Distributed sequential logic conflict resolution method for distributed control system

Abstract

The invention discloses a distributed sequential logic conflict resolution method for a distributed control system, which relates to the technical field of automatic control, and can accurately calculate the queuing retention time length of an instruction in a network by acquiring the real-time load and identifier of each network node in an instruction transmission path, thereby improving the accuracy of the queuing retention time length of the instruction. The local clock of the controller is subjected to high-precision correction by combining the basic transmission delay of the link, so that the real sequential relationship of the instruction is restored, and the interference of clock asynchronization and network delay fluctuation in the distributed system on sequential judgment is fundamentally eliminated; meanwhile, on the basis of a real-time working condition self-adaptive physical constraint model, the theoretical duration needed by equipment state change is quantified, the corrected clock difference value and the theoretical duration are dynamically matched, causal instructions can be effectively distinguished from the physical level, and the accuracy and reliability of conflict resolution are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of automation control technology, and in particular to a method for resolving distributed sequential logic conflicts in a distributed control system. Background Technology

[0002] Distributed Control Systems (DCS) are core infrastructure in modern industrial automation, widely used in continuous production processes such as petrochemicals, power energy, and metallurgical manufacturing. A DCS typically consists of multiple controllers, field devices, and a communication network. Each controller independently collects field data and generates control commands, which are then transmitted to actuators via the industrial network. With the increasing scale of industrial systems and the growing complexity of control tasks, multi-controller collaborative operation has become the norm. However, this brings with it the problem of command timing conflicts: when two or more controllers issue commands to the same target device or related devices, if the arrival time and execution order of the commands do not meet the process logic requirements, it may lead to equipment malfunctions, decreased production efficiency, or even safety accidents.

[0003] To address timing conflicts, existing technologies primarily focus on the following areas: First, employing global clock synchronization protocols (such as IEEE 1588 PTP) for high-precision time synchronization of controllers, enabling conflict resolvers to directly compare the local timestamps carried by instructions to determine their order; second, introducing priority arbitration mechanisms to pre-set static priorities for different controllers or instruction types, preserving higher-priority instructions in case of conflicts; and third, using network latency compensation methods to coarsely correct instruction timestamps by measuring average network latency or employing fixed compensation values. However, these technologies have numerous limitations in practical applications.

[0004] Existing global clock synchronization schemes require all controllers and network devices to support precision time protocols, resulting in high hardware costs and sensitivity to network topology. Furthermore, they struggle to guarantee long-term stability in complex electromagnetic environments, such as industrial settings. Static priority arbitration mechanisms lack flexibility and cannot adapt to changing operating conditions; low-priority commands involving emergency safety actions can lead to serious consequences. Methods based on fixed delay compensation ignore dynamic fluctuations in network load, and queuing delay, a major component of transmission delay, is simply averaged, leading to significant discrepancies between the corrected clock and the actual issuance time. More critically, none of these methods consider the time required for the physical process corresponding to the command: in industrial control, the time from issuing a command to its effect often requires a dynamic response process from the equipment itself (such as changes in valve opening or temperature rise). Ignoring this physical constraint can lead to fundamental errors in causal judgment; for example, misjudging a later-issued emergency shutdown command as the cause and an earlier-issued dangerous opening command as the effect, thus causing an accident. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of the embodiments of the present invention and to briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section, as well as in the abstract and title of the present application, to avoid obscuring the purpose of this section, the abstract and title of the invention. Such simplifications or omissions shall not be used to limit the scope of the present invention.

[0006] In view of the aforementioned existing problems, the present invention is proposed.

[0007] Therefore, the technical problem solved by this invention is: how to provide a distributed timing logic conflict resolution method that can accurately correct network transmission delay, integrate device physical dynamic characteristics, and make reasonable decisions when timing is ambiguous.

