A carbon intensity constraint energy management and direct current traction power supply safe operation method

Abstract

The present application belongs to the technical field of energy management and safe operation, and relates to a carbon intensity constraint energy management and DC traction power supply safe operation method, which realizes accurate space-time alignment of an external carbon intensity signal and an EMS global time axis by establishing a global monotonic clock on the EMS side, generates an effective window gate state in combination with signal update characteristics and coverage indexes, and provides reliable signal support for carbon constraint scheduling. Based on comparison of a power grid net exchange power increment and a historical significance threshold, marginal or average emission factors are adaptively selected as carbon constraint factors to improve carbon emission reduction accuracy. By calculating an effective voltage safety margin of a DC bus, candidate instructions are shaped in combination with device ramping capability, a multi-state adaptive state machine is constructed to dynamically adjust carbon constraint participation, and safe priority hierarchical scheduling is realized.

Description

Technical Field

[0001] This invention belongs to the field of energy management and safe operation and maintenance technology for rail transit vehicles, specifically relating to a carbon intensity-constrained energy management and DC traction power supply safe operation and maintenance method. Background Technology

[0002] The current Energy Management System (EMS) of urban rail transit DC traction power supply systems focuses on power supply stability and operational economy as its core optimization objectives. Typical functions include regenerative braking energy recovery and utilization, energy storage system charging and discharging control, voltage regulation of reversible substations / bidirectional converters, and comprehensive loss optimization of the traction power supply system. As the core support for traction power supply and regenerative braking energy recovery in rail transit vehicles, this EMS needs to adapt to the drastic power fluctuations caused by frequent vehicle starts and stops and alternating traction / braking. With the deepening of the "dual-carbon" strategy in the rail transit industry, incorporating carbon emission reduction indicators into core operation and maintenance objectives has become an industry consensus. In practice, this is commonly achieved by introducing carbon intensity data (such as carbon intensity) published by power grid companies and third-party carbon monitoring platforms. CI The average emission factor (AEF) and marginal emission factor (MEF) are used to optimize the scheduling decisions of the traction power supply system using carbon constraints. However, the unique operating characteristics of rail transit vehicles exacerbate technical contradictions: First, the instantaneous impact of vehicle traction load and the misalignment of carbon signal time scales can easily lead to conflicts between carbon constraint instructions and vehicle power supply needs. Second, the randomness of vehicle braking regenerative energy coupled with low-quality carbon signals can cause bus voltage oscillations, affecting vehicle traction stability. Third, existing technologies do not take into account the load differences during peak / off-peak operating periods when selecting carbon factors, making it difficult to balance carbon emission reduction and vehicle power supply safety.

[0003] Specifically, there is a natural incompatibility between the external carbon intensity signal and the inherent characteristics of the traction power supply system. Directly introducing it into the existing EMS closed-loop control would trigger a series of technical risks. Firstly, there is the issue of inconsistent timestamp systems. The release time and applicable period of the carbon intensity signal use the data source's own clock, while the Supervisory Control and Data Acquisition (SCADA) system for sampling and converter control of the traction power supply system uses the EMS local clock. This lack of synchronization can easily lead to a misalignment between the applicable period of the carbon intensity signal and the control window. For example, if the applicable period of a certain carbon intensity signal is 10:00:00~10:05:00 (data source clock), but the EMS local clock is 8 seconds ahead, and the current control window is 10:04:55~10:04:56 (EMS clock), the system will misjudge that the carbon intensity signal does not cover the current window, and thus use an expired carbon intensity signal for scheduling. Combined with the impact of vehicle traction load, the risk of conflict is significantly increased. Secondly, there are defects in the quality of carbon intensity signal transmission. During network transmission, carbon intensity signals suffer from random delays, update interval jitter, data out-of-order delivery, and missing data segments. Low-quality carbon signals, coupled randomly with vehicle regenerative energy, directly drive scheduling, leading to frequent command reversals and bus voltage fluctuations. Thirdly, there is a mismatch in the selection of carbon constraint factors. Carbon-constrained scheduling requires selecting either AEF or MEF based on the nature of the control action—AEF is suitable for statistical attribution, reflecting the average carbon emission level of the power grid over a period of time; MEF is suitable for real-time incremental scheduling, reflecting the incremental carbon emissions per unit of new power consumption. Existing technologies often use a single fixed factor without considering load differences during vehicle operation periods, easily leading to scheduling deviations and missed carbon reduction opportunities.

[0004] The combination of these problems directly leads to misjudgments in carbon-constrained scheduling, which in turn triggers safety boundary conflicts in the DC traction power supply system: the carbon optimization command and the bus voltage stabilization target are mutually antagonistic, resulting in overshoot or undershoot of the bus voltage, triggering malfunctions of overvoltage or undervoltage protection; frequent command reversals cause high-frequency oscillations in the bus voltage, affecting the stability of train traction power supply; during operation and maintenance, it is impossible to trace the basis of scheduling decisions, making it difficult to determine the reasons for the failure of carbon emission reduction to meet expectations or the responsibility for protection actions. Summary of the Invention

[0005] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution: Firstly, a carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method is proposed, including the following steps: Establish a global monotonic clock on the EMS side; obtain the arrival time, release time, and applicable time period of the carbon intensity signal from an external data source; calculate the time offset based on the difference between the arrival time and the release time, and use the time offset to map the applicable time period to the global time axis of the EMS to obtain the aligned carbon intensity signal. Determine whether the applicable time period of the aligned carbon intensity signal covers the current control window; if so, obtain the coverage index. Analyze the update interval jitter and continuity of the carbon intensity signal, and generate an effective window gating state by combining the coverage index; If the effective window gating state allows participation in scheduling, the net power exchange increment of the power grid is predicted based on the candidate control actions of this cycle; it is then determined whether the absolute value of the net power exchange increment of the power grid exceeds the significance threshold obtained from the statistics of historical stable operation data; if so, the marginal emission factor is selected as the carbon constraint factor, otherwise the average emission factor is selected as the carbon constraint factor. Based on the relative position of the selected carbon constraint factor in the historical time series distribution, and combined with the current adjustable resource constraints, candidate scheduling instructions containing energy storage charging and discharging power or converter voltage regulation are generated. Calculate the voltage safety margin of the DC bus voltage relative to the protection trip threshold, and obtain the effective safety margin after deducting measurement noise redundancy; obtain the allowable voltage change based on the effective safety margin; and perform slope limiting and smoothing shaping on candidate scheduling instructions according to the allowable voltage change and equipment ramping capability to obtain the final execution instruction. A state machine is constructed based on the effective window gating state, voltage safety margin, and protection event count. When the state machine is in a frozen state, it exits the carbon constraint scheduling and executes instruction maintenance or voltage stabilization and limiting actions. When the state machine is in a recovering state, the restriction on the change of instruction is linearly released over time.

[0006] Secondly, a carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance system is proposed to perform the carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance as described in the first aspect, including: The signal processing module is used to acquire external carbon intensity signals and perform time-scale alignment using time offsets; The gating and factoring module is used to generate effective window gating states and adaptively select carbon constraint factors based on the significance of the net power exchange increment in the power grid. The instruction generation module is used to generate candidate scheduling instructions based on the carbon constraint factor. The safety shaping module is used to calculate voltage and power safety margins and to perform slope limiting and smoothing shaping on candidate scheduling commands based on the allowable changes. The state machine control module is used to adjust the system operating state according to the gating state and safety margin, and to enforce the voltage stabilization priority strategy in the frozen state.

[0007] Thirdly, a computer-readable storage medium is proposed, on which instructions are stored, which, when executed on a computer, perform a carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method as described in the first aspect.

[0008] Compared with existing technologies, this invention has the following advantages and beneficial effects: By establishing a global monotonic clock on the EMS side, the applicable time period of the external carbon intensity signal is mapped to the global time axis of the EMS. Combined with signal update interval jitter, continuity, and coverage indicators, an effective window gating state is generated, achieving precise spatiotemporal alignment and effectiveness discrimination of the carbon intensity signal. This solves the technical pain points of inconsistent time bases and difficulty in quantifying the effectiveness of multi-source carbon signals, providing a reliable signal input basis for carbon-constrained scheduling. Furthermore, based on the comparison results of the net power exchange increment of the power grid and the historical significance threshold, the marginal emission factor or average emission factor is dynamically selected as the carbon constraint factor. This ensures a high degree of matching between the carbon factor type and the carbon emission impact mechanism of the scheduling action, avoiding scheduling deviations caused by factor misalignment, and improving the accuracy and effectiveness of carbon emission reduction constraints. Simultaneously, by calculating the effective voltage safety margin of the DC bus, By combining the ramp-up capability of equipment with slope limiting and smoothing shaping of candidate scheduling instructions, and relying on an adaptive state machine with multiple states, the carbon constraint participation is dynamically adjusted according to the gating state, voltage safety margin, and protection event count. When the carbon signal is abnormal or the system operating conditions become tight, the system switches to a voltage stabilization priority strategy in a timely manner, and linearly releases the instruction restriction when the system recovers. This establishes a safety-first hierarchical scheduling instruction generation mechanism, which not only ensures the safe and stable operation of the DC traction power supply system, but also achieves a deep integration of carbon intensity constraints and energy management. Finally, through the linkage of state machine output parameters with instruction generation, shaping, log traceability, and other links, as well as the accurate recording of state switching reason codes, a closed-loop control system for the entire process of scheduling strategy from signal input, factor selection, instruction generation to state adjustment is formed. This significantly improves the convenience and controllability of system operation and maintenance, and achieves synergistic optimization of carbon emission reduction benefits and power supply safety and stability. Attached Figure Description

[0009] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of a carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method provided in Embodiment 1 of the present invention. Detailed Implementation

[0010] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. The illustrative embodiments and descriptions of this invention are for illustrative purposes only and are not intended to limit the invention. The embodiments described below are some, but not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0011] In the following description, numerous specific details are set forth to provide a thorough understanding of the invention. However, it will be apparent to those skilled in the art that these specific details are not necessary to practice the invention. In other embodiments, well-known structures, materials, or methods are not specifically described to avoid obscuring the invention. Unless otherwise specified, the materials, instruments, and reagents used in the following embodiments are commercially available. Unless otherwise specified, the techniques used in the embodiments are conventional methods well known to those skilled in the art.

