time-series data processing and rolling optimization scheduling technology, specifically to a method and system for intelligent green electricity allocation that coordinates cold chain industrial parks and highway networks. Background Technology

[0002] In recent years, cold chain industrial parks have become highly dependent on refrigeration and temperature control loads in warehousing, distribution, and trunk transportation. Highway service areas also experience significant time-varying peak electricity consumption due to the refueling needs of new energy vehicles. With the increasing deployment of distributed photovoltaic or wind power and energy storage in these parks, energy management is gradually evolving from equipment-level control to data-driven collaborative scheduling. This includes the widespread adoption of unified time synchronization, multi-source time-series data fusion, prediction, and rolling optimization—digital data processing methods that enable the coordinated calculation and control of park load, green electricity output, and electricity price signals on computer equipment. This improves the local consumption and operational controllability of renewable energy.

[0003] Existing technologies still have several shortcomings. Cold chain parks and highway service areas are often operated independently by different platforms, with inconsistent data sampling frequencies, time scales, and quality standards. There is a lack of a unified time-slicing and confidence identification mechanism for scheduling computation, leading to unstable optimization inputs and difficulties in closed-loop cross-entity collaboration. Most methods treat cold storage or vehicle refrigeration as rigid loads or rely solely on empirical rules for pre-cooling, failing to transform the allowable temperature range and thermal inertia into optimizable state variables, capacity, and boundary constraints. This makes it difficult to input temperature control compliance constraints into the rolling solver in a computable boundary form. The linkage between charging loads in highway service areas and cooling loads in the park typically employs static quotas or separate optimization, lacking a multi-time-slice collaborative solution and instruction feedback closed-loop mechanism under power distribution capacity, power upper and lower limits, and ramping constraints. This makes it difficult to maintain scheduling feasibility and stability under green electricity fluctuations, electricity price changes, and operational event disturbances. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is: the existing methods for coordinated energy allocation between cold chain parks and highway service areas suffer from inconsistencies in the time scale and quality caliber of multi-source time-series data, leading to unstable scheduling inputs; the failure to convert the thermal inertia and allowable temperature range of the cold chain into a calculable virtual energy storage state and a dispatchable boundary; the separation and optimization of charging load and cooling load, resulting in difficulties in unifying constraints and a lack of closed-loop execution mechanisms; and the problem of how to achieve rolling time-domain coordinated allocation of green electricity in the park and load in the service area and generate executable instructions under the constraints of power distribution capacity and temperature control.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for intelligent green electricity allocation in coordination between cold chain parks and highway networks, including acquiring coordinated operation data of cold chain parks and highway networks, and constructing a unified time-slice scheduling input vector.

[0007] The temperature-based virtual energy storage state variables are calculated based on the scheduling input vector, and the schedulable boundary of the temperature-based virtual energy storage is calculated.

[0008] Perform rolling time-domain collaborative optimization under schedulable boundaries to generate and issue green electricity intelligent dispatch instructions.

[0009] As a preferred embodiment of the intelligent green electricity allocation method for the coordinated operation of cold chain parks and highway networks described in this invention, the acquisition of coordinated operation data between cold chain parks and highway networks includes: reading timestamped data frames from the data interfaces between the cold chain park side and the highway network side, including the current temperature of the cold storage or refrigerated truck compartment, the allowable temperature range, the operating power or gear status of the refrigeration equipment, and the real-time charging load and available charging resource status of the charging stations in highway service areas. The real-time value and predicted sequence of green electricity output from the park, as well as electricity price data, are read, and the data frames, the read real-time value and predicted sequence of green electricity output from the park, and the electricity price data are identified using the same clock source.

[0010] As a preferred embodiment of the intelligent green energy allocation method for the coordinated operation of cold chain industrial parks and highway networks described in this invention, the construction of a unified time-slice scheduling input vector includes: setting a scheduling period and resampling and aligning data frames, mapping data with different sampling frequencies to the same time-slice index. When there are missing measurements or anomalies, the data is replaced based on the previous valid value and the short-term predicted value at the same time-slice index, and a confidence marker is added. Temperature, power, load, green energy output, and electricity price are standardized in dimensions and their ranges are clipped to form a scheduling input vector that includes the current state quantity and the future rolling window prediction quantity.