[0008] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a distributed timing logic conflict resolution method for a distributed control system, characterized in that it includes: acquiring a first conflict command and a second conflict command issued by a first controller and a second controller respectively when a timing conflict occurs; the first conflict command includes a first feature pair consisting of a first local clock stamp and a first physical characteristic parameter, and a first forwarding trace set; the second conflict command includes a second feature pair consisting of a second local clock stamp and a second physical characteristic parameter, and a second forwarding trace set; calculating the first conflict command and the second conflict command in network transmission based on the first forwarding trace set and the second forwarding trace set. The first and second dwell times are determined; the first and second local clock stamps are corrected using the first and second dwell times respectively to obtain the first corrected clock and the second corrected clock; the first and second physical characteristic parameters are input into the physical constraint model to obtain the first theoretical time required for the first physical characteristic parameter to be converted to the second physical characteristic parameter, and the second theoretical time required for the second physical characteristic parameter to be converted to the first physical characteristic parameter; the clock difference between the first and second corrected clocks is calculated, the clock difference is matched with the first and second theoretical times, and the reserved instruction is selected as the conflict resolution result output according to the causal order.

[0009] In a preferred embodiment of the present invention, when the first conflict instruction passes through the first network node, the first network node collects its current first network node load data, its current first network node local timestamp, and its first network node identifier, and adds the three together as a first network node forwarding trace to the first conflict instruction to generate a first intermediate instruction; when the first intermediate instruction passes through the second network node, the second network node collects its current second network node load data, its current second network node local timestamp, and its second network node identifier, and adds the three together as a second network node forwarding trace to the first intermediate instruction to generate a first conflict instruction carrying a first feature pair and a first forwarding trace set consisting of the first network node forwarding trace and the second network node forwarding trace; the conflict resolver receives the first conflict instruction, records the local reception timestamp, and extracts the first feature pair and the first forwarding trace set from the first conflict instruction.

[0010] In a preferred embodiment of the present invention, the second conflict instruction issued by the second controller is subjected to steps symmetrical to the first conflict instruction to obtain a second feature pair, a second forwarding trace set, and a corresponding local receiving timestamp.

[0011] In a preferred embodiment of the present invention, the conflict resolver parses the first forwarding trace set to extract the first network node load sequence and the corresponding first network node identifier sequence arranged in forwarding order; similarly, it parses the second forwarding trace set to obtain the second network node load sequence and the corresponding second network node identifier sequence; it reads the service rate parameters corresponding to each network node and the sum of the transmission delays of each network link from the system configuration database as the basic transmission delay; based on the first network node load sequence and the corresponding first network node identifier sequence, it uses the service rate parameters of each network node to determine the queuing delay of the first conflict instruction at each forwarding network node, and accumulates all queuing delays to obtain the first dwell time; similarly, it determines the second dwell time based on the second network node load sequence and the corresponding second network node identifier sequence.

[0012] In a preferred embodiment of the present invention, the first dwell time is added to the basic transmission delay to obtain the first total transmission time; the first local reception timestamp recorded by the conflict resolver is subtracted from the first total transmission time to obtain the first corrected clock; similarly, the second dwell time is added to the basic transmission delay to obtain the second total transmission time, and the second local reception timestamp is subtracted from the second total transmission time to obtain the second corrected clock.

[0013] In a preferred embodiment of the present invention, the following steps are performed: Based on the first local clock stamp in the first conflict instruction and the second local clock stamp in the second conflict instruction, historical operating condition data of the target equipment at corresponding times are extracted from the historical database to form a first operating condition parameter set and a second operating condition parameter set. If there is no data at a precise time in the historical database, linear interpolation or nearest neighbor interpolation is used to obtain the data. A weighted average is performed on the first and second operating condition parameter sets to obtain the current operating condition parameter set. The first physical characteristic parameter is used as the initial state value, and the second physical characteristic parameter is used as the target state value. The current operating condition parameter set is input into the equipment state transition duration calculation model. The equipment state transition duration calculation model predicts the dynamic characteristic parameters of the equipment at the current time online based on the current operating condition parameter set, and determines the first theoretical duration required for the change from the initial state value to the target state value based on the dynamic characteristic parameters. The initial state value and the target state value are exchanged, and the second physical characteristic parameter is used as the new initial state value, and the first physical characteristic parameter is used as the new target state value. The current operating condition parameter set remains unchanged and is input into the equipment state transition duration calculation model to obtain the second theoretical duration.