[0012] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0013] Example 1: Due to issues such as inconsistent time scales, transmission delays, and update jitter in carbon intensity signals, and the mismatch between the selection of average carbon intensity factor and marginal carbon intensity factor and the nature of control actions, these factors together lead to misjudgments in carbon constraint scheduling, which in turn triggers safety boundary conflicts in the DC traction power supply system. Specifically, this manifests as mutual tension between the bus voltage stabilization target and the carbon optimization command, malfunctions of protection devices, and bus voltage oscillations.

[0014] To address the aforementioned issues, this embodiment provides a carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method, employing an overall technical framework of "credible carbon intensity signal loop insertion and safety gating fallback." This framework consists of two mutually coupled closed loops: the first closed loop is a carbon intensity signal reliability verification closed loop—by unifying timestamps, the external carbon intensity signal and the EMS control window are aligned to the same time axis, and then multi-dimensional evaluations of coverage, continuity, and jitter consistency are performed to form an effective window gating state. G CO2The system ensures that only qualified carbon intensity signals enter the scheduling process, and adaptively selects either the average carbon intensity factor (AEF) or the marginal carbon intensity factor (MEF) based on the net switching power increment to avoid incorrect factor selection. The second closed loop is a safety gating and fallback loop—after generating candidate carbon constraint commands, the commands are shaped according to the bus voltage safety margin and equipment ramp-up capability to ensure that candidate commands do not cross safety boundaries. If invalid carbon intensity signals, tightening safety margins, or an abnormal increase in protection events occur, freezing or degradation is triggered, and a voltage stabilization priority strategy is switched to. Carbon constraint participation is gradually restored after conditions recover. Finally, an evidence chain is formed throughout the entire process for easy operation and maintenance traceability.

[0015] Based on the above overall technical framework, the carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method includes: Figure 1 The following steps are shown: Step 1: Establish a global monotonic clock on the EMS side; obtain the arrival time, release time, and applicable time period of the carbon intensity signal from an external data source; calculate the time offset based on the difference between the arrival time and the release time, and use the time offset to map the applicable time period to the global time axis of the EMS to obtain the aligned carbon intensity signal.

[0016] The purpose of this step is to eliminate timescale misalignment caused by the inconsistency between the carbon intensity signal data source clock and the EMS local clock, and to avoid misusing carbon constraint factors that are not applicable to the current control window in scheduling decisions.

[0017] The specific implementation method for this step is as follows: In this step, the external data sources mainly include power grid companies and third-party carbon monitoring platforms. These external data sources can publish data such as carbon intensity. CI Carbon intensity data, including average emission factor (AEF) and marginal emission factor (MEF), are required. This step involves obtaining the arrival times of carbon intensity signals from these external data sources. t recv Release time t publish and applicable time period t valid,from ~ t valid,to .

[0018] Based on the obtained carbon intensity signal, the following processing is performed: First, construct a global monotonic clock for the EMS. Synchronize the EMS local clock via GPS timing or NTP protocol (time synchronization accuracy ≤1ms) to ensure the clock monotonically increases (no jumps, no backoffs), serving as the time base for the entire system. t local .

[0019] Subsequently, the offset of the carbon intensity signal timescale relative to the EMS timescale was calculated. offset Specifically, this is achieved by maintaining a sliding statistical window containing multiple recent carbon intensity signals (e.g., 30 carbon intensity signals), and recording the time difference Δ of each carbon intensity signal. t i = t recv,i - t publish,i This time difference is the difference between the EMS receiving time and the data source publishing time; the initial offset is calculated using the median algorithm. offset This algorithm can effectively resist the influence of network latency outliers (such as a single delay of 50 seconds). If the time difference Δ of a certain carbon intensity signal... t i Satisfy |Δ t i -median(Δ t )|>3•MAD(Δ t If a value is found to be outlier, it is removed. Here, median(Δt) is the median of the time difference sequence, representing the central tendency of the time difference sequence; MAD(Δt) is the absolute median deviation of the time difference, specifically the median of the absolute deviations of each value in the time difference sequence from its median, used to characterize the dispersion of the time difference sequence. It should be noted that in this embodiment, the offset is dynamically updated every time a predetermined number of new carbon intensity signals (e.g., 10) are received. offset This allows it to adapt to slow changes in network latency.

[0020] Finally, the applicable time period of the carbon intensity signal is mapped to the EMS time axis to obtain the local applicable time period. t valid,from,local , t valid,to,local The core of time-stamp mapping is to accurately convert the "absolute time interval" of the carbon intensity signal on the data source side into a "relative time interval" under the EMS local clock, ensuring the comparability of the applicable time period of the carbon intensity signal with the control window. This is specifically explained in detail in three scenarios: Additional explanation needed: The applicable time period for the carbon intensity signal must match the operating time of rail transit vehicles (peak operating time, off-peak operating time, and nighttime shutdown time) to ensure that carbon-constrained scheduling is adapted to the traction load requirements of vehicles at different times.

[0021] 1. Complete time-stamped information scenario If the carbon intensity signal carries a complete timestamp field (i.e., the external data source provides the publication time of the carbon intensity signal) t publish Applicable time period start time t valid,from and the end time of the applicable period tvalid,to If the offset is calculated, then it is directly based on the already calculated offset. offset Perform a linear mapping, the mapping formula is: Applicable time period start time mapping: tᵥ al ᵢ d,f ᵣ om,local = tᵥ al ᵢ d,f ᵣ om + offset ; Applicable time period end time mapping: tᵥ al ᵢ d,to,local = tᵥ al ᵢ d,to + offset ; The mapping logic is based on the assumption that "the deviation between the data source clock and the EMS clock is globally consistent," and the offset... offset The effects of random delays have been eliminated through robust statistics, therefore the mapped time period accurately reflects the effective interval of the carbon intensity signal on the EMS time axis. For example, if the start time of the applicable time period is obtained from an external data source... tᵥ al ᵢ d , f ᵣ om =10:00:00.000, End time of applicable period tᵥ al ᵢ d,to =10:05:00.000, the offset is calculated. offset =8 seconds, then the applicable time period for EMS locally is 10:00:08.000~10:05:08.000. If the current EMS control window is 10:02:00.000~10:02:01.000, it can be directly determined that the signal covers this window.

[0022] 2. Missing release moment scenario If the external data source does not provide the release time of the carbon intensity signal t p ᵤ bl ᵢ sh An equivalent release time needs to be constructed before mapping, as follows: First, estimate the update period of the carbon intensity signal. Tᵤ By statistically analyzing the arrival time interval Δ of the most recent multiple (e.g., 10) valid carbon intensity signals, tᵣ ec ᵥᵢ Take all arrival time intervals Δ tᵣ ec ᵥᵢ The median is used as the update period Tᵤ.

[0023] Then, construct the equivalent release time of the carbon intensity signal. t p ᵤ bl ᵢ sh , eq . t p ᵤ bl ᵢ sh , eq = tᵥ al ᵢ d , f ᵣ om - Tᵤ / 2, which assumes that the carbon intensity signal is released at the midpoint of the applicable period, conforms to the release logic of most carbon data platforms.

[0024] Finally, the equivalent time difference Δ is calculated. tᵢ ,eq Δ tᵢ ,eq = tᵣ ec ᵥᵢ - t p ᵤ bl ᵢ sh,eq , tᵣ ec ᵥᵢ The EMS system receives the external carbon intensity signal based on a globally monotonic clock, and subsequently calculates the offset according to the logic of a complete time-scaled scenario. offset Then, the applicable time period for the local area is obtained through the mapping formula.

[0025] This scenario is adapted to situations where some external data sources only provide applicable time periods and arrival times. Through reasonable assumptions and statistical estimations, the mapping accuracy is ensured, and the error can be controlled within ±10 seconds, meeting the real-time requirements of traction power supply scheduling.

[0026] 3. Only applicable scenarios at the start time are provided. If the carbon intensity signal only indicates the start time of the applicable period but not the end time, the applicable period must be completed before mapping. First, the update period of the carbon intensity signal is estimated using the method described in Scenario 2 (missing release time scenario). Tᵤ .

[0027] Then, complete the end time of the applicable time period. tᵥ al ᵢ d,to,eq , tᵥ al ᵢ d,to,eq = tᵥ al ᵢ d,f ᵣ om + Tᵤ This means that the default applicable duration is equal to the update cycle, which conforms to the industry practice of "periodic validity" of carbon intensity data; Finally, construct the equivalent release time of the carbon intensity signal. t p ᵤ bl ᵢ sh,eq , t p ᵤ bl ᵢ sh,eq = tᵥ al ᵢ d,f ᵣ om - Tᵤ / 2, the subsequent process is the same as in scenario 2, and the time-stamp mapping is completed.

[0028] This scenario can cover the output format of some simplified carbon data interfaces, and fill in missing fields through industry practices to ensure the integrity and consistency of time stamp mapping.