[0011] As a preferred embodiment of the green energy intelligent allocation method for the coordinated operation of cold chain parks and highway networks described in this invention, the calculation of the virtual energy storage state based on the scheduling input vector includes: for each cold storage or refrigerated truck, reading the current temperature and the allowable temperature range from the scheduling input vector; calculating the ratio using the difference between the maximum allowable temperature and the current temperature as the numerator and the difference between the maximum allowable temperature and the minimum allowable temperature as the denominator; and limiting the ratio to a closed interval between zero and one to obtain the virtual energy storage state. When multiple temperature measurement points exist for the same object, the temperatures of the measurement points are first fused according to a preset weight to obtain an equivalent current temperature before calculating the ratio. The virtual energy storage state is updated at each time slice based on the latest collected temperature.

[0012] As a preferred embodiment of the green energy intelligent allocation method for the coordinated operation of cold chain parks and highway networks described in this invention, the calculation of the schedulable boundary of virtual temperature energy storage includes: calculating the corresponding virtual temperature energy storage capacity scalar based on the equivalent heat capacity parameters of the cold storage or refrigerated truck compartment, the allowable temperature range, and the coefficient of performance (COP). The cooling power range for each time slot is determined based on the maximum allowable power, minimum allowable power, and ramp-up limits of the refrigeration equipment. Combining the upper and lower limit constraints of the virtual temperature energy storage state quantity, the cooling power range is converted into the allowable variation range of the state quantity in the next time slot, thus obtaining the schedulable boundary of virtual temperature energy storage. When a door opening event or loading change indicator is detected, the heat loss estimate is increased and the allowable variation range is simultaneously tightened according to preset rules, which are determined based on the duration of the door opening or the loading change indicator.

[0013] As a preferred embodiment of the intelligent green energy allocation method for the coordinated operation of cold chain industrial parks and highway networks described in this invention, the step of performing rolling time-domain collaborative optimization under the schedulable boundary includes: constructing a multi-time-slice optimization problem with a rolling window length; using the refrigeration power sequence of the cold chain industrial park and the charging power quota sequence or dynamic power upper limit sequence of the highway service area as decision variables; and using the green energy output prediction and electricity price data of the industrial park as exogenous inputs. Under the conditions of satisfying the distribution capacity constraint, power upper and lower limit constraint, and temperature virtual energy storage schedulable boundary constraint, the objective function including the power purchase term and the power fluctuation term is solved. After the solution is completed, the optimal control quantity of the current time slice is output, and the construction and solution are repeated in the next time slice.

[0014] As a preferred embodiment of the intelligent green electricity allocation method for the coordinated operation of cold chain industrial parks and highway networks described in this invention, the generation and issuance of intelligent green electricity allocation instructions includes mapping the optimal control quantity of the current time slice into an executable instruction set, including power setpoints or operating level setpoints for refrigeration equipment in the cold chain industrial park, and charging power quotas or dynamic power limits for charging stations in highway service areas. The instruction set is sent to the corresponding execution devices through the park-side control interface and the service area-side control interface, respectively, and execution receipts are received, with an allocation log containing timestamps, instruction parameters, and receipt status recorded.

[0015] As a preferred embodiment of the green electricity intelligent dispatching system for the coordinated operation of cold chain parks and highway networks as described in this invention, it includes a coordinated data time-series modeling module, a temperature virtual energy storage boundary module, and a rolling optimization instruction issuance module.

[0016] The collaborative data time series modeling module is used to acquire collaborative operation data between the cold chain park and the highway network, and to construct a unified time slice scheduling input vector.

[0017] The temperature virtual energy storage boundary module is used to calculate the temperature virtual energy storage state quantity based on the scheduling input vector, and to calculate the schedulable boundary of the temperature virtual energy storage.

[0018] The rolling optimization instruction issuing module is used to perform rolling time-domain collaborative optimization solutions under schedulable boundaries, and generate and issue green electricity intelligent allocation instructions.

[0019] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement a green energy intelligent dispatching method for coordinating cold chain industrial parks and highway networks.

[0020] A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of a green electricity intelligent allocation method for coordinated cold chain industrial parks and highway networks are disclosed.

[0021] The beneficial effects of this invention are: Through multi-source acquisition, unified time synchronization, time-slicing, and quality management, a computable input vector was constructed from heterogeneous data across entities, operating on the same time axis and with the same quality caliber. This input vector serves as the sole data entry point for subsequent S2 state variable calculations and S3 optimization solutions, transforming raw data that cannot be directly used for optimization into a constrained, traceable, and reproducible scheduling data structure. This avoids optimization infeasibility or control jitter caused by timescale offsets, inconsistent sampling frequencies, and the propagation of missing measurements / anomalies, providing a stable data foundation for cross-park-service area collaborative scheduling.