[0014] As a preferred embodiment of the present invention, the equipment state transition time calculation model is a first-order inertial plus pure time delay model. The model parameters are corrected online according to the current operating condition parameter set, and the theoretical time required for the change from the initial state value to the target state value is calculated based on the step response inverse calculation principle.

[0015] In a preferred embodiment of the present invention, the conflict resolver calculates the clock difference between the first corrected clock and the second corrected clock; acquires network latency fluctuation data in the most recent n historical conflict resolution cases, and determines the network latency fluctuation range based on the statistical distribution of the fluctuation data; acquires the prediction deviation data of the device state transition duration calculation model in historical conflict resolution cases, and determines the model prediction error range based on the statistical distribution of the prediction deviation data; combines the network latency fluctuation range and the model prediction error range to obtain the current uncertainty tolerance interval; determines whether the clock difference falls within the uncertainty tolerance interval based on the first theoretical duration, and if so, determines that the first theoretical duration matches; determines whether the clock difference falls within the uncertainty tolerance interval based on the second theoretical duration, and if so, determines that the second theoretical duration matches.

[0016] As a preferred embodiment of the present invention, the causal order is determined based on the matching result: if only the first theoretical duration matches, the first conflicting instruction is determined to be the cause and the second conflicting instruction to be the effect; if only the second theoretical duration matches, the second conflicting instruction is determined to be the cause and the first conflicting instruction to be the effect; if both the first and second theoretical durations match or neither matches, the process proceeds to dynamic criticality assessment; the retained instruction selected based on the causal order determination or dynamic criticality assessment is used as the conflict resolution result and output to the actuator of the distributed control system.

[0017] In a preferred embodiment of the present invention, the dynamic criticality assessment includes: acquiring real-time operating data of the target equipment, wherein the real-time operating data includes at least load rate, medium temperature, and wear coefficient; assessing the criticality of the first conflicting instruction and the second conflicting instruction based on the real-time operating data; the criticality is determined by analyzing the deviation of the first or second physical characteristic parameter involved in the instruction from the safety boundary under the current operating conditions, the urgency of the action represented by the instruction, and the sensitivity of the equipment's current health status to the execution of the instruction; comparing the first and second conflicting instructions in the order of deviation priority, urgency second, and response sensitivity last, and selecting the instruction with higher priority as the reserved instruction; if all indicators are the same, randomly selecting one of them as the reserved instruction.

[0018] The beneficial effects of this invention are as follows: By collecting the real-time load and identifiers of each network node in the instruction transmission path, this invention can accurately calculate the queuing and delay time of instructions in the network, and combine the basic transmission delay of the link to perform high-precision correction of the controller's local clock, thereby restoring the true timing relationship of instructions and fundamentally eliminating the interference of clock asynchrony and network delay fluctuations on timing judgment in distributed systems. At the same time, based on a real-time adaptive physical constraint model, the theoretical duration required for equipment state changes is quantified, and the corrected clock difference is dynamically matched with the theoretical duration, which can effectively distinguish causal instructions from a physical perspective, significantly improving the accuracy and reliability of conflict resolution. Furthermore, when the model cannot make a unique determination, a hierarchical criticality assessment mechanism that comprehensively considers the degree of deviation from the safety boundary, the urgency of the action, and the sensitivity of the equipment ensures that the instruction most beneficial to the system's safety and stability is always selected for execution, greatly enhancing the robustness and intelligence level of the distributed control system under complex operating conditions, and effectively avoiding equipment malfunctions or operating condition oscillations caused by instruction conflicts. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a flowchart illustrating a distributed timing logic conflict resolution method for a distributed control system, as shown in this invention.