[0029] Example explanation: If the release time of a certain carbon intensity signal t publish =10:00:00.000 (Data source clock), start time of the applicable time period t valid,from =10:00:00.000, the end time of the applicable time period. t valid,to =10:05:00.000, EMS receiving time t recv=10:00:08.000; Δ of the most recent 30 carbon intensity signals t i The median was 7.8 seconds after removing outliers. offset =8s; after mapping t valid,from,local =10:00:08.000, t valid,to,local =10:05:08.000, the current EMS control window is [10:02:00.000, 10:02:01.000]. Since 10:00:08.000≤10:02:00.000 and 10:02:01.000≤10:05:08.000, it is determined that the signal covers the current window.

[0030] It should be noted that all mapping operations in this step are completed locally in the EMS without changing the original data of the carbon intensity signal. Only the time scale of the multi-source data is unified through time axis conversion. The mapped time period field will be synchronously stored in the data structure of the carbon intensity signal to provide a unified time reference for subsequent coverage determination, jitter assessment and continuity assessment, and to ensure the objectivity of subsequent quality assessment.

[0031] Step 2: Determine whether the applicable time period of the aligned carbon intensity signal covers the current control window; if so, obtain the coverage index.

[0032] The purpose of this step is to ensure that the carbon constraint factor used in the current period (the core quantitative indicator for carbon constraint scheduling in the carbon intensity signal) is effective for the current control window, and to avoid using carbon constraint factors from expired periods or future periods.

[0033] The specific implementation method for this step is as follows: The EMS system stores the most recent aligned carbon intensity signals in a circular buffer, preferably 5 to 10 aligned carbon intensity signals, matching the carbon signal update cycle; the stored aligned carbon intensity signals are then mapped to the start time of the applicable time period. t valid,from,local Sort in ascending order, using a self-balancing binary search tree standard data structure (such as a red-black tree structure), and maintain the local valid start timestamp of the carbon signal after alignment with the EMS global clock. t valid,from,local Index; with t valid,from,local As the unique key of the red-black tree, node insertion is performed for newly added valid carbon signals, and node deletion is performed for expired or invalid carbon signals. The red-black tree automatically maintains structural balance during data addition and deletion, preventing performance degradation due to tree skew. This indexing method optimizes the time complexity of a single carbon signal retrieval to [missing information]. O ( log K ),in K This represents the total number of valid carbon signals currently maintained by the system; compared to the O(n) time complexity of traditional linear traversal retrieval... K The time complexity is reduced to logarithmic order, which can significantly shorten the retrieval time. In this scheme, the EMS system uses a fixed control cycle of 1 second, and all carbon signal retrieval and validity determination operations must be completed within a single cycle. O ( log K The retrieval efficiency ensures that the time consumed by related operations is much less than the duration of a single control cycle, which is suitable for the strong real-time requirements of DC traction power supply energy management and scheduling.

[0034] The coverage determination logic is as follows: for the current control window [ t k , t k + T c (This needs to be combined with the rail transit vehicle operation and scheduling plan, such as train arrival and departure times.) T c To control the period duration, retrieve the start times of all time periods that satisfy the applicable mapping time period. t valid,from,local ≤The start time of the current control window t k And the end time of the applicable mapping period t valid,to,local ≥ The end time of the current control window t k + T c If at least one carbon pre-intensity signal that meets the above constraints exists, a coverage index is generated. is covered =1 (coverage), otherwise generate coverage metrics. is covered =0 (not covered).

[0035] It should be noted that if the applicable period of the carbon intensity signal mapping only partially overlaps with the current control window, it will still be judged as not covered, in order to avoid scheduling deviations caused by insufficient effectiveness of the carbon constraint factor.

[0036] Example explanation: If the control cycle T c =1s, current control window [ t k =10:02:00.000,10:02:01.000); The start time of the applicable time period for mapping a certain carbon intensity signal in the cache. t valid,from,local=10:00:08.000, the end time of the applicable time period. t valid,to,local =10:05:08.000, determined because coverage conditions are met. is covered =1; if the mapping of another carbon intensity signal applies to the end of the time period t valid,to,local =10:02:00.500, if the window is not fully covered, then a judgment is made. is covered =0.

[0037] Coverage metrics is covered The "hard condition" for generating the gating state in subsequent step 3 is the coverage index. is covered =0, regardless of other quality indicators, the effective window gating status is directly determined. G CO2 =INVALID provides the core input for subsequent gating.

[0038] Step 3: Analyze the update interval jitter and continuity of the carbon intensity signal, and generate an effective window gating state by combining the coverage index.

[0039] The purpose of this step is to explicitly quantize the "carbon intensity signal quality" and form an executable gating system to prevent low-quality carbon signals from directly driving the scheduling closed loop and causing bus control oscillations.

[0040] The specific implementation method for this step is as follows: 1. Update interval jitter analysis First, calculate the arrival time interval Δ between adjacent carbon intensity signals. T update,i Δ T update,i = t recv,i - t recv,i-1 .in, i This is the sequence number of the carbon intensity signal. i ≥2; t recv,i Indicates the first i The timing of receiving the carbon intensity signal. t recv,i Indicates the first i -1 (previous) carbon intensity signal reception time.

[0041] Then, within the arrival time interval Δ of the most recent (preferably 10) adjacent carbon intensity signals T update Median (Δ) Tupdate ) and absolute median difference MAD(Δ T update If the latest arrival time interval Δ T update,i Exceeding [med(Δ T update )-3•MAD(Δ T update ),med(Δ T update )+3•MAD(Δ T update If )], it is determined to be an abnormal jitter ( is jitter,abnormal =1).

[0042] It should be noted that: if the external data source declares the update cycle of the carbon intensity signal... T u Additional thresholds can be set. J th = T u ×20% (e.g.) T u =300s J th =60s); At the same time, considering the tolerance for carbon signal update interval jitter caused by the frequent starts and stops of rail transit vehicles during peak hours, the first threshold and the second threshold are dynamically adjusted according to the vehicle operation period. At this time, if |Δ T update,i - T u |> J th This is also judged as abnormal jitter. A threshold is set. J th The aim is to provide a more practical engineering-oriented judgment benchmark for the "jitter anomaly" of carbon signal update interval, to make up for the limitations of relying solely on statistical methods (median + MAD), and to further improve the rigor and reliability of carbon signal quality assessment.

[0043] 2. Continuity Analysis If the mapping of the new carbon intensity signal applies to the start time of the time period t valid,from,local <Starting time of the mapping period applicable to the previous carbon intensity signal> t valid,from,local If so, it is determined to be out of order ( is out_of,order =1), count the out-of-order proportion of the most recent multiple (e.g., 10) carbon intensity signals, if the out-of-order proportion > the first threshold. O th (like O th=30%), then the continuity is judged to be poor; if the mapping of the new carbon intensity signal is applicable to the start time of the time period. t valid,from,local >The end time of the mapping period applicable to the previous carbon intensity signal t valid,to,local And the time difference between the two is greater than the second threshold. G th (Recommended) G th =2× T u If a segment is missing, it is marked as missing. is gap =1.

[0044] 3. Generate valid window gating states G CO2 In this embodiment, a valid window gating state is generated. G CO2 This embodiment provides the following three effective window gating states based on the priority of "coverage > continuity > jitter": First gating state V ALID This indicates that the carbon intensity signal quality is acceptable (coverage index). is covered =1. When there is no disorder, no missing segments, and no abnormal jitter, carbon-constrained scheduling can operate at full capacity without limiting the adjustment intensity to avoid signal quality risks. First gated state V ALID This ensures that carbon emission reduction is maximized while maintaining the bottom line of power supply safety through equipment constraints and safety verification.

[0045] The second gating state, DEGRADED1, indicates that there is a Type I anomaly in the carbon intensity signal (coverage indicator). is covered =1. When there is no disorder or missing segments, but abnormal jitter exists, carbon-constrained scheduling still participates in the decision-making process. However, to avoid the risk of low-quality carbon intensity signals, the adjustment intensity of the command is compressed by a first preset ratio (e.g., 30%) based on the normal maximum regulation capacity. In the second gating state DEGRADED1, power supply safety is prioritized, and carbon emission reduction targets are taken into account.

[0046] The third gating state, DEGRADED2, indicates the presence of a Type II anomaly (coverage indicator) in the carbon intensity signal. is covered =1, but the proportion of disordered items is greater than the first threshold. O th or is gapWhen =1), carbon-constrained scheduling still participates in decision-making regardless of the presence of abnormal jitter. However, to avoid the risk of low-quality carbon intensity signals, the adjustment intensity of the command is further compressed by a second preset ratio (e.g., 20%) based on the normal maximum regulation capacity. In the third gating state DEGRADED2, power supply safety is given higher priority, and only the basic carbon emission reduction target is achieved.

[0047] The fourth gating state, INVALID1, indicates the presence of a Type III anomaly (coverage indicator) in the carbon intensity signal. is covered =1, but the proportion of disordered items is greater than the first threshold. O th and is gap When =1), regardless of whether there is abnormal jitter, the carbon intensity signal does not participate in the carbon constraint scheduling decision at all. The system directly cuts off the influence of the carbon constraint factor on energy storage charging and discharging, converter voltage and power command, and switches to the pure voltage regulation mode or power regulation mode with power supply safety as the priority. This is equivalent to temporarily shutting down the carbon constraint function.

[0048] The fifth gated state, INVALID2, indicates a Type IV anomaly in the carbon intensity signal (as long as the coverage index is within the range). is covered When =0), regardless of whether the disorder ratio is greater than the first threshold. O th Whether there is a missing segment or abnormal jitter, the carbon intensity signal does not participate in the carbon constraint scheduling decision at all. The system directly cuts off the influence of the carbon constraint factor on energy storage charging and discharging, converter voltage and power command, and switches to a pure voltage regulation mode or power regulation mode that prioritizes power supply safety.

