An industrial enterprise energy consumption intelligent management and control SaaS service management method and system

Abstract

The application relates to the technical field of network communication, and discloses an industrial enterprise energy consumption intelligent management and control SaaS service management method and system, which comprises the following steps: a gateway module receives discrete network data packets transmitted by a network element link and determines the current queuing delay of an input buffer; a timing control module calls the queuing delay and the transient induced voltage drop fluctuation rate of a transmission cable to calculate a load discrete state index; when the load discrete state index exceeds a fixed discrete threshold and the fluctuation rate is not over-standard, a distributed sensing network time slice matrix is reconstructed according to the network topology identification code of the network element node, and an asynchronous sending delay compensation time is dispatched; and a communication control module distributes timing configuration instructions. The application adjusts the data throughput state of the gateway module, avoids channel deadlock caused by queue congestion, and improves the transmission stability of the control data link.

Description

Technical Field

[0001] This invention relates to a SaaS service management method and system for intelligent energy consumption control in industrial enterprises, belonging to the field of network communication technology. Background Technology

[0002] Currently, industrial enterprise energy consumption data acquisition networks typically consist of a centralized gateway node aggregating polling messages reported by multiple edge sensor nodes. Daily communication of the data link is maintained through time-division multiplexing topology or carrier sense multiple access protocol, enabling the factory's intelligent management and control system to acquire time-series status data from the factory's power consumption side. Sensor nodes collect bus incoming parameters according to a fixed high-frequency sampling period, sending discrete energy consumption packets to the gateway's input buffer. The communication protocol stack then handles packet parsing and command forwarding. This is a common architecture in industrial field data interconnection. However, when high-load equipment in industrial enterprises is started... When a power supply bus experiences a transient fault, edge sensing nodes will generate a large number of reporting messages in a short period of time, quickly consuming the input buffer of the gateway node. When the gateway buffer receives such message streams with unstable arrival times and high concurrency, the internal queue is prone to queuing and blockage. Since existing communication control rules usually only process messages based on their arrival time and queue status, it is difficult to identify the corresponding power bus status changes behind the messages. Emergency control commands related to power supply bus anomalies that urgently need to be reported may be queued together with regular energy consumption statistics groups, which may lead to high transmission delays or even blockage of network control links.

[0003] To alleviate the aforementioned communication deadlock, simply increasing buffer capacity or widening channel bandwidth often further prolongs the waiting time for messages at the tail of the queue, causing some data packets to lose their practical control significance due to timeouts. If conventional indiscriminate packet dropping or fixed flow limiting strategies are adopted, critical fault messages may be directly discarded during sudden changes in energy consumption, causing the central control system to miss the optimal opportunity to intervene in excessive electricity demand. The disconnect between communication protocol scheduling rules and the characteristics of changes in the physical state of the bus is a structural problem that existing flow control mechanisms struggle to overcome under extreme bursts of traffic. It also causes a lag in the feedback of network scheduling methods to changes in operating conditions. For example, the publication number CN108965160A... A Chinese invention patent application discloses a dynamic differential weighted round-robin scheduling method applicable to multi-service transmission in mining IoT. This prior art relies on pure network layer message arrival speed and queuing delay to adjust bandwidth allocation. Its underlying premise is that traffic changes are only constrained by network statistical behavior. However, in industrial enterprises, sudden pulses of sensor data are often deeply coupled with physical power consumption anomalies such as the start-up and shutdown of heavy load equipment or transient bus faults. Because this scheduling mechanism fails to interlock network flow control with physical parameters such as voltage drop variation rate of the power supply bus across domains, it is difficult to distinguish between normal network congestion and power consumption anomalies under sudden energy consumption conditions, and it is impossible to open up a deterministic low-latency transmission channel for critical control commands.

[0004] Therefore, the technical problem to be solved by this invention is how to distinguish between regular congestion and power consumption anomalies when dense energy consumption data is reported in bursts, and how to reconstruct an asynchronous timing arrangement time slice matrix with feedback capability to ensure low-latency deterministic transmission of high-priority control command streams. Summary of the Invention

[0005] To address the problems in the background art, the technical solution of the present invention is as follows: A method for intelligent energy consumption management and control SaaS service in industrial enterprises, comprising the following steps: Step S101: The gateway module receives discrete network data packets transmitted via the network element link. The discrete network data packets carry high-frequency sensor data messages. Step S102: The congestion calculation module counts the total number of queued packets in the input buffer of the gateway module and determines the queuing delay in the buffer. Step S103: The timing control module reads the buffer queuing delay and the transient induced voltage drop variation rate of the transmission cable obtained by the power grid state parameter acquisition module. Step S104: The timing control module calculates the load dispersion index based on the deviation ratio between the buffer queuing delay and the fixed reference threshold. The load dispersion index is the ratio of the difference between the buffer queuing delay and the fixed reference threshold to the fixed reference threshold. Step S105: When the load discrete state index exceeds the fixed discrete threshold and the transient induced voltage drop variation rate is less than or equal to the fixed voltage distortion threshold, the timing control module reconstructs the distributed sensor network time slice matrix based on the load discrete state index and the network topology identifier of the network element node, and assigns asynchronous transmission delay compensation time to each network element node. In step S106, the communication control module encapsulates the reconstructed distributed sensor network time slice matrix into timing configuration instructions and distributes them to each network element node to adjust the data throughput status of the gateway module.

[0006] Preferably, the reconstruction of the distributed sensor network time slice matrix in step S105 includes the following sub-steps: Step S1051, using a first-order discrete difference trend extrapolation algorithm to calculate the queue length change rate of the gateway module input buffer in the past three control cycles, and predicting the expected traffic throughput trend in the next control cycle; Step S1052, combining the expected traffic throughput trend to calculate the timing peak compensation parameters, and reconstructing the peak transmission time slice allocation matrix in the distributed sensor network time slice matrix.