[0022] Through temperature, state variables, capacity, and boundary mapping steps, the hard constraints of cold chain temperature control are transformed into schedulable boundary constraints that can be directly used by the optimizer. This transforms the temperature control problem, which originally required complex dynamic temperature differential modeling and was difficult to solve in a unified manner with power constraints on the electrical side, into a state interval constraint that can be applied within discrete time slices. This allows S3's rolling optimization to handle both cold chain and charging loads simultaneously under the same constraint language. Thus, while ensuring the solvability of the model and the executableness of the constraints, the thermal inertia of the cold chain is truly treated as a schedulable resource for coordinated allocation.

[0023] By employing rolling time-domain collaborative optimization and closed-loop execution steps, an end-to-end closed loop is achieved, from solvable scheduling results to executable device instructions. The purpose of the closed loop is to align the mathematically optimal solution with on-site execution capabilities and to incorporate abnormal communication and computation timeouts into rule-based processing, ensuring uninterrupted rolling execution. Attached Figure Description

[0024] 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.

[0025] Figure 1 The above is an overall flowchart of a green electricity intelligent allocation method for cold chain industrial parks and highway networks provided in Embodiment 1 of the present invention. Detailed Implementation

[0026] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0027] Example 1, referring to Figure 1 As an embodiment of the present invention, a method for intelligent green electricity allocation in coordination between cold chain industrial parks and highway networks is provided, comprising: S1: Obtain data on the coordinated operation of the cold chain park and the highway network, and construct a unified time-slice scheduling input vector.

[0028] Data frames with timestamps are read from the data interfaces between the cold chain park and the highway network. These frames include the current temperature of the cold storage or refrigerated truck compartment, the allowable temperature range, the operating power or gear status of the refrigeration equipment, and the real-time charging load and available charging resource status of the highway service area charging stations. Real-time values ​​and predicted sequences of the park's green electricity output, as well as electricity price data, are read and identified using the same clock source.

[0029] Furthermore, the data interfaces include the park energy management system interface and the service area charging operation platform interface. Timestamped data frames must include at least the data source identifier, sampling time, measurement value, and quality identifier. The same clock source is either NTP or GPS time synchronization. The processor re-timestamps the arriving data using the synchronized system clock. To ensure the executability of the same clock source identifier, the processor performs a threshold comparison of the timescale drift of each data source. When the difference between the timescale of any data source and the system clock exceeds 2 seconds, the timescale is deemed unreliable, and the data source is marked as low confidence. Simultaneously, re-synchronization or switching to a backup time source is triggered. The 2-second threshold is derived from the typical round-trip delay upper limit of the communication link between the park and the service area and is fixed as a configurable parameter. Low confidence corresponds to setting the confidence identifier of the relevant field in the scheduling input vector of that data source to prediction or replacement.

[0030] Furthermore, the predicted sequence of green electricity output in the park is a future rolling window sequence output according to a unified time slice, generated by the park-side prediction module or a third-party prediction service. When the time index of the predicted sequence is inconsistent with the current system clock, the processor performs index correction according to the nearest time slice alignment rule, and simultaneously retains fields of real-time value and predicted value in the scheduling input vector for subsequent modeling calls.

[0031] A scheduling period is set, and data frames are resampled and aligned, mapping data with different sampling frequencies to the same time slice index. When missing measurements or outliers exist, the data is replaced with the previous valid value and the short-term predicted value at the same time slice index, and a confidence flag is marked. Temperature, power, load, green electricity output, and electricity price are standardized in dimensions and their ranges are clipped to form a scheduling input vector containing the current state quantity and the future rolling window prediction quantity. The short-term predicted value is obtained by linear extrapolation based on the measurement values ​​of the most recent three valid time slices. If there are fewer than three valid samples, the previous valid value is used as the short-term predicted value.

[0032] Furthermore, in one implementation, the scheduling cycle is set to 5 minutes, and the time slice index is generated by rounding down to the 5-minute boundary. Resampling adopts zero-order hold or linear interpolation rules. Zero-order hold is used for state-type quantities such as gear position and start / stop, and linear interpolation is used for continuous quantities such as temperature, power, and load. Interpolation is only enabled when the interval between adjacent valid samples does not exceed 15 minutes. If it exceeds 15 minutes, it enters the missing measurement processing stage.