[0049] Based on the above five gating states, this embodiment follows the priority order of "coverage > continuity > jitter" to generate the effective window gating state. G CO2 The specific logic is shown in Table 1 below.

[0050] Table 1 shows the logical relationship between the five gating states and signal usage permissions: Coverage s_covered Continuity (out of order / missing segments) Is there any abnormal shaking? <![CDATA[Gating state G CO2 > Signal usage permissions 1 No disordered order and no missing segments 0 <![CDATA[ V ALID ]]> Normal participation in scheduling, without amplitude limitations. 1 No disordered order and no missing segments 1 DEGRADED1 Conservative participation in scheduling, with instruction range limited to 30%. 1 <![CDATA[The scrambling ratio ≤ 30% or the missing segment ≤ G th > any DEGRADED2 Conservative participation in scheduling, with instruction range limited to 20%. 1 <![CDATA[Out-of-order ratio > 30% or missing segment > G th > any INVALID1 Prohibited from participating in scheduling 0 any any INVALID2 Prohibited from participating in scheduling The valid window gating state generated in this step G CO2 This will serve as the core triggering condition for the state machine in subsequent steps: if the valid window gating state... G CO2 for V ALID Then carbon constraints are allowed to enter the loop normally; if the effective window is under gating... G CO2If it is DEGRADED1 or DEGRADED2, a downgraded entry into the loop is triggered (e.g., reducing the carbon constraint weight or extending the hold time); if the effective window gating state is... G CO2 If the value is INVALID1 or INVALID2, a freeze to exit carbon constraints will be triggered to ensure safety takes priority.

[0051] Step 4: If the effective window gating state allows participation in scheduling, predict the net power exchange increment of the power grid based on the candidate control actions of this cycle; determine whether the absolute value of the net power exchange increment of the power grid exceeds the significance threshold obtained from the statistics of historical stable operation data; if so, select the marginal emission factor as the carbon constraint factor, otherwise select the average emission factor as the carbon constraint factor.

[0052] The purpose of this step is to avoid using mismatched carbon constraint factors for scheduling due to incorrect selection of the average emission factor (AEF) or marginal emission factor (MEF), which could lead to scheduling direction deviations or artificially high or low carbon gains, thereby inducing unnecessary power disturbances.

[0053] The specific implementation method for this step is as follows: First, the system's stable operating periods are screened from the historical operating data of SCADA. In this embodiment, the determination of stable operation requires the simultaneous fulfillment of three quantitative conditions: First, there are no overvoltage, undervoltage, or overcurrent protection actions triggered within 5 consecutive minutes; second, the DC bus voltage fluctuation amplitude is ≤5V (this threshold is set based on the P90 quantile of historical operating data and can cover 90% of the stable operating voltage fluctuation range); and third, the energy storage device's charging and discharging power fluctuation amplitude is ≤10% of the rated power.

[0054] Based on the selected stable operating period data, the noise standard deviation σ of the net exchange power between the power grid and the traction power supply system was statistically analyzed. P (This standard deviation reflects the natural power fluctuation level during normal system operation); then, based on the statistical 3σ principle, a significance threshold ε for the power increment is set, ε = 3⋅σ. P The statistical significance of this threshold is that it covers 99.7% of the normal power fluctuation range. Its core function is to define the degree of impact of control actions on the net power exchange of the grid—only when the absolute value of the net power exchange increment predicted by the candidate control action is |Δ P grid,pred When |≥ε, the control action is determined to be an incremental regulation behavior that "significantly affects grid power," and the marginal emission factor (MEF) is selected for carbon-constrained scheduling; conversely, if |Δ P grid,pred If | < ε, it is determined that the control action belongs to the internal energy transfer or fluctuation within the noise range. The average emission factor (AEF) is selected for carbon statistical attribution, thereby achieving a precise match between the carbon constraint factor type and the nature of the control action.

[0055] Subsequently, based on the candidate control actions to be executed in this cycle, the incremental Δ of the candidate control actions on the net power exchange of the power grid is predicted. P grid,pred Based on the real-time power balance relationship of the power system, the change in net power exchange of the grid is jointly determined by the power command adjustment of the controllable units within the system and the power fluctuation of uncontrollable renewable energy. Therefore, the quantitative calculation formula can be obtained: Δ P grid,pred ≈Δ P inv,cmd +Δ P ess,cmd −Δ P regen,est In the formula, Δ P inv,cmd Δ represents the change in converter power command, specifically the difference between the converter power command value corresponding to the candidate control action in the current cycle and the final power command value actually executed in the previous cycle. Its positive or negative sign indicates the direction of increase or decrease in converter power (positive for power increase, negative for power decrease). P ess,cmd This represents the change in power command for the energy storage system, specifically the difference between the energy storage charge / discharge power command value corresponding to the candidate control action in the current cycle and the final charge / discharge power command value actually executed in the previous cycle. Its sign must be determined in accordance with the energy storage operating mode definition (negative for charging, positive for discharging, or vice versa, consistent with the system power flow calibration rules); Δ P regen,est This parameter represents the real-time power fluctuation of uncontrollable renewable energy sources such as photovoltaic and wind power. It is calculated by using the most recent (e.g., 10 seconds) measured value of vehicle braking regenerative power and filtering out high-frequency noise through time smoothing to ensure the stability and reliability of the estimation of regenerative power change trend.

[0056] It should be noted that the approximate relationship of the above quantitative calculation formula is based on the engineering assumption that "the change in internal system losses is much smaller than the power adjustment of controllable units". This assumption is fully reasonable in the scenario of coordinated scheduling of electrochemical energy storage and converter, and can meet the accuracy requirements of carbon constraint factor selection and scheduling decision.

[0057] Finally, an adaptive carbon constraint factor selection process is executed once per scheduling cycle, with the following specific rules: 1. Two-factor optional scenarios When both the average emission factor (AEF) and the marginal emission factor (MEF) are provided by the external data source: If the absolute value of the predicted increase in net power exchange in the power grid satisfies |Δ P grid,predIf |≥ε, then the candidate control action in this cycle is determined to significantly alter the scale of energy interaction between the power grid and the system (such as high-power charging and discharging of energy storage, or significant adjustment of converter power commands). The marginal emission factor (MEF) is chosen in this case because it accurately characterizes the marginal rate of change of carbon emissions per unit incremental power, which is highly compatible with the carbon emission calculation requirements of "incremental actions," avoiding the carbon emission accounting bias caused by using the statistically averaged attribute (AEF).

[0058] If the absolute value of the predicted increase in net power exchange in the power grid satisfies |Δ P grid,pred If | < ε, then the candidate control action for this cycle is determined to be within the power fluctuation noise range or an internal energy transfer within the system (such as minor adjustments to energy storage or slight corrections to converter power commands, without changing the overall energy interaction scale between the system and the grid). The average emission factor (AEF) is chosen in this case because it possesses statistical stability, effectively filtering out power fluctuation noise from minor actions and improving the reliability of carbon emission calculations.

[0059] It should be noted that the threshold ε is a preset threshold for determining the significance of power increment. Its value needs to be calibrated in conjunction with the system's rated power and the fluctuation characteristics of renewable energy. Its value range is preferably 0.5% to 2% of the system's rated power to ensure accurate differentiation between "significant actions" and "fine-tuning actions".

[0060] 2. Single-factor optional scenario When the external data source does not simultaneously provide both the average emission factor (AEF) and the marginal emission factor (MEF), and only outputs a single carbon intensity index (CI): Select carbon strength value directly CI As the basis for the scheduling calculation in this cycle, the factor type (such as ONLYONE) is marked in the system log to clarify that the current factor selection is limited by the data source capability, rather than an active selection based on action characteristics.

[0061] 3. Requirements for Selecting Results Recording and Traceability Each scheduling cycle requires the synchronous recording of the carbon constraint factor selection results and calculation basis. The carbon constraint factor selection results include: the type and specific value of the selected carbon constraint factor; the calculation basis includes: |Δ P grid,pred | Calculated values, threshold ε, and the types of carbon constraint factors provided by the data source are used to form a complete traceability ledger for factor selection. The ledger must meet the requirements of being searchable and traceable, providing data support for subsequent carbon emission accounting audits, scheduling strategy optimization, and factor selection rule iteration.

[0062] This step outputs the average emission factor (AEF) or marginal emission factor (MEF) through a significance discrimination rule. These will be directly used as the core carbon constraint input parameters in the candidate command generation stage of the subsequent step 5. This ensures that the carbon emission reduction constraint direction of the candidate command is accurately matched with the carbon emission impact mechanism of the scheduling action, thus avoiding the mismatch between the average emission factor and the marginal emission factor from the source of carbon constraint factor selection.

[0063] Step 5: Based on the relative position of the selected carbon constraint factor in the historical time series distribution, and combined with the current adjustable resource constraints, generate candidate scheduling instructions that include energy storage charging and discharging power or converter voltage regulation.

[0064] The purpose of this step is to transform the qualitative goals of carbon constraints into quantitative instructions that can be executed by the equipment, while taking into account the physical constraints of the equipment and ensuring the feasibility of the instructions.