[0007] Preferably, after assigning asynchronous transmission delay compensation time to each network element node in step S105, the following steps are included: Step S301, calculating the cumulative standard deviation of each transmission delay compensation time within the sliding detection window; Step S302, if the cumulative standard deviation exceeds the fixed variable gate, issuing a benchmark adjustment instruction to the congestion calculation module to dynamically deduct the fixed benchmark threshold in the next control cycle from the fixed correction step size; if the cumulative standard deviation does not exceed the fixed variable gate, maintaining the fixed benchmark threshold at a steady-state initial value.

[0008] Preferably, the method further includes the following steps: Step S401, deploying a discrete bypass preemption control channel driven directly by a hardware timer at the physical network adapter layer of the gateway module to bypass the operating system interrupt queue; Step S402, in the case of hardware input buffer overflow pin triggering, directly broadcasting a fixed-duration pause transmission frame to the network element link through the discrete bypass preemption control channel to suspend the regular status statistics messages of low-priority network element nodes.

[0009] Preferably, in step S402, a fixed-duration pause frame is broadcast to open a low-latency transmission window for the downlink distribution of timing configuration instructions and to prevent deadlock of cached data in the gateway module protocol stack under overflow conditions.

[0010] Preferably, the method further includes the following steps: Step S601, in the gateway module, the discrete network data packets are decoupled into high-priority congestion control frames and regular non-instantaneous status statistics frames; Step S602, when the gateway module generates a mild queuing delay state, the software-defined topology is started, and the regular non-instantaneous status statistics frames are redirected to a low-rate bypass communication link to isolate the transmission of high-priority congestion control frames.

[0011] Preferably, the rule for determining that the gateway module has a slight queuing delay state in step S602 is: when the load discrete state index is greater than 0 and less than or equal to the fixed discrete threshold, the software-defined topology is triggered to start.

[0012] Preferably, before reading the transient induced voltage drop variation rate of the transmission cable in step S103, the method further includes: using the power grid state parameter acquisition module to collect the voltage waveform of the transmission cable at a fixed sampling frequency of 50Hz, calculating the absolute value of the slope of the transient voltage drop through waveform distortion feature analysis, and determining the transient induced voltage drop variation rate of the transmission cable.

[0013] Preferably, after distributing the timing configuration instruction in step S106, the method further includes: sending the timing configuration instruction to the distributed energy consumption monitoring node, restricting the distributed energy consumption monitoring node to report high-frequency sensor data packets within a specified time slice, and smoothing out the transient induced voltage drop fluctuation rate of the transmission cable.

[0014] An intelligent energy consumption management and control SaaS service management system for industrial enterprises, which is used to implement an intelligent energy consumption management and control SaaS service management method for industrial enterprises, including: The gateway module is used to receive discrete network data packets transmitted through network element links. These discrete network data packets carry high-frequency sensor data messages. The congestion calculation module is used to count the total number of queued packets in the input buffer of the gateway module and determine the queuing delay in the buffer. The timing control module is used to read the buffer queuing delay and the transient induced voltage drop variation rate of the transmission cable obtained by the power grid state parameter acquisition module, and calculate the load discrete state index based on the deviation ratio of the buffer queuing delay and the fixed reference threshold. The load discrete state index is the ratio of the difference between the buffer queuing delay and the fixed reference threshold to the fixed reference threshold. When the load discrete state index exceeds the fixed discrete threshold and the transient induced voltage drop variation rate is less than or equal to the fixed voltage distortion threshold, the timing control module reconstructs the distributed sensor network time slice matrix based on the load discrete state index and the network topology identifier of the network element node, and assigns asynchronous transmission delay compensation time to each network element node. The communication control module is used to encapsulate the reconstructed distributed sensor network time slice matrix into timing configuration instructions and distribute them to each network element node, and adjust the data throughput status of the gateway module.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. In the intelligent energy consumption management SaaS service of industrial enterprises, the gateway module determines the queuing delay based on the number of queued messages in the input buffer and makes a judgment based on the transient voltage drop of the incoming bus. This enables the system to distinguish between congestion caused by ordinary network overload and sudden reporting caused by power consumption abnormalities. Thus, when the pressure in the buffer increases but the voltage abnormality does not exceed the set range, it can promptly trigger off-peak reporting, slowing down the concentrated entry speed of regular energy consumption data and gradually reducing the gateway buffer queue to below the safe level. This not only alleviates the congestion caused by concentrated message reporting but also avoids the loss of critical energy consumption data due to simple rate limiting or packet loss.

[0016] 2. By observing the fluctuations in the transmission delay compensation time of each monitoring node over a period of time, the timing control module can more promptly identify unstable characteristics of the centralized arrival of sensor data. When the fluctuation exceeds the set range, the system lowers the fixed reference threshold for the next control cycle, enabling congestion judgment to intervene earlier and forming feedback adjustment oriented towards the gateway buffer state. Compared with the flow control method that relies solely on a fixed reference, this method reduces the risk of scheduling response lag, enabling the industrial site to maintain a relatively stable control command transmission state even under non-ideal conditions such as strong electromagnetic interference and concurrent reporting by multiple nodes.

[0017] 3. The gateway module sets up a discrete bypass preemption control channel directly driven by a hardware timer at the physical network adapter layer. When the hardware input buffer overflows, it can bypass the operating system interrupt queue and directly broadcast a fixed-duration pause transmission frame to the network element link. After the non-critical sensor data reporting of low-priority nodes is temporarily suspended, the timing configuration command obtains a lower latency downlink transmission window, thereby reducing the risk of the gateway protocol stack software queue continuing to back up and form a deadlock in the overflow state. This fast yielding mechanism on the hardware side, together with the asynchronous peak-shifting scheduling of the protocol layer, helps to ensure the stability of the control data link and the smooth flow of the field control channel under extreme operating conditions. Attached Figure Description

[0018] Fig. 1 This is a flowchart of the gateway data throughput timing configuration method of the present invention; Fig. 2 This is a system architecture diagram of the gateway data throughput timing configuration system of the present invention.