[0033] The missing measurement determination rule is that if no corresponding measurement is obtained within ±2.5 minutes of the current time slice center point, the measurement is considered missing. Anomaly point determination includes two categories: physical boundary violations and sudden boundary violations. The physical boundary violation thresholds for temperature are set to [-40℃, 30℃], for power and load to [0, 1.2 × rated value], and for electricity price to [0, 10] yuan / kWh. In one implementation, the sudden boundary violation threshold is set as a temperature jump exceeding 2℃ within 5 minutes or a power / load jump exceeding 30% of the rated value in a single time slice. The thresholds are derived from the sensor accuracy and the maximum ramp-up capability limit of the equipment and are fixed as configurable parameters. The rated value is the rated power on the equipment nameplate or the rated power configured on the operating platform. The physical boundary violation thresholds are written as default values ​​in the configuration file and can be adjusted according to the cold chain equipment range and the local electricity price limit.

[0034] Furthermore, the confidence flag is a quality marker written along with each field into the scheduling input vector, encompassing three states: valid, replacement, and prediction. When a missing value occurs, the previous valid value is preferentially used for replacement, and the replacement lasts for no more than two consecutive time slices. If it exceeds two consecutive time slices, the short-term predicted value at the same time slice index is used instead, and the confidence flag is set to prediction. Standardized dimensionality includes normalization to rated values ​​and linear scaling to allowable ranges. Range pruning is performed according to the aforementioned physical boundary thresholds, and the confidence flag of the pruned field is set to replacement. Normalization to rated values ​​is used for power and load fields, linear scaling to allowable ranges is used for temperature fields, and the electricity price field is linearly scaled to a preset upper limit of 10 yuan / kWh.

[0035] It should be noted that this step uses unified time synchronization to align the park's cold chain and service area load, green electricity and electricity price data to a fixed time slice, and manages data quality through missing measurement anomaly handling and confidence labeling to form an input vector that can be directly used for subsequent temperature virtual energy storage calculation and rolling optimization, thereby reducing the interference of heterogeneous sampling and time scale drift on scheduling calculation.

[0036] S2: Calculate the temperature virtual energy storage state quantity based on the scheduling input vector, and calculate the schedulable boundary of the temperature virtual energy storage.

[0037] For each cold storage room or refrigerated truck, the current temperature and allowable temperature range are read from the scheduling input vector. A ratio is calculated using the difference between the maximum allowable temperature and the current temperature as the numerator, and the difference between the maximum allowable temperature and the minimum allowable temperature as the denominator. This ratio is then confined to a closed interval between zero and one to obtain the virtual temperature energy storage state. When multiple temperature measurement points exist for the same object, the temperatures of these points are first fused according to preset weights to obtain an equivalent current temperature before calculating the ratio. The virtual temperature energy storage state is updated at each time slice based on the latest acquired temperature.

[0038] The equivalent current temperature is obtained by weighted fusion of the temperatures measured at the same temperature measurement points in the same cold storage or refrigerated truck:

[0039] in, Indicates cold storage or refrigerated truck compartment In time slice The equivalent current temperature. This is an object identifier for cold storage or refrigerated truck compartments, used to distinguish different temperature-controlled objects. This represents the unified time slice index, which is derived from the time slice scheduling input vector constructed by S1. Indicates the temperature measurement point number, used to identify the object. Inner There are several temperature measurement points. Representation Object The number of temperature measurement points. Representation Object The The fusion weight of each temperature measurement point satisfies . Representation Object The Each temperature measurement point in time slice The actual measured temperature.

[0040] The default weights are 0.5 for the return air vent, 0.3 for the center of the storage unit, and 0.2 for the door area. If a temperature measurement point is marked as low confidence in S1, then the weights of the temperature measurement points are merged. Set it to 0, and normalize the remaining weights before recalculating. .

[0041] For each cold storage unit or refrigerated truck, the allowable temperature range and equivalent current temperature are read from the scheduling input vector. The virtual temperature energy storage state is obtained by subtracting the current temperature from the maximum allowable temperature range, and is limited to between 0 and 1. The virtual temperature energy storage state is represented as:

[0042] in, Representation Object In time slice The temperature virtual energy storage state quantity has a value range of [0,1]. Represents the interval clipping function, if Then take 0, if Then take 1, otherwise take . Representation Object The maximum permissible temperature is the upper limit of the permissible temperature range collected by S1. Representation Object The minimum allowable temperature is the lower limit of the allowable temperature range collected by S1. Representation Object In time slice The equivalent current temperature.