[0065] The specific implementation method for this step is as follows: First, calculate the relative position coefficients used to characterize the current carbon intensity level. s — (1) Statistical quantiles: Collect and statistically analyze the effective carbon constraint factors in the most recent 24 hours. CI selected Generate effective carbon constraint factors CI selected Time series data; based on effective carbon constraint factor CI selected Calculate the 10th percentile (denoted as 10%) from the time series data. CI low ) and 90th percentile (denoted as CI high It should be noted that the above effective carbon constraint factors... CI selected That is, the average emission factor AEF or marginal emission factor MEF output in step 4; selecting the 10th percentile or 90th percentile instead of extreme values ​​can effectively filter extreme outliers (such as abrupt changes caused by communication interference) in the carbon constraint factor time series data, ensuring the stability of the baseline interval. (2) Normalization calculation: based on statistical results CI low and CI high The current carbon constraint factor CI selected Normalized to the [0,1] interval, the relative position coefficient of carbon intensity, s, is obtained. The normalization formula is: ,in, clip (·) is a cutoff function used to restrict the calculation result to the interval [0,1]. That is, s=0 when the numerator is negative and s=1 when the calculated value is greater than 1. s =1, to avoid the current carbon constraint factor CI selectedExceeding [ CI low , CI high The interval can cause distortion in the normalization results.

[0066] It should be noted that the value of the carbon intensity relative position coefficient s directly maps to the carbon intensity scheduling strategy: when s When =0, it indicates that the current carbon intensity is in the low-carbon range of the past 24 hours, and the dispatch strategy is to prioritize purchasing electricity from the grid and charging it through energy storage systems to maximize the utilization of low-carbon electricity; when s When =1, it indicates that the current carbon intensity is in the high carbon range of the last 24 hours. The dispatch strategy is to prioritize the release of energy storage power, reduce the amount of electricity purchased from the grid, and reduce the carbon emissions from electricity purchases during the high carbon emission phase.

[0067] Then, combining the operating status of the energy storage system with the constraints of the equipment itself, the flexible power regulation capability is calculated, specifically taking into account the state of charge of the energy storage system. SOC Battery temperature T ess and rated power P rated,ess The formula for calculating the available charging power under these three constraints is as follows: The corresponding formula for calculating the available discharge power is: ;in, SOC max This represents the maximum permissible state of charge for the energy storage system. SOC min The minimum permissible state of charge for an energy storage system. SOC max and SOC min These are the safety operating boundary parameters for energy storage batteries; f ( T ess This is a battery temperature correction factor used to compensate for the impact of temperature on battery charge and discharge performance. Based on this, the flexible power amplitude of the energy storage system is calculated. P flex =min( P chg,avail , P dis,avail By taking the minimum value between the available charging power and the available discharging power, the energy storage's bidirectional charging and discharging regulation capabilities are kept balanced, avoiding the failure of scheduling command execution due to excessive unidirectional power margin and insufficient margin in the other direction. This provides reliable flexible regulation capability boundary parameters for the generation of subsequent carbon-constrained scheduling commands.

[0068] It should be further noted that, based on engineering practice of electrochemical energy storage systems and battery thermal characteristics, f ( Tess The value of ) is determined based on a clearly defined temperature sensitivity threshold for battery charge and discharge performance, specifically as follows: The temperature thresholds (45℃ and 55℃) are based on the "upper limit of normal operating temperature (45℃)" and "lower limit of thermal runaway warning temperature (55℃)" specified in the general technical manual for energy storage batteries and the national standard "Lithium-ion Batteries for Electric Energy Storage" (GB / T36276-2018). T ess ≤45℃ is the optimal operating temperature range. This range represents the high-efficiency charge-discharge range for mainstream energy storage batteries, characterized by low internal resistance, weak polarization, and the ability to reach rated charge-discharge power. Therefore, a correction factor of 1 is applied, and there is no power attenuation constraint. 45℃ < T ess ≤55°C is the performance degradation range. When the temperature exceeds 45°C, the rate of side reactions inside the battery accelerates, the internal resistance increases, and continuous full-power charging and discharging will accelerate battery aging. At the same time, this temperature has not reached the thermal runaway warning threshold, and it is permissible to operate at 50% of the rated power. Therefore, the correction factor is taken as 0.5 to achieve reasonable power limitation. T ess >55℃ is the safe protection range. When the temperature exceeds 55℃, the battery separator is prone to thermal shrinkage, which poses a risk of short circuit and thermal runaway. According to the mandatory requirements of the "Safety Regulations for Electrochemical Energy Storage Power Stations" (GB / T42288-2022), the charging and discharging circuit must be cut off. Therefore, the correction factor is taken as 0, and the battery is prohibited from participating in power regulation.

[0069] Next, based on the effective carbon constraint factor and energy storage flexible power amplitude determined by the aforementioned screening, candidate power commands for the energy storage system are generated. P ess,cand —(1) Define the rules for positive and negative signs of instructions: Define P ess,cand >0 indicates that the energy storage system is in a charging state, corresponding to an increase in the power purchased by the grid. P ess,cand <0 indicates that the energy storage system is in a discharging state, corresponding to a reduction in the power purchased by the grid. The symbol definition is consistent with the power flow direction of the grid and the carbon emission reduction scheduling requirements. (2) The basic calculation formula for the candidate power command is: P ess,cand =(0.5− s )×2× P flex In the formula, s is a carbon intensity status indicator parameter, the value of which is related to the current carbon intensity level of the power grid (e.g., s=0 represents a low-carbon period, s=1 represents a high-carbon period), and is used to distinguish the direction of scheduling strategies; P flexThe energy storage flexible power amplitude calculated above is the equilibrium constraint boundary of the available power for charging and discharging, ensuring that the command does not exceed the actual adjustment capability of the energy storage system.

[0070] Finally, the mass gating state is combined with the carbon intensity signal. G CO2 Implement hierarchical constraints on the instruction amplitude: (1) If the quality gate state G CO2 =DEGRADED1 or DEGRADED2 (i.e., the carbon intensity signal covers the current control window, but has quality defects such as slight jitter, a small amount of out-of-order signal, or short-duration missing segments), and needs to be processed by the amplitude truncation function ( clip The function limits the absolute value of the candidate power command to 30% of the base command amplitude. The specific constraint logic is as follows: P ess,cand,limited = clip ( P ess,cand ,−0.3×∣ P ess,cand |,0.3×| P ess,cand |). By limiting the amplitude, the "conservative participation" mode of carbon-constrained scheduling is realized, which retains the right to participate in carbon emission reduction scheduling while avoiding the risk of power fluctuations that may be caused by low-quality carbon signals. (2) If the quality gate state G CO2 = V ALID Then the candidate power command directly uses the basic calculation result without additional amplitude restrictions. (3) If the quality gate state G CO2 If INVALID1 or INVALID2 is selected, the carbon constraint scheduling logic is frozen, and the energy storage system executes the instructions according to the normal safe operation strategy.

[0071] In addition, it should be noted that for application scenarios without energy storage, the scheduling strategy only generates voltage and / or power coordinated control commands for the converter. The specific command generation logic and parameter setting rules are as follows: 1. Set voltage command reference value Converter voltage command reference value V ref The voltage rating is directly set according to the rated value of the system bus voltage, with an example value of 850V. The setting of this reference value must strictly match the grid-side voltage level standard to ensure that the converter output voltage is consistent with the grid operation requirements.

[0072] 2. Calculate the voltage fine-tuning amount Voltage adjustment Δ V cand Sensitivity curve based on device power-voltage kvp The quantitative conversion of power deviation into voltage regulation is achieved, and the calculation formula is: Δ V cand = k vp ×Δ P grid,pred In the formula, k vp The power-voltage sensitivity coefficient of the converter (unit: V / kW) characterizes the converter voltage regulation amplitude corresponding to a unit amplitude of grid power fluctuation. This coefficient needs to be obtained through on-site offline identification or online adaptive correction to ensure its matching degree with the actual operating characteristics of the equipment; Δ P grid,pred The power prediction deviation value is the difference between the predicted power and the actual power. Its positive or negative sign corresponds to different operating conditions of power surplus or shortage.

[0073] 3. Synthesize the final candidate voltage command The final candidate voltage command for the converter is obtained by superimposing the reference value and the fine-tuning amount, and the expression is: V cand = V ref +Δ V cand The generated V cand The converter voltage output range constraint (such as ±5% of the rated value) must be met simultaneously. If the constraint boundary is exceeded, the amplitude is truncated to ensure that the command is within the safe operating range of the equipment.

[0074] The energy storage candidate power command generated in this step P ess,cand and converter candidate voltage command V cand This will be directly used as the input parameter for instruction security verification and shaping in step 6; in addition, the aforementioned energy storage candidate power instructions P ess,cand and converter candidate voltage command V cand The device must pass the safety margin verification of the device itself, the operation boundary verification of the grid side, and the smoothing and shaping of the command in sequence. After the verification is passed and the shaping and optimization are completed, it can be used as the final executable command to be issued to the execution unit of the energy storage system and the converter, so as to ensure the safe operation of the device and the stable operation of the grid during the execution of the command.

[0075] Step 6: Calculate the voltage safety margin of the DC bus voltage relative to the protection trip threshold, and obtain the effective safety margin after deducting measurement noise redundancy; obtain the allowable voltage change based on the effective safety margin; based on the allowable voltage change and the equipment ramp-up capability, perform slope limiting and smoothing shaping on the candidate scheduling instructions to obtain the final execution instructions.

[0076] The purpose of this step is to establish a quantifiable safety isolation zone between the carbon-constrained candidate control action and the DC traction power supply safety boundary, so as to prevent the candidate command from causing protection malfunctions or voltage oscillations near the boundary.

[0077] The specific implementation method for this step is as follows: First, calculate the voltage safety margin. M V The calculation formula is: M V =min( V trip,high - V dc , V dc - V trip,low ),in, V trip,high This is the overvoltage protection trip threshold. V trip,low The undervoltage protection trip threshold (complies with the requirements of GB50157 "Code for Design of Metro"). V dc This is the measured value of the bus voltage. M V It directly reflects the minimum margin between the current voltage and the tripping threshold.

[0078] Then, the measurement noise confidence redundancy is calculated. U V The calculation formula is: U V =3·σ V , where σ V The standard deviation of voltage measurement noise is obtained from historical data during stable periods, ensuring coverage of 99.7% of measurement fluctuations.