[0019] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0021] A method for managing intelligent energy consumption control SaaS services in industrial enterprises includes the following steps: Step S101: The gateway module receives discrete network data packets transmitted via the network element link, wherein the discrete network data packets carry high-frequency sensor data messages. Step S102: The congestion calculation module counts the total number of queued packets in the input buffer of the gateway module and determines the buffer queuing delay. Step S103: The timing control module reads the buffer queuing delay and the transient induced voltage drop variation rate of the transmission cable obtained by the power grid state parameter acquisition module. Step S104: The timing control module calculates the load dispersion state index based on the deviation ratio between the buffer queuing delay and the fixed reference threshold. The load dispersion state index is the ratio of the difference between the buffer queuing delay and the fixed reference threshold to the fixed reference threshold. Step S105: When the load discrete state index exceeds the fixed discrete threshold and the transient induced voltage drop variation rate is less than or equal to the fixed voltage distortion threshold, the timing control module reconstructs the distributed sensor network time slice matrix based on the load discrete state index and the network topology identifier of the network element node, and assigns asynchronous transmission delay compensation time to each of the network element nodes. In step S106, the communication control module encapsulates the reconstructed distributed sensor network time slice matrix into timing configuration instructions and distributes them to each of the network element nodes to adjust the data throughput status of the gateway module.

[0022] Preferably, the reconstruction of the distributed sensor network time slice matrix in step S105 includes the following sub-steps: Step S1051, using a first-order discrete difference trend extrapolation algorithm to calculate the queue length change rate of the gateway module input buffer in the past three control cycles, and predicting the expected traffic throughput trend in the next control cycle; Step S1052, combining the expected traffic throughput trend to calculate the timing peak compensation parameters, and reconstructing the peak transmission time slice allocation matrix in the distributed sensor network time slice matrix.

[0023] Preferably, after assigning asynchronous transmission delay compensation time to each of the network element nodes in step S105, the following steps are included: Step S301, calculating the cumulative standard deviation of each of the transmission delay compensation times within the sliding detection window; Step S302, if the cumulative standard deviation exceeds the fixed variable gate, issuing a benchmark adjustment instruction to the congestion calculation module to dynamically deduct a fixed correction step size from the fixed benchmark threshold in the next control cycle; if the cumulative standard deviation does not exceed the fixed variable gate, maintaining the fixed benchmark threshold at a steady-state initial value.

[0024] Preferably, the method further includes the following steps: Step S401, deploying a discrete bypass preemption control channel driven directly by a hardware timer at the physical network adapter layer of the gateway module to bypass the operating system interrupt queue; Step S402, in the case of hardware input buffer overflow pin triggering, directly broadcasting a fixed-duration pause transmission frame to the network element link through the discrete bypass preemption control channel to suspend the regular status statistics messages of low-priority network element nodes.

[0025] Preferably, in step S402, a fixed-duration pause frame is broadcast to open a low-latency transmission window for the downlink distribution of the timing configuration instruction, thereby preventing the deadlock of the cached data in the overflow state of the gateway module protocol stack.

[0026] Preferably, the method further includes the following steps: Step S601, in the gateway module, the discrete network data packets are decoupled into high-priority congestion control frames and regular non-instantaneous status statistics frames; Step S602, when the gateway module generates a mild queuing delay state, a software-defined topology is started to redirect the regular non-instantaneous status statistics frames to a low-rate bypass communication link to isolate the transmission of the high-priority congestion control frames.

[0027] Preferably, the rule for determining that the gateway module has a slight queuing delay state in step S602 is: when the load discrete state index is greater than 0 and less than or equal to the fixed discrete threshold, the software-defined topology is triggered to start.

[0028] Preferably, before reading the transient induced voltage drop variation rate of the transmission cable in step S103, the method further includes: using the power grid state parameter acquisition module to collect the voltage waveform of the transmission cable at a fixed sampling frequency of 50Hz, calculating the absolute value of the slope of the transient voltage drop through waveform distortion feature analysis, and determining the transient induced voltage drop variation rate of the transmission cable.

[0029] Preferably, after distributing the timing configuration instruction in step S106, the method further includes: sending the timing configuration instruction to the distributed energy consumption monitoring node, restricting the distributed energy consumption monitoring node to report the high-frequency sensor data message within a specified time slice, and smoothing out the transient induced voltage drop fluctuation rate of the transmission cable.

[0030] An intelligent energy consumption management and control SaaS service management system for industrial enterprises, which is used to implement an intelligent energy consumption management and control SaaS service management method for industrial enterprises, including: The gateway module is used to receive discrete network data packets transmitted through network element links. These discrete network data packets carry high-frequency sensor data messages. The congestion calculation module is used to count the total number of queued packets in the input buffer of the gateway module and determine the queuing delay in the buffer. The timing control module is used to read the buffer queuing delay and the transient induced voltage drop variation rate of the transmission cable obtained by the power grid state parameter acquisition module, and calculate the load discrete state index based on the deviation ratio of the buffer queuing delay and the fixed reference threshold. The load discrete state index is the ratio of the difference between the buffer queuing delay and the fixed reference threshold to the fixed reference threshold. When the load discrete state index exceeds the fixed discrete threshold and the transient induced voltage drop variation rate is less than or equal to the fixed voltage distortion threshold, the timing control module reconstructs the distributed sensor network time slice matrix based on the load discrete state index and the network topology identifier of the network element node, and assigns asynchronous transmission delay compensation time to each network element node. The communication control module is used to encapsulate the reconstructed distributed sensor network time slice matrix into timing configuration instructions and distribute them to each network element node, and adjust the data throughput status of the gateway module.

[0031] Example 1: When the intelligent energy consumption management system of an industrial enterprise is running, the industrial internet platform system receives discrete network data packets transmitted by the network element link through the gateway module. The discrete network data packets are formed by the convergence of multiple edge sensing nodes and carry high-frequency sensing data messages. In large-scale high-energy-consuming industrial sites, the start-up and shutdown of heavy-load equipment is easily accompanied by strong electromagnetic interference, and the input buffer of the gateway module will be impacted by non-stationary concurrent data streams. At this time, the congestion calculation module counts the current total number of queued messages in the input buffer in real time, and converts the total number of queued messages into the buffer queuing delay according to the processing cycle of a single message in the gateway module or the calibrated queue service time.