[0043] Update in real time for each time slice And write it to the state cache as input for subsequent temperature-based virtual energy storage schedulable boundaries. When detected... When the default threshold of 0.5 degrees Celsius is determined by the temperature sensing accuracy and control resolution configuration rules, the processor will... Marked as non-adjustable and fixed This avoids division by zero and unstable calculations from propagating to subsequent boundaries.

[0044] Based on the equivalent heat capacity parameters, permissible temperature range, and coefficient of performance (COP) of the cold storage or refrigerated truck compartment, the corresponding virtual temperature energy storage capacity scalar is calculated. According to the maximum and minimum permissible power of the refrigeration equipment and ramp limits, the selectable range of refrigeration power for each time slice is determined. Combining the upper and lower limits of the virtual temperature energy storage state variables, the selectable range of refrigeration power is converted into the permissible variation range of the state variables in the next time slice, thus obtaining the schedulable boundary of the virtual temperature energy storage. When a door opening event or loading change indicator is detected, the heat loss estimate is increased and the permissible variation range is simultaneously tightened according to preset rules. The preset rules are determined based on the duration of the door opening or the loading change indicator.

[0045] Based on the equivalent heat capacity parameters, allowable temperature range, and coefficient of performance (COP) of the cold storage or refrigerated truck, the storable heat range is mapped to the electrical energy domain, resulting in a scalar of virtual temperature energy storage capacity. This scalar is used to convert refrigeration power constraints into state variable change constraints. The virtual temperature energy storage capacity scalar is expressed as:

[0046] in, Representation Object The temperature-based virtual energy storage capacity scalar is used to characterize the space of schedulable energy corresponding to the temperature-allowed range. Representation Object The equivalent heat capacity parameters are derived from the equipment commissioning and calibration or online identification results and written into the configuration table. Representation Object The maximum permissible temperature. Representation Object The minimum permissible temperature. Representation Object The coefficient of performance (COP) of a refrigeration system is derived from the equipment nameplate or energy efficiency curve and can be adjusted according to operating conditions.

[0047] processor By converting cooling power × time slice length into the magnitude of state variable changes, subsequent rolling optimization can directly apply constraints in the state space without introducing temperature differential equations into the optimizer.

[0048] Based on the maximum and minimum allowable power of the refrigeration equipment and the ramp limit, the selectable range of refrigeration power for each time slice is determined. Combined with the virtual energy storage state quantity of temperature and the heat loss deduction item, the power range is converted into the allowable variation range of the state quantity of the next time slice, which serves as the schedulable boundary output of the virtual energy storage of temperature.

[0049]

[0050] in, Representation Object In the next time segment Temperature virtual energy storage state quantity. Representation Object In the current time slice Temperature virtual energy storage state quantity. Representation Object The equivalent charging efficiency is set to 0.95 by default. Representation Object In time slice The lower limit of the selectable cooling capacity range is consistent with the S1 power field. Representation Object In time slice The upper limit of the selectable cooling power range is consistent with the S1 power field. This indicates the length of the scheduling cycle, with a default value of 5 minutes. Representation Object In time slice The heat loss is deducted from the state quantity calculation.

[0051] This interval is written into the constraint set as the schedulable boundary of temperature-based virtual energy storage, for subsequent S3 rolling optimization. , The ramp threshold is determined by the upper and lower limits of the equipment power and the ramp constraint. In this embodiment, the ramp threshold is equal to 0.2 times the rated power on the equipment nameplate, which means that the power change within a single time slice does not exceed 20% of the rated power.

[0052] When a door opening event or a load change indicator is detected, the processor increases the heat loss estimate according to preset rules and simultaneously tightens the allowable variation range in the schedulable boundary of the virtual temperature energy storage. The preset rules are determined based on the duration of the door opening or the load change indicator and provide clear threshold segments.

[0053]

[0054] in, Representation Object In time slice The heat loss is deducted from the state quantity calculation. Representation Object In time slice The baseline heat loss deduction item is derived from historical statistics or identification parameter tables under the same operating conditions. Representation Object In time slice The event trigger amplification factor. Representation Object In time slice The duration of the door being open inside is recorded by the door magnetic sensor. Representation Object In time slice The loading change flag is set to 0 or 1. Once a loading change is detected, it remains at 1 for the next two time slices. 30 and 120 represent the threshold for door opening duration; the default values ​​are derived from the short-opening or long-opening door boundary rules in the park's standard operating procedures and are fixed as configurable parameters. 0.3, 0.6, and 0.2 represent the default amplification factor, derived from the statistical proportion of door opening or loading to the temperature recovery rate in historical operating data and are fixed as configurable parameters.