[0079] Next, based on voltage safety margin M V and noise redundancy U V Calculate the effective safety margin M V,eff The calculation formula is: M V,eff = M V - UV If the effective safety margin M V,eff ≤0 indicates that the voltage is approaching the safety boundary, and voltage adjustment should be prohibited. Therefore, an allowable voltage change Δ is set. V allow =0; if the effective safety margin M V,eff If the voltage change is greater than 0, then the allowable voltage change Δ V allow The minimum value between the equipment's climbing ability and effective safety margin can be taken, i.e., Δ V allow =min( R amp,V × T c , M V,eff ),in, R amp,V The voltage ramp-up capability of the converter (provided by the equipment's factory specifications). T c To control the cycle.

[0080] To ensure that the voltage regulation process of the converter meets the requirements for safe operation of the equipment and the system stability constraints, based on the preset allowable voltage change Δ V allow For the generated candidate voltage command V cand Shaping processing is performed, including amplitude limiting and smoothing filtering—(1) Calculate the voltage command difference. Select the final execution voltage command of the previous control cycle. V cmd,prev As a benchmark, the difference Δ between the current candidate voltage command and the benchmark command is calculated. V cand Δ V cand = V cand - V cmd,prev The difference represents the adjustment range of the candidate instruction relative to the historical instruction, and serves as the basis for subsequent limiting processing. (2) Based on the allowable voltage change Δ V allow Candidate voltage command V cand Amplitude limiting – via amplitude truncation function ( clip The function calculates the difference Δ between the current candidate voltage command and the reference command. V cand Limited to ±Δ V allow Within the safe range, to avoid the risk of overvoltage / undervoltage of equipment caused by excessively large single voltage adjustment, the voltage adjustment amount after limiting is Δ.V clipped Δ V clipped = clip (Δ V cand ,−Δ V allow ,+Δ V allow Based on the voltage adjustment amount Δ after limiting. V clipped The initial shaping voltage command is calculated. V cmd,temp , V cmd,temp = V cmd,prev +Δ V clipped (3) Initial shaping voltage command V cmd,temp A first-order low-pass filter is used for smoothing to avoid system voltage oscillations caused by the step change in the initial shaping command. The calculation formula for the first-order low-pass filter is: V cmd =α× V cmd,prev +(1−α)× V cmd,temp Where α is the filter coefficient, and its value is determined by the control cycle. T c Determined by the target response time constant τ, i.e. It should be noted that the time constant... τ The preferred value is 5s. This value can achieve the optimal balance between response speed and adjustment smoothness, ensuring that the command can quickly track the changing trend of the power deviation of the power grid, and effectively suppressing high-frequency voltage fluctuations, thus avoiding equipment losses and system oscillations caused by frequent operation of the converter.

[0081] Furthermore, the shaping logic of the power command of the energy storage system is consistent with the shaping logic of the voltage command of the converter—(1) Calculate the power safety margin. Take into account the operating temperature of the converter or energy storage device and the state of charge of the energy storage ( SOC The dual constraints define the power safety margin. M P The calculation formula is as follows: In the formula, Prated is the rated power of the converter or energy storage system, which is the reference boundary for power regulation; T Tmax is the real-time operating temperature of the equipment; Tmax is the maximum allowable operating temperature of the equipment. This is the temperature margin coefficient; the closer the temperature is to the threshold, the smaller the margin and the stricter the constraint. State of charge for energy storage SOCThe power safety margin is calculated by adjusting the state of charge (SBC) coefficient. The closer the SBC is to the upper limit, the smaller the charging safety margin, thus avoiding the risk of overcharging. The aforementioned power safety margin... M P The calculation formula ensures that the power regulation is always within the safe operating range of the equipment by coupling the constraints of temperature and energy storage state of charge, thus avoiding the risk of power surges under high temperature or extreme energy storage state of charge. (2) Define the allowable power variation. The allowable power variation Δ is calculated by combining the equipment ramp rate constraint and the safety margin. P allow The calculation formula is: Δ P allow =min( R amp,P × T c , M P - U P In the formula: R amp,P The maximum permissible ramp rate of the equipment (unit: kW / s). T c To control the cycle duration, R amp,P × T c The maximum power variation limit based on ramp capability within a single control cycle; U P The uncertainty of the current power fluctuation (introduced by power disturbances or measurement errors on the grid side). M P - U P The actual safety margin after deducting disturbance margin; allowable power variation Δ P allow Pick R amp,P × T c and M P - U P The minimum value between is used as the allowable power variation to achieve ramp rate constraint and safety margin constraint. (3) Generate the final power command. A first-order low-pass filtering algorithm is used, combined with an amplitude truncation mechanism, to generate the final executable power command. P cmd The calculation formula is: P cmd =α× P cmd,prev +(1−α)×( P cmd,prev + clip ( Pess,cand - P cmd,prev ,−Δ P allow ,+Δ P allow In the formula, α is the filtering coefficient (with a value range of 0≤α≤1), which is used to balance the response speed and smoothness of the instruction. The closer α is to 1, the smoother the instruction is, but the greater the response delay. P cmd,prev The power command execution value of the previous control cycle is used to ensure the time continuity of the command; clip (·) is the truncation function, which calculates the difference between the candidate instruction and the instruction in the previous cycle. P ess,cand - P cmd,prev ) limited to ±Δ P allow Within the specified range, ensure that a single power adjustment does not exceed the allowable variation. By introducing a first-order filter, command jitter caused by grid-side disturbances and carbon signal fluctuations can be effectively suppressed, thereby improving the operational stability of the energy storage system.

[0082] After command shaping is completed, the final output converter voltage command is... V cmd and energy storage power command P cmd Perform safety boundary verification and correction to ensure that the instructions strictly match the safety operation threshold of the equipment and the grid connection specifications—(1) For converter voltage instructions V cmd Apply safety boundary constraints. Voltage command. V cmd The required safety boundary constraint range is: V cmd ∈[ V trip,low + U V , V trip,high - U V In the formula, V trip,low The threshold for low voltage protection operation of the converter. V trip,high The high voltage protection threshold for the converter. V trip,low and V trip,high These are the core safety parameters set at the factory for the equipment. Once the voltage reaches the threshold, the protection system will trigger a trip and shutdown. U VFor voltage safety margin, it is used to avoid the risk of commands approaching the protection threshold caused by factors such as grid voltage fluctuations and measurement errors, and to ensure the stability and redundancy of voltage regulation. The upper and lower limits of the above safety boundary constraint range are set by “protection threshold ± safety margin”, which not only reserves sufficient voltage regulation space, but also avoids protection malfunctions from the root. (2) For energy storage power commands P cmd Apply safety boundary constraints. Energy storage power command. P cmd The required safety boundary constraint range is: P cmd ∈[− P dis,avail , P chg,avail In the formula, P chg,avail Available power for charging energy storage, − P dis,avail The negative value of the available power for energy storage discharge corresponds to the direction of the discharge power flow. This range directly matches the actual charge and discharge capacity boundary of the energy storage system under the current operating conditions, ensuring that the power command does not exceed the flexible adjustment range and avoiding the risk of damage to battery life such as overcharging, over-discharging, or high temperature overload. (3) For the converter voltage command V cmd and energy storage power command P cmd Make corrections. On the one hand, if V cmd < V trip,low + U V Then the converter voltage command V cmd Revised to V trip,low + U V ;like V cmd > V trip,high - U V Then the converter voltage command V cmd Revised to V trip,high - U V On the other hand, if P cmd <− P dis,avail Then the energy storage power command will be sent. P cmd Corrected to - P dis,avail ;likeP cmd > P chg,avail Then the energy storage power command will be sent. P cmd Revised to P chg,avail The above modifications ensure that output commands remain within a safe and feasible range, providing a final technical guarantee for stable equipment operation and reliable grid connection.

[0083] Step 6: Calculate the power safety margin M P Noise redundancy U V With the allowable voltage change Δ V allow By limiting candidate commands to a safety margin beyond noise redundancy, the risk of protection malfunction and oscillation is reduced; when the allowable voltage change Δ V allow When the value approaches 0, step 7 can be triggered to enter a frozen or downgraded state, thus providing a safety net.

[0084] Step 7: Construct a state machine based on the effective window gating state, voltage safety margin, and protection event count; when the state machine is in a frozen state, exit the carbon constraint scheduling and execute the instruction maintenance or voltage stabilization and limiting action; when the state machine is in a recovering state, linearly release the restriction on the change of the instruction over time.

[0085] The purpose of this step is to dynamically adjust the carbon constraint participation by constructing a state machine to ensure power supply safety as a priority, while achieving a smooth transition after anomaly recovery.

[0086] The specific implementation method for this step is as follows: The state machine includes four states: NORMAL, DEGRADED, FROZEN, and RECOVERING. Referring to step 3 above, the state triggering conditions are defined as follows: (1) The triggering condition for the NORMAL state is: carbon constraints are normally involved in scheduling, the instruction amplitude is only constrained by the equipment itself without additional restrictions, and the carbon constraint control enable flag is displayed. enable carbon,control Set to true; (2) The triggering condition for the DEGRADED state is: carbon constraint degradation participates, the command amplitude is limited to 30%~50% of the normal maximum adjustment, and the carbon constraint control enable flag is displayed. enable carbon,controlSet to true, filter coefficient α=0.9 (enhancing smoothness); (3) The trigger condition for the FROZEN state is: the system completely exits carbon constraint scheduling, switches to voltage regulation priority control strategy, keeps the current voltage command and power command stable, and the carbon constraint control enable flag is displayed. enable carbon,control Set to false; (4) The trigger condition for the recovery state RECOVERING is: release the instruction amplitude limit and the carbon constraint control enable flag. enable carbon,control Maintaining the value as true, the amplitude limitation is linearly released by introducing a recovery coefficient β. The formula for calculating β is: β = min(1,( k - k 0) / K 2), where, k This represents the current control cycle number. k 0 represents the number of cycles the system takes to enter recovery mode. K 2. To determine the number of cycles required for recovery (e.g., set to 60 control cycles), ensure that β gradually increases linearly from its initial value to 1, thereby achieving a smooth recovery in carbon constraint participation.