[0032] The timing control module retrieves the buffer queuing delay and the transient induced voltage drop variation rate of the transmission cable obtained by the power grid state parameter acquisition module. Based on the deviation ratio of the buffer queuing delay from a fixed reference threshold, it calculates the load discrete state index using the formula... Calculate, where, As an indicator of load discrete state, This refers to the queuing delay in the buffer zone. With a fixed baseline threshold, when the load discrete state index exceeds the fixed discrete threshold and the transient induced voltage drop variation rate is less than or equal to the fixed voltage distortion threshold, the timing control module determines that the current queue accumulation is a communication congestion caused by conventional network overload, triggering asynchronous peak-shaving control. The timing control module reconstructs the distributed sensor network time slice matrix based on the load discrete state index and the network topology identifier code of the network element node, and assigns asynchronous transmission delay compensation time to each network element node, so that the reporting windows of each network element node are staggered on the time axis, thereby reducing the number of packets entering the gateway module input buffer in the same period.

[0033] When reconstructing the time-slice matrix of the distributed sensor network, the timing control module first retrieves the historical queue length data of the gateway module input buffer in the current control cycle, the previous control cycle, and the two previous control cycles. Then, it calculates the first difference in queue length between the current control cycle and the previous control cycle, and the second difference in queue length between the previous control cycle and the two previous control cycles. The first difference is then subtracted from the second difference to obtain the current queue length change rate, and this change rate is used as the slope to predict the expected traffic throughput trend in the next control cycle. If the expected traffic throughput trend shows an increasing trend, the timing control module calculates a timing peak-shifting compensation parameter that is proportional to the expected traffic throughput trend, and determines the row and column position or time slice number of the corresponding node in the matrix by combining the network topology identifier code of each network element node. The timing peak-shifting compensation parameter is then superimposed on the original time slice position of the corresponding node to form a peak-shifting transmission time slice allocation matrix.

[0034] To track the transient cluster pulse characteristics of sensor network data, the timing control module initiates feedback adjustment after allocating asynchronous transmission delay compensation times. It calculates the cumulative standard deviation of each transmission delay compensation time within the sliding detection window. When the cumulative standard deviation exceeds a fixed variable threshold, the timing control module sends a baseline adjustment command to the congestion calculation module, adjusting the fixed baseline threshold for the next control cycle. Dynamically reduce the fixed correction step size to improve the detection sensitivity of the pre-congestion judgment; when the cumulative standard deviation does not exceed the fixed variable gate, the fixed benchmark threshold is applied. Maintain the initial steady-state value.

[0035] The communication control module encapsulates the reconstructed distributed sensor network time slice matrix into timing configuration instructions and distributes them to each network element node. Each distributed energy consumption monitoring node responds to the timing configuration instructions and reports high-frequency sensor data packets according to the specified off-peak time slices, thereby adjusting the data throughput state of the gateway module and causing the communication packets backed up in the input buffer to converge below the water level, thus preventing the gateway module protocol stack software queue from overflowing and deadlocking.

[0036] In practice, the network adapter chip of the gateway module connects the overflow pin of the internal hardware input buffer to the trigger terminal of the dedicated timer control register via hardwire. When the total number of queued messages in the input buffer reaches the chip's hardware overflow threshold, the hardware input buffer overflow pin outputs a high-level signal. Without going through the operating system interrupt handler and protocol stack queuing, the hardware timer directly drives the physical layer media access controller. Through a discrete bypass preemption control channel independent of the regular data channel, a pause transmission frame with a duration of 50μs is broadcast to the network element link, suspending the regular status statistics messages of low-priority network element nodes and opening a low-latency transmission window for timing configuration instructions. Through the above timing peak-shifting control and discrete bypass preemption control channel, under the impact of traffic loads crossing fixed discrete thresholds, the transmission delay of fault-critical data frames under channel congestion is stabilized within 12ms, avoiding the problem of energy consumption data omission caused by passive discarding in conventional flow control, and outputting energy consumption management decision flow data to the terminal execution mechanism.

[0037] Specifically, the energy consumption management decision flow data is output to the power capacitor bank and dynamic voltage restorer on the substation bus side via a low-latency control link. After receiving the decision flow data, the terminal actuators perform reactive power compensation and active voltage injection. Since the main communication link avoids data deadlock under peak-shaving control, the transmission delay of critical fault data frames is controlled within 12ms. The terminal actuators can complete protective action response in the initial stage of transient faults caused by the start-up and shutdown of heavy load equipment. By adjusting the reactive power flow, the local distortion of the power grid is suppressed from spreading to the transmission cables, thereby smoothing the transient induced voltage drop rate of the transmission cables.

[0038] Example 2: Performance verification testing of the current industrial enterprise's intelligent energy consumption management system was conducted on a semi-physical simulation test bench with multiple Ethernet interfaces. The test bench received high-frequency digital sampling signals output from the field power distribution monitoring bus, with a data stream sampling rate maintained at 10kHz, a voltage measurement resolution of 0.01V, and a voltage measurement accuracy of ±0.1%. This was used to reproduce the transmission status of discrete network data packets entering the gateway module under heavy load and complex topology conditions, in order to determine the sliding detection window parameters during the test. The system uses the arrival frequency of sensor data under high network concurrency and the dynamic allocation overhead of the gateway module's input buffer as the basis for its values, and makes a trade-off between data real-time performance and processor computational scheduling load. When the packet concurrency intensity continues to increase with the expansion of the monitoring node scale, in order to enable the timing control module to track the discrete changes in asynchronous transmission delay compensation time, the sliding detection window parameters are adjusted. The value range should move closer to the lower limit of the range that satisfies the anti-statistical aliasing constraint; in a typical industrial network element topology scenario containing 128 distributed monitoring terminals, the sliding detection window parameters... Take 50 sampling periods.

[0039] The testing process uses the complete scheme of this invention as the sample group, and three control sample groups are set up. The fixed benchmark threshold is removed in the first control sample group. The feedback adjustment loop was compared with the branch in which the transient induced voltage drop variation rate of the transmission cable was removed from the interlock judgment in sample group two, and the fixed discrete threshold setting was changed in sample group three. Under different congestion load intensities, the average transmission delay, message drop rate and buffer queue fluctuation variance of high frequency sensor data packets were recorded for each sample group.