[0055] When the duration of the door being open falls into different threshold ranges, the processor calculates the corresponding... And enlarge This causes the boundary intervals of the virtual energy storage state quantity at temperature to tighten downwards simultaneously. When load changes occur simultaneously, additional [features / effects] are added. This feature makes the rolling optimization of the cooling power control output in the next time slice more conservative during the event, avoiding aggressive reductions that do not conform to the temperature control model during door opening or loading phases.

[0056] It should be noted that the design concept of this step is to transform the cold chain temperature control problem from a hard temperature constraint and experience-based adjustment into an optimizable virtual energy storage state and schedulable boundary. The equivalent temperature is obtained through multi-point fusion, and the temperature margin is then mapped to a state variable of 0-1. Furthermore, the upper and lower limits of equipment power, ramp-up limits, and heat loss events are uniformly converted into the allowable variation range of the state variable for the next time slice. This allows subsequent rolling optimization to simultaneously allocate cold chain and charging loads under the same constraint language. Compared to conventional solutions that only perform pre-cooling or peak shaving, this invention provides a computable intermediate layer of state, capacity, and boundary, enabling the optimizer to directly invoke and maintain consistency for cross-time peak shifting and boundary tightening during events.

[0057] S3: Perform rolling time-domain collaborative optimization under schedulable boundaries to generate and issue green electricity intelligent dispatch instructions.

[0058] A multi-time-slice optimization problem is constructed using a rolling window length. The refrigeration power sequence of the cold chain industrial park and the charging power quota sequence or dynamic power upper limit sequence of the highway service area are used as decision variables, while the predicted green electricity output and electricity price data of the industrial park are used as exogenous inputs. Under the constraints of distribution capacity, power upper and lower limits, and the dispatchable boundary of temperature-based virtual energy storage, the objective function, which includes a power purchase term and a power fluctuation term, is solved. After the solution is completed, the optimal control quantity for the current time slice is output, and the process is repeated in the next time slice. The power purchase term and the power fluctuation term are normalized and weighted, with the default weights set to 1 for the power purchase term and 0.2 for the power fluctuation term. The weights are derived from configuration rules that ensure cost dominance and suppress fluctuations, and can be adjusted in the configuration file.

[0059] Furthermore, the length of the rolling window is represented by the number of time slices. In this invention, 12 time slices are used, with each time slice having a length of 5 minutes, thus corresponding to multi-time slice optimization for the next 60 minutes. The default value of the length is determined based on the controllable error range of the green electricity output prediction within 60 minutes and the configuration rule that the dynamic response of the cold chain temperature is adjustable at the hour level, and can be adjusted in the configuration file within the range of 30 minutes to 120 minutes.

[0060] Distribution capacity constraints include upper limits on access capacity for both the park and service areas, and upper limits on tie-line power. The tie-line power limit is derived from the tie-line rated capacity or protection setting configuration table. The upper limits on access capacity for the park side and the service area side are given by the capacity parameter table of the distribution cabinet or operation platform. In each time slice, the processor calculates the purchased power based on the exogenous green electricity output prediction, electricity price data, and current state variables, and limits it within the corresponding capacity limit. Simultaneously, it combines the refrigeration power sequence of the cold chain park side with the charging power quota sequence or dynamic power limit sequence of the highway service area into the same constraint set. The purchased power is determined according to the power balance relationship of the same time slice, calculated as the sum of the refrigeration power of the cold chain park side and the charging power of the highway service area minus the real-time value of the park's green electricity output. If the calculation result is less than zero, it is counted as zero, and the result is used as the restricted quantity of the distribution capacity constraint.

[0061] Furthermore, in one implementation, the solution process transforms the objective function and constraint set into a quadratic programming problem and solves it using the interior-point method or the active set method. When the refrigeration equipment in the cold chain park uses discrete speed control, the corresponding decision variables are set to integers and the solution is switched to mixed-integer quadratic programming. To ensure rolling execution, the solver sets a single solution timeout threshold of 1 second (the default value is derived from the rule that the reserved calculation time under a 5-minute scheduling cycle does not exceed 0.3% of the scheduling cycle). If the timeout occurs, the optimal control quantity that has been verified to be feasible in the previous time slice is output, and the solution is reconstructed in the next time slice.

[0062] The optimal control values ​​for the current time slice are mapped to an executable instruction set, including power setpoints or operating level setpoints for refrigeration equipment in the cold chain park, and charging power quotas or dynamic power limits for charging stations in highway service areas. The instruction set is sent to the corresponding execution devices through the park-side control interface and the service area-side control interface, respectively. Execution receipts are received, and a dispatch log containing timestamps, instruction parameters, and receipt status is recorded.