[0087] The switching methods between the above four states are as follows: (1) The triggering condition for switching from the NORMAL state to the DEGRADED state is: any one of the following conditions is met: 1.1, 1.2, or 1.3; where, condition 1.1: valid window gating state. G CO2 from V ALID Change to DEGRADED1 or DEGRADED2; Condition 1.2: Effective voltage safety margin M V,eff Below the first voltage threshold (e.g., 20V); Condition 1.3: Protection event count within the most recent preset time period (e.g., within the most recent 10 seconds). N prot Among the three conditions, it exceeds the P80 quantile of historical running data.

[0088] (2) The triggering condition for switching from the DEGRADED state to the FROZEN state is: satisfying any of conditions 2.1, 2.2, or 2.4; where, condition 2.1: the effective window gating state of the carbon intensity signal. G CO2 To become INVALID, condition 2.2: Effective voltage safety margin M V,eff If the voltage drops to the second voltage threshold (e.g., 0V) or below, condition 2.3: the number of protection events within the most recent preset time period (e.g., within the most recent 10 seconds). Nprot Exceeding the P95 quantile of historical operating data, condition 2.4: Allowable voltage variation Δ over three consecutive control cycles. V allow =0.

[0089] (3) The triggering condition for switching from the FROZEN state to the RECOVERING state is: conditions 3.1 to 3.3 are met simultaneously; where condition 3.1 is: the effective window gating state for multiple consecutive control cycles (e.g., 30). G CO2 Keep V ALID Condition 3.2: Effective voltage safety margin for multiple consecutive control cycles (e.g., 5 cycles). M V,eff Not lower than the first voltage threshold (e.g., 20V), condition 3.3: the number of protection events within the most recent preset time period (e.g., within the most recent 10 seconds). N prot It does not exceed the P50 percentile of historical operating data.

[0090] (4) The triggering condition for switching from RECOVERING state to NORMAL state is: when conditions 4.1 to 4.2 are met; where, condition 4.1 is: the recovery coefficient β reaches 1, and condition 4.2 is: the effective voltage safety margin of multiple consecutive (e.g., 10) control cycles. M V,eff Not lower than the third voltage threshold (e.g., 30V). (5) The triggering condition for switching from DEGRADED state to NORMAL state is: all triggering conditions of DEGRADED state are eliminated, and the effective voltage safety margin is maintained for multiple (e.g., 10) consecutive control cycles. M V,eff Not lower than the third voltage threshold (e.g., 30V).

[0091] When the above four operating states switch, the corresponding state switch reason code is recorded synchronously. reason code The reason code reason code Based on the core factors that trigger the switching, the output should be in the format of "state source-trigger type-condition number". For example, the state switching reason code. reason code =01-01-01 corresponds to the state transition scenario of NORMAL→DEGRADED (carbon signal degradation), indicating the initial state is NORMAL and the carbon intensity signal is... G CO2 from V ALID The state changes to DEGRADED, triggering a downgrade; state transition reason code. reasoncode =01-02-01 corresponds to the state transition scenario from NORMAL to DEGRADED (insufficient voltage margin), indicating the initial state as NORMAL and the effective voltage safety margin. M V,eff <20V triggers degradation, etc. State transition reason codes for other state transition scenarios. reason code For the corresponding meanings, please refer to Table 2.

[0092] Table 2 lists the state transition reason codes for each state transition scenario. reason code A table comparing the meanings of the symbols: It should be noted that the carbon constraint control enable flag output by the state machine... enable carbon,control Recovery coefficient β and state transition reason code reason code This needs to be synchronized in real time to steps 5, 6, and the subsequent step 8, achieving coordinated actions across all stages through parameter linkage, and constructing a closed-loop control system with safety as a safety net. Among these, the carbon constraint control enable label... enable carbon,control As the core switching signal for carbon constraint scheduling, it works in conjunction with the recovery coefficient β in the step 5 command amplitude limitation stage—in the NORMAL state, β=1, subject only to hard equipment constraints with no additional amplitude limitation; in the DEGRADED state, β corresponds to an amplitude limitation ratio of 30%~50%, precisely controlling the weak participation of carbon constraints; in the RECOVERING state, β linearly increases to 1 with the control cycle, smoothly releasing the command amplitude to adapt to the system recovery process. Simultaneously, the carbon constraint control enable flag... enable carbon,control In conjunction with the filter coefficient α, it works in the filter coefficient adjustment step 6. In the DEGRADED state, α is simultaneously set to 0.9 to enhance command smoothness. In the NORMAL and RECOVERING states, the α value is adapted as needed to avoid system disturbances caused by command jitter. Additionally, the state switching reason code... reason code The log recording process in step 8 is synchronized in real time, and the triggering reasons, time nodes and corresponding parameters of the state switch are fully retained. This provides verifiable evidence for fault tracing and strategy optimization, and finally forms a closed-loop control of "state perception - parameter output - collaborative action - reverse tracing", which takes into account both carbon emission reduction targets and power supply safety and stability.

[0093] Step 8: Record the aligned carbon intensity signal timestamp, effective window gating status, factor selection, candidate instruction, final instruction, safety margin, state machine status, and cause code in each control cycle to form a traceable log.

[0094] The purpose of this step is to address the "unexplainable" problem of carbon constraint scheduling through full-process logging, providing support for operational review and compliance auditing.

[0095] The specific implementation method for this step is as follows: Periodic records include time information ( t cycle ), carbon intensity signal information ( CI selected , t valid,from,local , G CO2 , is covered ), quality assessment information (Δ T update , is jitter,abnormal , is out_of,order Factor selection information () factor type Δ P grid,pred , ε), instruction information ( P ess,cand , V cand , P cmd , V cmd Safety margin information ( V dc , M V , M V,eff , N prot ), state machine information (state, reason code , β), Equipment status information ( SOC , T ess Core fields such as ) are stored in CSV or JSON format.

[0096] Log data can be used as training data for subsequent machine learning optimization of carbon-constrained scheduling strategies, while also meeting carbon emission reduction compliance audit requirements.

[0097] In summary, this embodiment provides a carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method. Based on the core control logic of a four-state adaptive state machine, it constructs a complete technical system encompassing "precise carbon factor selection, flexible power quantification of energy storage, hierarchical command generation and shaping, dynamic state switching, and closed-loop traceability." This method first combines the grid power increment characteristics and scenario requirements to output the target carbon factor type as the carbon constraint input for candidate command generation; secondly, it integrates energy storage... SOC Temperature and rated power constraints are considered to calculate the available charging and discharging power and flexible power amplitude. Simultaneously, a converter voltage / power coordinated control strategy is designed for scenarios without energy storage. Then, command shaping is achieved through safety margin constraints, ramp rate limits, and first-order filtering algorithms, followed by boundary checks and forced corrections. Finally, relying on an adaptive state machine containing four states—NORMAL, DEGRADED, FROZEN, and RECOVERING—the carbon constraint participation is dynamically adjusted based on quantification conditions, and the state machine output parameters are simultaneously updated. enable carbon,control ,β, reason code This method links the control to command amplitude limits, filter coefficient adjustments, and log recording, forming a closed-loop control with a safety fallback. On one hand, through precise carbon factor matching and hierarchical command control, this method effectively improves the carbon emission reduction capability of the DC traction power supply system, achieving a deep integration of carbon intensity constraints and energy management. On the other hand, it leverages multi-dimensional safety constraints (temperature, ... SOC The dynamic switching mechanism between voltage and power boundaries and state machines ensures the core objective of prioritizing power supply safety, mitigating scheduling risks in scenarios such as abnormal carbon signals and equipment deterioration. Simultaneously, through... reason code The full-process traceability improves the convenience and controllability of system operation and maintenance, and ultimately achieves synergistic optimization of carbon emission reduction benefits and power supply safety and stability.

[0098] The above-mentioned effective window gating state GCO2 belongs to the input side signal quality state; the state machine state belongs to the system operation control state, and the state machine performs state transitions based on GCO2.

[0099] Example 2: This example provides a carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance system, used to perform carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance as described in Example 1, including: The signal processing module is used to acquire external carbon intensity signals and perform time-scale alignment using time offsets; The gating and factoring module is used to generate effective window gating states and adaptively select carbon constraint factors based on the significance of the net power exchange increment in the power grid. The instruction generation module is used to generate candidate scheduling instructions based on the carbon constraint factor. The safety shaping module is used to calculate voltage and power safety margins and to perform slope limiting and smoothing shaping on candidate scheduling commands based on the allowable changes. The state machine control module is used to adjust the system operating state according to the gating state and safety margin, and to enforce the voltage stabilization priority strategy in the frozen state.

[0100] Example 3: This example provides a computer-readable storage medium that stores instructions that include the method described in Example 1 or any other method that may involve the method described in Example 1. Specifically, the computer-readable storage medium stores instructions that, when executed on a computer, perform the method described in Example 1 or any other method that may involve the method described in Example 1. The computer-readable storage medium refers to a data storage medium, which may include, but is not limited to, floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or Memory Sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0101] The working process, working details and technical effects of the aforementioned computer-readable storage medium provided in this embodiment can be found in the method described in Embodiment 1 or any method that may involve the method described in Embodiment 1, and will not be repeated here.