[0040] In the sample of this invention, the fixed discrete threshold was set at three support points of 1.5, 2.5, and 3.5 depending on the congestion load intensity. Under low load gradient, with a fixed discrete threshold of 1.5, the average transmission delay was 6.2 ms, the packet drop rate was 0.00%, and the buffer queue fluctuation variance was 0.12. Under normal load gradient, with a fixed discrete threshold of 2.5, the average transmission delay was 8.4 ms, the packet drop rate was 0.00%, and the buffer queue fluctuation variance was 0.15. Under high load gradient, with a fixed discrete threshold of 3.5, the average transmission delay was 11.8 ms, the packet drop rate was 0.00%, and the buffer queue fluctuation variance was 0.19. The above data show that at the three test points of low load, normal load, and high load, the buffer queuing delay is... All were kept below the preset safety level, and no high-frequency sensor data packets were dropped.

[0041] In the first comparative sample, the fixed discrete threshold was set to 2.5. Under low load gradient, after canceling the feedback adjustment loop, the average transmission delay was 9.3ms, the message drop rate was 0.00%, and the buffer queue fluctuation variance was 0.28. Under high load gradient, the average transmission delay increased to 24.6ms, the message drop rate increased to 1.15%, and the buffer queue fluctuation variance increased to 0.54. Compared with the test data of the sample of this invention in adjacent load intervals, the cumulative standard deviation of the transmission delay compensation time failed to continue to affect the fixed baseline threshold. The dynamic deduction results in a lag in the front-end congestion determination under high load gradients, and the packet accumulation fluctuation in the input buffer increases.

[0042] In the second comparative sample, the fixed discrete threshold was also set to 2.5. Under normal load gradient, after removing the branch in which the transient induced voltage drop variation rate of the transmission cable was used for interlock judgment, the average transmission delay was 15.2ms, the message drop rate was 0.42%, and the buffer queue fluctuation variance was 0.41. Under high load gradient, the average transmission delay further increased to 38.7ms, the message drop rate increased to 4.12%, and the buffer queue fluctuation variance increased to 0.73. The phenomenon corresponding to this set of data is that the system only schedules based on the queue status on the communication side, making it difficult to distinguish between communication congestion caused by normal network overload and sensor data surges caused by sudden power failures. After different types of messages are mixed in the queue, the transmission delay and packet loss increase simultaneously.

[0043] In the third comparative sample, the fixed discrete threshold was set outside the working window. Under low load gradient, the fixed discrete threshold was 1.0, the average transmission delay was 18.5ms, the packet drop rate was 0.00%, and the buffer queue fluctuation variance was 0.22. Due to the overly sensitive judgment condition, the system triggered off-peak reporting under low load, and the energy consumption data refresh delay increased accordingly. Under high load gradient, the fixed discrete threshold was 4.5, the average transmission delay was 46.3ms, the packet drop rate was 8.65%, and the buffer queue fluctuation variance was 0.91. Due to the lag in congestion judgment, the gateway module input buffer failed to reduce the high-frequency sensor data packet surge in time, eventually resulting in overflow and large-scale packet loss.

[0044] The correspondence between the data sets shows that when the fixed discrete thresholds are 1.5, 2.5, and 3.5, the average transmission delays of the sample group are 6.2ms, 8.4ms, and 11.8ms, respectively, all below 12ms, and the packet drop rate is 0.00%. In contrast, sample group 1 has an average transmission delay of 24.6ms and a packet drop rate of 1.15% under high load gradient, indicating that the fixed baseline threshold... The feedback regulation loop supports queue convergence under high load. Compared with sample group 2, the average transmission delay under normal load gradient has reached 15.2ms, and further reaches 38.7ms under high load gradient, indicating that after the transient induced voltage drop variation rate of the transmission cable participates in the interlock judgment, it can reduce the indiscriminate mixing of conventional network overload messages and sudden power consumption fault messages. Compared with sample group 3, when the fixed discrete threshold is 1.0 and 4.5, the refresh delay increases and the packet loss rate increases, respectively, indicating that the fixed discrete threshold needs to fall within the working window that matches the load discrete state index.

[0045] In the average transmission latency test, the semi-physical simulation test architecture included 128 distributed monitoring terminals. The network protocol analyzer deployed high-precision timestamp hooks at the network adapter sending end of the network element node and the kernel protocol stack receiving end of the gateway module. When a critical fault data frame was sent, the hardware timer recorded the first absolute timestamp. After the data frame passed through the physical layer and was successfully parsed by the gateway module protocol stack, the second absolute timestamp was recorded. After continuously collecting 5000 critical fault data frames, the difference between the second absolute timestamp and the first absolute timestamp was used as the single frame transmission latency, and the arithmetic mean of the 5000 differences was calculated. The test results showed that the average transmission latency of the sample group of this invention was controlled within 12ms under various congestion load gradients, and the energy consumption management decision flow data could maintain a reliable output state to the terminal execution mechanism.

[0046] Example 3: During the operation of the current intelligent energy consumption management system in industrial enterprises, the industrial sites where distributed energy consumption monitoring nodes are located experience heavy equipment start-ups and shutdowns, as well as strong electromagnetic interference with high harmonics. The input buffer of the gateway module is subjected to the impact of non-stationary, randomly distributed concurrent data streams under high network concurrency. To eliminate communication scheduling baseline inaccuracies caused by long-term operation or signal drift of network components, the timing control module within the gateway module retrieves the current total number of queued messages in the input buffer and calculates the buffer queuing delay. and combined with a fixed benchmark threshold And the transient induced voltage drop rate of the transmission cable obtained by the power grid status parameter acquisition module, according to the formula Calculate the load discrete state index ,in, As an indicator of load discrete state, This refers to the queuing delay in the buffer zone. A fixed baseline threshold.