[0063] Furthermore, mapping to an executable instruction set includes quantizing continuous optimal control quantities at the device control granularity. In one embodiment, the quantization step size for the cooling power setpoint is 0.1kW, and the quantization step size for the charging power quota is 1kW. The default step size is derived from the configuration rules of commonly used inverter power resolution and charging pile power metering resolution. When using the operating gear setpoint, the processor maps the power setpoint to the corresponding gear according to the nearest gear mapping rule, and performs power upper and lower limit constraint verification again after mapping to ensure that the instruction can be executed by the device.

[0064] Sending and receiving execution receipts involves request-response communication with a 3-second timeout threshold. This 3-second threshold is derived from the upper limit of the typical round-trip latency of the communication link between the campus and the service area and is fixed as a configurable parameter. When a receipt times out or returns a failure status, the processor retransmits it twice according to the retry threshold. If it still fails, a degradation instruction is issued to the object. The degradation instruction maintains the power setting or operating gear setting value executed in the previous time slice and simultaneously locks the decision variables of the object in the next time slice as non-adjustable to avoid control jitter caused by consecutive failures. The duration of the non-adjustable lock is one time slice. If the receipt recovers normally in the next time slice, the lock is released and the object is re-included in the solution.

[0065] Furthermore, in addition to timestamps, command parameters, and receipt status, the allocation log also records the rolling window length used for reproducible calculations, the version number of exogenous inputs, and a summary identifier of the schedulable boundary of the temperature virtual energy storage. A hash checksum is generated for each log entry to prevent tampering. The processor rotates the logs according to a storage cycle threshold of 180 days. This 180-day threshold is derived from the configuration rules of the park's operation audit and settlement reconciliation cycle and can be adjusted in the configuration file.

[0066] It should be noted that this step incorporates cold chain refrigeration and service area charging into the same optimization problem using a rolling window, solving it under the constraint of temperature-based virtual energy storage schedulable boundary conditions. Only the control quantity for the current time slice is issued, and the problem is continuously rebuilt, balancing real-time performance and feasibility. Furthermore, closed-loop control and traceable logs are formed through quantization mapping, retrieval timeout retry, and degradation locking, ensuring a stable and implementable execution path for collaborative dispatch.

[0067] Example 2, an embodiment of the present invention, provides a green energy intelligent dispatching system that coordinates cold chain parks and highway networks, including a collaborative data time-series modeling module, a temperature virtual energy storage boundary module, and a rolling optimization instruction issuing module.

[0068] The collaborative data time series modeling module is used to acquire collaborative operation data between cold chain parks and highway networks, and to construct a unified time slice scheduling input vector.

[0069] The temperature virtual energy storage boundary module is used to calculate the temperature virtual energy storage state variables based on the scheduling input vector, and to calculate the schedulable boundary of temperature virtual energy storage.

[0070] The rolling optimization instruction issuance module is used to perform rolling time-domain collaborative optimization solutions under schedulable boundaries, and generate and issue green electricity intelligent dispatch instructions.

Claims

1. A method for intelligent green electricity allocation in coordination between cold chain industrial parks and highway networks, characterized in that, include: Acquire data on the coordinated operation of cold chain parks and highway networks, and construct a unified time-slice scheduling input vector; The temperature virtual energy storage state variables are calculated based on the scheduling input vector, and the schedulable boundary of the temperature virtual energy storage is calculated. Perform rolling time-domain collaborative optimization under schedulable boundaries to generate and issue green electricity intelligent dispatch instructions.

2. The green energy intelligent allocation method for cold chain industrial parks and highway networks as described in claim 1, characterized in that: The acquisition of data on the coordinated operation of the cold chain park and the highway network includes... Read timestamped data frames from the data interfaces between the cold chain park side and the highway network side, including the current temperature of the cold storage or refrigerated compartment, the allowable temperature range, the operating power or gear status of the refrigeration equipment, and the real-time charging load and available charging resource status of the charging stations in the highway service area. Read the real-time value and prediction sequence of green power output in the park, as well as the electricity price data, and identify the data frame and the read real-time value and prediction sequence of green power output in the park, as well as the electricity price data, according to the same clock source.