[0102] Example 4: Based on the method provided in Example 1 and the system provided in Example 2, this example provides a computer device that executes the method described in Example 1 or any other method that may involve the method described in Example 1. The device includes a memory, a processor, and a transceiver connected in sequence. The memory stores a computer program, the transceiver sends and receives messages, and the processor reads the computer program and executes the method described in Example 1 or any other method that may involve the method described in Example 1. Specifically, the memory may include, but is not limited to, random-access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out (FIFO) memory, and / or last-in-first-out (FILO) memory, etc.; the processor may include, but is not limited to, a microprocessor of the STM32F105 series. Furthermore, the computer device may also include, but is not limited to, a power module, a display screen, and other necessary components.

[0103] Example 5: This example provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform the method described in Example 1 or any method that may involve the method described in Example 1. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0104] It should be understood that the terms "system," "device," "unit," and / or "module" as used in this specification are a method of distinguishing different components, elements, parts, sections, or assemblies at different levels. However, if other terms can achieve the same purpose, they may be replaced by other expressions.

[0105] As indicated in this specification and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of expressly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0106] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0107] It should be noted that the structures, proportions, sizes, etc., illustrated in the accompanying drawings are merely for illustrative purposes to aid those skilled in the art and are not intended to limit the scope of the invention. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in proportions, or adjustments to size, without affecting the effectiveness and purpose of the invention, should still fall within the scope of the disclosed technical content. Furthermore, terms such as "upper," "lower," "left," "right," and "middle" used in this specification are merely for clarity and not intended to limit the scope of the invention. Changes or adjustments to their relative relationships, without substantially altering the technical content, should also be considered within the scope of the invention.

Claims

1. A method for carbon intensity-constrained energy management and safe operation and maintenance of DC traction power supply, characterized in that, Includes the following steps: Establish a global monotonic clock on the EMS side; obtain the arrival time, release time, and applicable time period of the carbon intensity signal from an external data source; calculate the time offset based on the difference between the arrival time and the release time, and use the time offset to map the applicable time period to the global time axis of the EMS to obtain the aligned carbon intensity signal. Determine whether the applicable time period of the aligned carbon intensity signal covers the current control window; if so, obtain the coverage index. Analyze the update interval jitter and continuity of the carbon intensity signal, and generate an effective window gating state by combining the coverage index; If the effective window gating state allows participation in scheduling, the net power exchange increment of the power grid is predicted based on the candidate control actions of this cycle; it is then determined whether the absolute value of the net power exchange increment of the power grid exceeds the significance threshold obtained from the statistics of historical stable operation data. If so, the marginal emission factor is selected as the carbon constraint factor; otherwise, the average emission factor is selected as the carbon constraint factor. Based on the relative position of the selected carbon constraint factor in the historical time series distribution, and combined with the current adjustable resource constraints, candidate scheduling instructions containing energy storage charging and discharging power or converter voltage regulation are generated. Calculate the voltage safety margin of the DC bus voltage relative to the protection trip threshold, and obtain the effective safety margin after deducting measurement noise redundancy; The allowable voltage variation is determined based on the effective safety margin; Based on the allowable voltage variation and the equipment ramping capability, the candidate scheduling instructions are subjected to slope limiting and smoothing shaping to obtain the final execution instructions; A state machine is constructed based on the effective window gating state, voltage safety margin, and protection event count; When the state machine is in a frozen state, it exits the carbon constraint scheduling and executes instruction maintenance or voltage stabilization limiting actions; when the state machine is in a recovering state, it linearly releases the restriction on the amount of instruction change over time.

2. The carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method according to claim 1, characterized in that, The method for calculating the time offset is as follows: Maintain a sliding statistical window containing multiple recent carbon intensity signals, and record the difference between the arrival time and the release time of each carbon intensity signal; The median algorithm is used to calculate the statistical median of the difference as the initial offset, and the time offset is updated after outliers are removed based on the absolute median difference. If the external data source lacks the release time, the median of the arrival time intervals of the most recent valid carbon intensity signals is used as the update period, and an equivalent release time is constructed based on the midpoint of the update period and the applicable time period.

3. The carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method according to claim 1, characterized in that, The method for generating a valid window gating state is as follows: If the coverage indicator shows no coverage, or the continuity analysis shows that the out-of-order ratio exceeds the first threshold or the missing segment duration exceeds the second threshold, then the gating state is determined to be invalid. If the coverage indicator shows that the coverage and continuity meet the requirements, but the update interval jitter exceeds the normal range determined based on historical statistics, the gating state is determined to be a degraded state, and the amplitude of the subsequently generated candidate scheduling instructions is limited. If the coverage, continuity, and jitter indicators are all normal, the gating status is determined to be valid.

4. The carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method according to claim 1, characterized in that, The method for determining the significance threshold is as follows: The stable operating period of the screening system; the stable operating period simultaneously meets the following conditions: no protection action is triggered, the bus voltage fluctuation is less than the historical P90 percentile value, and the energy storage power fluctuation is less than the rated power setting ratio. The noise standard deviation σ of the net exchange power of the power grid during stable operation. P And set the significance threshold to 3·σ P .

5. The carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method according to claim 1, characterized in that, The method for predicting the increase in net power exchange capacity of the power grid is as follows: Based on the real-time power balance relationship of the power system, the algebraic sum of the changes in converter power command, energy storage power command, and regenerative feedback power estimation is calculated as the predicted value of the net power exchange increment of the power grid; among which, the regenerative feedback power estimation is calculated by moving average based on the measured regenerative power value in the most recent time window.

6. The carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method according to claim 1, characterized in that, The method for generating candidate scheduling instructions is as follows: Calculate the 10th percentile of the effective carbon factor over the most recent 24 hours. CI low and 90th percentile CI high ; The selected carbon factor is normalized to the [0,1] interval using a cutoff function to obtain the carbon intensity relative position coefficient s; Calculate the flexible power amplitude of the energy storage system P flex ; According to the formula P ess,cand =(0.5−s)×2× P flex Generate energy storage candidate power commands; Among them, when s A charging command is generated when the value is less than 0.

5. s A discharge command is generated when the value is greater than 0.

5.

7. The carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method according to claim 6, characterized in that, Flexible power amplitude of energy storage system P flex The calculation method is as follows: Obtain the real-time temperature of the energy storage battery T ess and state of charge SOC ; Determine the temperature correction factor based on the real-time temperature. f ( T ess ): The coefficient is 1 when the temperature is in the optimal operating range, 0.5 when the temperature is in the performance degradation range, and 0 when the temperature is in the safety protection range. Combination SOC Given the battery's rated power, calculate the maximum available charging power and the maximum available discharging power respectively, and take the minimum value of the maximum available charging power and the maximum available discharging power, multiply it by the temperature correction factor to obtain the flexible power amplitude.

8. The carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method according to claim 1, characterized in that, For scenarios without energy storage, the method for generating candidate scheduling instructions is as follows: Obtain the power-voltage sensitivity coefficient of the converter; the power-voltage sensitivity coefficient is used to characterize the converter voltage regulation amplitude corresponding to a unit amplitude of grid power fluctuation. Multiply the predicted net power exchange increment of the power grid by the power-voltage sensitivity coefficient to obtain the voltage fine-tuning amount; The voltage fine-tuning amount is superimposed on the system rated voltage reference value to generate the converter candidate voltage command.

9. A method for carbon intensity-constrained energy management and safe operation and maintenance of DC traction power supply according to claim 1, characterized in that, The methods for security assessment and instruction shaping are as follows: Calculate voltage safety margin M V =min( V trip,high - V dc , V dc - V trip,low ),in, V trip,high This is the overvoltage protection trip threshold. V trip,low This is the undervoltage protection trip threshold; Calculate noise redundancy U V =3·σ V , where σ V The standard deviation of voltage measurement noise; Calculate the allowable voltage change Δ V allow =min( R amp,V × T c , M V,eff ),in, R amp,V For the voltage ramping capability of the converter, T c To control the cycle, M V,eff For effective safety margin; The truncation function is used to limit the difference between the candidate scheduling instruction and the instruction of the previous cycle to -Δ. V allow With +Δ V allow Within the range, and perform first-order low-pass filtering smoothing on the limited instruction, Δ V allow This represents the allowable voltage variation.

10. A method for carbon intensity-constrained energy management and safe operation and maintenance of DC traction power supply according to claim 1, characterized in that, It also includes the following steps: The power safety margin is calculated by comprehensively considering the equipment's rated power, temperature, and state of charge. M P ; Determining power fluctuation uncertainty by combining current measurement noise U P ; Calculate the allowable power change Δ P allow =min( R amp,P × T c , M P - U P ); in, R amp,P This is the maximum permissible ramp rate of the equipment. T c To control the cycle duration, R amp,P × T c The maximum power variation limit based on ramp capability within a single control cycle; The allowable power variation is used to limit and filter the candidate power command for energy storage, and to verify whether the final command is within the available power range for energy storage charging and discharging.

11. The carbon intensity-constrained energy management and DC traction power supply safety operation and maintenance method according to claim 1, characterized in that, The state machine's transition logic and actions include: When the effective window gating state changes to the degraded state, the effective safety margin is lower than the first voltage threshold, or the protection event count exceeds the historical P80 percentile, switch to the degraded state and limit the command amplitude proportionally. When the effective window gating status becomes invalid, the effective safety margin is exhausted, or the protection event count exceeds the historical P95 percentile, switch to the frozen state; When the gating state is valid for multiple consecutive control cycles and the effective safety margin is restored, the system switches from the frozen state to the restored state and introduces a release linear instruction amplitude limit that increases over time.

Images