[0047] When the load is discrete state index When the transient induced voltage drop rate exceeds a fixed discrete threshold and is less than or equal to a fixed voltage distortion threshold, the timing control module reconstructs the distributed sensor network time slice matrix according to the network topology identifier of the network element node, allocates asynchronous transmission delay compensation time, and dynamically adjusts the reporting timing of each distributed energy consumption monitoring node within the sliding detection window. The duration of the sliding detection window is determined by the sliding detection window parameters. Confirm, sliding detection window parameters Adjustments are made based on changes in network data concurrency intensity and dynamic allocation overhead of the input buffer; in a typical network topology environment containing 128 monitoring nodes, the sliding detection window parameters are... The sampling period is set to 50.

[0048] The timing control module calculates the cumulative standard deviation of each transmission delay compensation time within the sliding detection window. When the cumulative standard deviation exceeds the fixed variable threshold, it sends a baseline adjustment command to the congestion calculation module, setting the fixed baseline threshold for the next control cycle. Dynamically reduce the fixed correction step size to improve the sensitivity of the preceding congestion judgment; when the cumulative standard deviation does not exceed the fixed variable threshold, the fixed benchmark threshold is applied. Maintain the initial steady-state value.

[0049] Furthermore, when the calculated load discrete state index is greater than 0 and less than or equal to a fixed discrete threshold, the gateway module is in a state of mild queuing delay. The core control program within the gateway module initiates a software-defined topology. This mechanism reads the flow table control policy and performs deep packet inspection on the received discrete network data packets at the application layer, decoupling high-priority congestion control frames carrying key alarm control information and regular non-real-time status statistics frames. Subsequently, the virtual switching architecture of the gateway module issues a redirection control command, changing the forwarding port of the regular non-real-time status statistics frames to a low-rate bypass communication link. The main Ethernet physical communication link isolates the transmission of high-priority congestion control frames, completing dual-channel flow splitting from both the logical path and the physical link.

[0050] The communication control module encapsulates the reconstructed distributed sensor network time-slice matrix into timing configuration instructions and distributes them to each network element node. This restricts the reporting of high-frequency sensor data packets within the specified off-peak time slices, reduces the queuing delay of the gateway module's input buffer, and prevents deadlock due to overflow in the gateway module's protocol stack software queue. After the above processing, the transmission delay of critical fault data frames under channel congestion conditions is stabilized within 12ms, energy consumption data omissions are avoided, and energy consumption management decision flow data is output to the terminal execution mechanism.

[0051] Example 4: This example combines Figs. 1-2 This document describes the management methods and systems for intelligent energy consumption control SaaS services in industrial enterprises, such as... Fig. 1As shown, in step S101, the gateway module receives discrete network data packets transmitted from the network element link. These discrete network data packets carry high-frequency sensor data messages and are transmitted to step S102. The congestion calculation module then counts the total number of queued messages in the gateway module's input buffer to determine the buffer queuing delay. Subsequently, in step S103, the timing control module reads the buffer queuing delay and the transient induced voltage drop rate of the transmission cable obtained by the power grid state parameter acquisition module, and proceeds to step S104. The timing control module calculates the load discrete state index based on the deviation ratio of the buffer queuing delay from a fixed reference threshold. The load discrete state index is the buffer... The ratio of the difference between the queuing delay and the fixed reference threshold to the fixed reference threshold is further defined in step S105. When the load discrete state index exceeds the fixed discrete threshold and the transient induced voltage drop variation rate is less than or equal to the fixed voltage distortion threshold, the timing control module reconstructs the distributed sensor network time slice matrix based on the load discrete state index and the network topology identifier of the network element node, and assigns asynchronous transmission delay compensation time to each network element node. Finally, in step S106, the communication control module encapsulates the reconstructed distributed sensor network time slice matrix into a timing configuration instruction and distributes it to each network element node to complete the data throughput state adjustment of the gateway module.

[0052] like Fig. 2As shown, the gateway module receives discrete network data packets from the network element link and parses out high-frequency sensor data packets. Simultaneously, the congestion calculation module determines the buffer queuing delay based on the total number of queued packets. During this period, the power grid state parameter acquisition module performs voltage waveform distortion characteristic analysis on the transmission cable at a fixed sampling frequency of 50Hz to obtain its transient induced voltage drop variation rate. Then, the timing control state calculation unit combines a fixed reference threshold and the buffer queuing delay to determine the load discrete state index. When the timing time slice matrix reconstruction unit is triggered, it reconstructs and allocates the distributed sensor network time slice matrix using a first-order discrete differential trend extrapolation algorithm. Furthermore, the timing control feedback adjustment architecture monitors the cumulative standard deviation of the allocated asynchronous transmission delay compensation time and the fixed variable gate in real time within a specified sliding detection window. When facing extremely high concurrent traffic, the hardware... The path preemption control mechanism directly triggers the discrete bypass preemption control channel by using the overflow pin output signal of the hardware input buffer. Combined with the software-defined topology diversion mechanism, the received discrete network data packets are decoupled and separated into high-priority congestion control frames and regular non-real-time status statistics frames on the logical and physical paths to achieve dual-channel diversion. Subsequently, the communication control command distribution unit sends the generated timing configuration command downlink to the distributed energy consumption monitoring node to limit its reporting rhythm, so that the terminal actuator on the substation bus side responds to the control and uses the power capacitor bank and restorer to adjust the reactive power flow and inject active power. The modules and technical elements at all levels achieve system-wide collaborative interaction through the underlying signal input and output logic, physical connection relationship and process step flow sequence, so as to ultimately achieve the core goal of avoiding gateway buffer blockage and data deadlock and ensuring low-latency deterministic transmission of control links.

[0053] Example 5: Before the current industrial enterprise's intelligent energy consumption management and control system is deployed on-site, the gateway module is in an unloaded commissioning state. The timing control module within the gateway module retrieves the pre-baseline calibration algorithm to calculate the fixed reference threshold. The steady-state initial value is used as the input of the pre-baseline calibration algorithm, which takes 600 cycles of background communication residual data continuously collected by the gateway module under no-load conditions. The embedded microprocessor inside the gateway module calculates the moving average of the background communication residual data to filter out random interference caused by stray electromagnetic noise in the field. Under the condition of satisfying the anti-clock drift constraint, the timing control module determines the product of the moving average and the preset redundancy safety factor as the fixed reference threshold. The zero-load baseline quantity.