3. The green energy intelligent allocation method for cold chain industrial parks and highway networks as described in claim 2, characterized in that: The construction of the unified time-slice scheduling input vector includes: Set the scheduling period and resample and align the data frames to map data with different sampling frequencies to the same time slice index; When there are missing values ​​or outliers, replace them with the short-term predicted values ​​based on the previous valid value and the index of the same time slice, and mark them with a confidence flag. Temperature, power, load, green electricity output, and electricity price are standardized and their ranges are clipped to form a scheduling input vector that includes the current state variables and the future rolling window prediction variables.

4. The green energy intelligent allocation method for cold chain industrial parks and highway networks as described in claim 3, characterized in that: The calculation of the temperature virtual energy storage state quantity based on the scheduling input vector includes, For each cold storage or refrigerated truck, the current temperature and the temperature allowable range are read from the scheduling input vector. The ratio is calculated using the difference between the maximum allowable temperature and the current temperature as the numerator and the difference between the maximum allowable temperature and the minimum allowable temperature as the denominator. The ratio is then limited to a closed interval between zero and one to obtain the virtual temperature energy storage state quantity. When there are multiple temperature measurement points for the same object, the temperatures of the temperature measurement points are first fused according to the preset weights to obtain the equivalent current temperature, and then the ratio is calculated. The virtual energy storage status is updated based on the latest acquired temperature in each time slice.

5. The green energy intelligent allocation method for cold chain industrial parks and highway networks as described in claim 4, characterized in that: The computational temperature virtual energy storage schedulable boundary includes, Based on the equivalent heat capacity parameters, allowable temperature range, and coefficient of performance of the cold storage or refrigerated truck, calculate the corresponding virtual energy storage capacity scalar. Based on the maximum allowable power, minimum allowable power, and ramp limit of the refrigeration equipment, determine the range of refrigeration power for each time slot; By combining the upper and lower limits of the temperature virtual energy storage state variables, the cooling power range is converted into the allowable variation range of the state variables in the next time slice, thus obtaining the schedulable boundary of temperature virtual energy storage. When a door opening event or a load change indicator is detected, the heat loss estimate is increased according to preset rules and the allowable change range is tightened simultaneously. The preset rules are determined based on the duration of the door opening or the load change indicator.

6. The green energy intelligent allocation method for cold chain industrial parks and highway networks as described in claim 5, characterized in that: The step of performing rolling time-domain collaborative optimization under schedulable boundaries includes, A multi-time-slice optimization problem is constructed using the rolling window length. The refrigeration power sequence of the cold chain park and the charging power quota sequence or dynamic power upper limit sequence of the highway service area are used as decision variables, and the green electricity output forecast of the park and the electricity price data are used as exogenous inputs. Under the conditions of satisfying the power distribution capacity constraint, power upper and lower limit constraint, and temperature virtual energy storage dispatchable boundary constraint, the objective function containing the power purchase term and power fluctuation term is solved. After the solution is completed, the optimal control quantity of the current time slice is output and the construction and solution are repeated in the next time slice.

7. The green energy intelligent allocation method for cold chain industrial parks and highway networks as described in claim 6, characterized in that: The generation and issuance of green electricity intelligent dispatch instructions include: Map the optimal control quantity of the current time slice to an executable instruction set, including power setting values ​​or operating gear setting values ​​for refrigeration equipment in cold chain parks, and charging power quotas or dynamic power limits for charging stations in highway service areas. The instruction set is sent to the corresponding execution device through the park-side control interface and the service area-side control interface, respectively. The execution receipt is received and the allocation log containing timestamp, instruction parameters and receipt status is recorded.

8. A green energy intelligent dispatching system for cold chain industrial parks and highway networks, employing the green energy intelligent dispatching method for cold chain industrial parks and highway networks as described in any one of claims 1 to 7, characterized in that: This includes a collaborative data time-series modeling module, a temperature virtual energy storage boundary module, and a rolling optimization command issuance module; The collaborative data time series modeling module is used to acquire collaborative operation data between the cold chain park and the highway network, and to construct a unified time slice scheduling input vector. The temperature virtual energy storage boundary module is used to calculate the temperature virtual energy storage state quantity based on the scheduling input vector, and to calculate the schedulable boundary of the temperature virtual energy storage. The rolling optimization instruction issuing module is used to perform rolling time-domain collaborative optimization solutions under schedulable boundaries, and generate and issue green electricity intelligent allocation instructions.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the green electricity intelligent allocation method for coordinating cold chain parks and highway networks as described in any one of claims 1 to 7.

10. 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 green electricity intelligent dispatching method for coordinating cold chain parks and highway networks as described in any one of claims 1 to 7.

Images