[0054] In practice, the sliding average window width of the pre-baseline calibration algorithm is set to 50 control cycles. The timing control module performs a segmented sliding average on the background communication residual data for 600 consecutive cycles, removes high-frequency spike pulses that deviate from the normal range, and uses the smoothed background delay as the reference. The phase shift timing distributor adopts a dedicated hardware shift register bus architecture. After the pre-baseline calibration algorithm outputs a zero-load reference value, the phase shift timing distributor, driven by the hardware clock edge, writes the reference value in binary form to the target configuration register inside the congestion calculation module through the parallel data bus to complete the adaptive initial calibration.

[0055] The phase-shift timing distributor uses the calculated zero-load baseline value as the initial parameter value, writes it into the configuration register address space of the congestion calculation module, and uses it as the subsequent calculation of load discrete state indicators. The gateway module uses the zero-load baseline quantity during the initial operation phase to suppress the false triggering of pre-congestion judgment caused by the inherent communication background overhead on site, so that the distributed monitoring network topology can maintain the adaptability of communication link operation under the impact of traffic burst load, and limit the transmission delay of fault key data frames carrying energy consumption messages to within 12ms under channel congestion conditions, and output energy consumption control decision flow data to the terminal execution mechanism.

[0056] Example 6: To determine the incremental parameters for peak-shaving control under strong electromagnetic interference conditions, the timing control module adopts an offline optimization test method. Using the high-frequency sensor data packet drop rate under high-concurrency conflict conditions as the objective function, a control simulation loop is established on a semi-physical simulation test bench with multiple Ethernet interfaces. During the test, the communication attenuation operator is determined by adjusting the physical channel bit error rate variation sequence between the gateway module and each network element node read from the network protocol stack. Compensation time for the original transmission delay The incremental correction pattern and the enabling environment for offline optimization testing include a processor with a main frequency of 1.2 GHz. This processor is used to collect the discrete distribution uncertainty of the physical channel bit error rate variation sequence within a 10-second sliding time window in real time.

[0057] In the first sampling gradient, when the gateway module detects that the bit error rate of the network element node's raw physical channel changes from 10s to 10s, the network element node's bit error rate changes from 10s to ... Mutation to At that time, the processor calculates the normalized logarithmic integral of the bit error rate variation sequence to obtain the communication attenuation operator. The current quantization value is 1.24, which is used as the input variable for offline parameter finding and passed to the phase shift timing distributor.

[0058] The phase-shift timing distributor introduces a discrete step damping rule to determine the additional time delay compensation operator. It is used to compensate for the demodulation delay lag and high-order intermodulation distortion adjustment deviation caused by external strong harmonic electromagnetic radiation interference to network element nodes. When the communication attenuation operator When less than or equal to 1.0, an additional delay compensation operator is applied. Set to 0ms; when the communication attenuation operator When the value is greater than 1.0 and less than or equal to 2.0, according to the formula... Calculate the additional delay compensation operator ,in, To add a delay compensation operator, As the baseline delay increment, For communication attenuation operator, This is the attenuation correction factor. and All are dimensionless scalars.

[0059] In the reference delay increment Under the condition of curing time of 2.0ms, the attenuation correction coefficient is changed. The discrete gradient, when the attenuation correction coefficient When set to 0.35, substitute it into the communication attenuation operator. The quantized value is 1.24, and the additional delay compensation operator is obtained by recalculation. The normalized logarithmic integral is calculated as follows: The base-10 logarithmic value of the original physical channel bit error rate collected within a 10-second sliding time window is logarithmically calculated. This logarithmic value is then mapped to a preset dimensionless scalar interval for definite integration, outputting a communication attenuation operator that reflects the degree of channel quality degradation. Attenuation correction coefficient The value of 0.35 was determined after conducting multiple extreme overload experiments on a semi-physical simulation test bench. This value represents the attenuation correction coefficient. When the value is below 0.10, the calculated additional delay compensation is too small to offset the demodulation delay lag, and the peak transmission windows will still overlap; when the attenuation correction coefficient... When the value is higher than 0.60, the additional delay compensation is too large, the transmission window of the network element node is excessively delayed, and cascading transmission congestion is triggered.

[0060] When the communication attenuation operator When the value is greater than 2.0, the discrete step damping rule enables the upper limit blocking, and an additional time delay compensation operator is added. Limiting to a single scalar cutoff boundary of 1.5ms, and adding this additional delay compensation operator. Added to the original transmission delay compensation time In this context, it is used to update the time slice matrix of a distributed sensor network, adding a delay compensation operator. The truncation is set to 1.5ms to match the maximum allowable dynamic delay limit of the gateway module's input buffer; in the communication attenuation operator In extreme and severe operating conditions with a latency greater than 2.0, if the compensation latency continues to increase linearly, it will exceed the frame lifecycle window of the data link, causing the retransmission queue to accumulate rapidly and the control command to fail. By using a 1.5ms upper limit lockout, and in conjunction with the physical isolation measure of triggering the suspension of frame transmission by the hardware input buffer overflow pin, the overall queuing latency is limited to the safe waterline of the gateway module protocol stack, avoiding scheduling loss of control caused by large-scale timing jumps.

[0061] Each distributed energy consumption monitoring node responds to the timing configuration command issued based on the updated distributed sensor network time slice matrix, adjusts the data throughput status of the gateway module, and restricts the reporting of high-frequency sensor data packets within the specified off-peak time slice. This causes the communication packets accumulated in the gateway module's input buffer to converge below the preset water level. Under heavy load fluctuation conditions, the average transmission delay of key energy consumption control commands under retransmission congestion is stabilized within 12ms, data loss anomalies are avoided, and energy consumption management decision flow data is output to the terminal execution mechanism.

[0062] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0063] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for intelligent energy consumption control SaaS service management in industrial enterprises, characterized in that, Includes the following steps: Step S101: The gateway module receives discrete network data packets transmitted via the network element link. The discrete network data packets carry high-frequency sensor data messages. Step S102: The congestion calculation module counts the total number of queued packets in the input buffer of the gateway module and determines the queuing delay in the buffer. Step S103: The timing control module reads the buffer queuing delay and the transient induced voltage drop variation rate of the transmission cable obtained by the power grid state parameter acquisition module. Step S104: The timing control module calculates the load dispersion index based on the deviation ratio between the buffer queuing delay and the fixed reference threshold. The load dispersion index is the ratio of the difference between the buffer queuing delay and the fixed reference threshold to the fixed reference threshold. Step S105: When the load discrete state index exceeds the fixed discrete threshold and the transient induced voltage drop variation rate is less than or equal to the fixed voltage distortion threshold, the timing control module reconstructs the distributed sensor network time slice matrix based on the load discrete state index and the network topology identifier of the network element node, and assigns asynchronous transmission delay compensation time to each network element node. In step S106, the communication control module encapsulates the reconstructed distributed sensor network time slice matrix into timing configuration instructions and distributes them to each network element node to adjust the data throughput status of the gateway module.

2. The method for intelligent energy consumption control SaaS service management of industrial enterprises according to claim 1, characterized in that, Step S105 involves reconstructing the distributed sensor network time slice matrix, including the following sub-steps: Step S1051, using a first-order discrete differential trend extrapolation algorithm to calculate the rate of change of the queue length of the gateway module input buffer in the past three control cycles, and predicting the expected traffic throughput trend in the next control cycle; Step S1052, combining the expected traffic throughput trend to calculate the timing peak compensation parameters, and reconstructing the peak transmission time slice allocation matrix in the distributed sensor network time slice matrix.

3. The method for intelligent energy consumption control SaaS service management of industrial enterprises according to claim 1, characterized in that, After assigning asynchronous transmission delay compensation time to each network element node in step S105, the following steps are included: Step S301, calculate the cumulative standard deviation of each transmission delay compensation time within the sliding detection window; Step S302, if the cumulative standard deviation exceeds the fixed variable gate, issue a benchmark adjustment instruction to the congestion calculation module to dynamically deduct the fixed correction step size from the fixed benchmark threshold in the next control cycle; if the cumulative standard deviation does not exceed the fixed variable gate, maintain the fixed benchmark threshold at the steady-state initial value.

4. The method for intelligent energy consumption control SaaS service management of industrial enterprises according to claim 1, characterized in that, It also includes the following steps: Step S401: Deploy a discrete bypass preemption control channel directly driven by a hardware timer at the physical network adapter layer of the gateway module to bypass the operating system interrupt queue; S402, when triggered by the hardware input buffer overflow pin, directly broadcasts a fixed-duration pause transmission frame to the network element link through the discrete bypass preemption control channel, suspending the regular status statistics messages of low-priority network element nodes.

5. The method for intelligent energy consumption control SaaS service management of industrial enterprises according to claim 4, characterized in that, In step S402, a fixed-duration pause frame is broadcast to open a low-latency transmission window for the downlink distribution of timing configuration instructions and to prevent deadlock of buffered data in the gateway module protocol stack when it is in an overflow state.

6. The method for intelligent energy consumption control SaaS service management of industrial enterprises according to claim 1, characterized in that, It also includes the following steps: Step S601: In the gateway module, discrete network data packets are decoupled into high-priority congestion control frames and regular non-real-time status statistics frames. Step S602: In the case of a slight queuing delay state generated by the gateway module, the software-defined topology is started, and the regular non-real-time status statistics frames are redirected to a low-rate bypass communication link to isolate the transmission of high-priority congestion control frames.

7. The method for intelligent energy consumption control SaaS service management of industrial enterprises according to claim 6, characterized in that, The rule for determining that the gateway module has a slight queuing delay state in step S602 is: when the load discrete state index is greater than 0 and less than or equal to the fixed discrete threshold, the software-defined topology is triggered to start.

8. The method for intelligent energy consumption control SaaS service management of industrial enterprises according to claim 1, characterized in that, Before reading the transient induced voltage drop variation rate of the transmission cable in step S103, the method further includes: using the power grid state parameter acquisition module to collect the voltage waveform of the transmission cable at a fixed sampling frequency of 50Hz, calculating the absolute value of the slope of the transient voltage drop through waveform distortion characteristic analysis, and determining the transient induced voltage drop variation rate of the transmission cable.

9. The method for intelligent energy consumption control SaaS service management of industrial enterprises according to claim 1, characterized in that, After distributing the timing configuration instruction in step S106, the method further includes: sending the timing configuration instruction to the distributed energy consumption monitoring node, restricting the distributed energy consumption monitoring node to report high-frequency sensor data messages within a specified time slice, and smoothing out the transient induced voltage drop rate of the transmission cable.

10. An intelligent energy consumption management and control SaaS service management system for industrial enterprises, used to implement the intelligent energy consumption management and control SaaS service management method for industrial enterprises as described in claim 1, characterized in that, include: The gateway module is used to receive discrete network data packets transmitted through network element links. These discrete network data packets carry high-frequency sensor data messages. The congestion calculation module is used to count the total number of queued packets in the input buffer of the gateway module and determine the queuing delay in the buffer. The timing control module is used to read the buffer queuing delay and the transient induced voltage drop variation rate of the transmission cable obtained by the power grid state parameter acquisition module, and calculate the load discrete state index based on the deviation ratio between the buffer queuing delay and the fixed reference threshold. The load discrete state index is the ratio of the difference between the buffer queuing delay and the fixed reference threshold to the fixed reference threshold. When the load discrete state index exceeds the fixed discrete threshold and the transient induced voltage drop variation rate is less than or equal to the fixed voltage distortion threshold, the timing control module reconstructs the distributed sensor network time slice matrix based on the load discrete state index and the network topology identifier of the network element node, and assigns asynchronous transmission delay compensation time to each network element node. The communication control module is used to encapsulate the reconstructed distributed sensor network time slice matrix into timing configuration instructions and distribute them to each network element node, thereby adjusting the data throughput status of the gateway module.

Images