[0020] Figure 2 This is a logic diagram of a distributed timing logic conflict resolution method for a distributed control system according to the present invention. Detailed Implementation

[0021] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0022] Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort should fall within the scope of protection of this invention.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] According to an embodiment of the present invention, in combination Figure 1 and Figure 2 The flowchart shown illustrates a distributed sequential logic conflict resolution method for distributed control systems, comprising: S1: Obtain the first conflict instruction and the second conflict instruction issued by the first controller and the second controller respectively when there is a timing conflict; the first conflict instruction includes a first feature pair consisting of a first local clock stamp and a first physical feature parameter, and a first forwarding trace set; the second conflict instruction includes a second feature pair consisting of a second local clock stamp and a second physical feature parameter, and a second forwarding trace set.

[0025] It should be noted that in existing distributed control systems (DCS), when the time order in which multiple controllers receive instructions from the conflict resolver differs from the actual order in which the instructions are issued due to asynchronous distributed clocks or network transmission delays, traditional conflict resolution methods typically rely solely on the local clock stamps carried by the instructions for sorting or use preset priority rules for adjudication. However, such methods have the following drawbacks: First, the local clock stamps are not corrected for network transmission delays and cannot reflect the true timing relationship of the instructions; second, the various network nodes (such as switches and gateways) that the instructions pass through in the transmission path introduce queuing delays, but existing technologies do not collect and utilize the status information of these nodes, making it impossible to accurately estimate the actual transmission time of the instructions; third, there is a lack of consideration for the physical meaning represented by the instructions (such as the theoretical time required for equipment status changes), making it difficult to distinguish causal instructions. To solve the above problems, this invention, during the instruction transmission process, requires each network node to actively attach its own load data, local timestamp, and node identifier when forwarding instructions, forming a traceable transmission path status record. These forwarding traces form the basis for subsequent calculations of latency and clock correction, enabling the conflict resolver to restore the true issuance time of the instructions. Specifically: S1.1: When the first conflict instruction passes through the first network node, the first network node collects its current first network node load data, current first network node local timestamp, and first network node identifier, and adds the combination of the three as the first network node forwarding trace to the first conflict instruction to generate the first intermediate instruction.

[0026] It should be noted that, in this embodiment of the invention, the conflict resolver acts as the central adjudication unit in the distributed control system, continuously monitoring the command flow from each controller. When it is detected that commands issued by two or more controllers for the same target device or related device overlap in time (i.e., timing conflict), the conflict resolver initiates the conflict resolution process, acquiring the first conflict command issued by the first controller and the second conflict command issued by the second controller.

[0027] The first network node refers to the first forwarding device on the transmission path from the first controller to the conflict resolver. After receiving the original first conflict command from the first controller, this node does not forward it immediately, but first triggers a data acquisition and additional process.

[0028] Specifically, the first network node performs the following operations: First, the first network node reads its internally maintained real-time load status, including but not limited to the queue depth of the output port, the number of packets waiting to be forwarded in the current buffer, and the recent bandwidth utilization. This data quantifies the congestion level of the node at the moment of forwarding the first conflict command. In this embodiment, the first network node's load data is specifically selected as the real-time queue length of the output port, expressed in units of data packets.

[0029] Secondly, the first network node accesses its internal hardware clock or the system time maintained by software to obtain the precise value of the current moment. This timestamp is used to subsequently verify the authenticity of the forwarding traces and the rationality of the time series in multi-node forwarding scenarios. Since the clocks of each network node may be independent of each other, this timestamp does not require global synchronization and only serves as a local time reference for the node.

[0030] Furthermore, the first network node appends a pre-defined unique identifier to its forwarding trace. This identifier is used in subsequent steps to query the node's service rate parameters, physical location, and other information from the system configuration database.

[0031] After collecting the above three data points, the first network node encapsulates them into a single first network node forwarding trace data block according to a predefined data format and appends it to the end of the original first conflict instruction (or inserts it into the instruction's extension field). At this point, the content of the original instruction (including the first feature pair) remains unchanged; only the new trace data block is added.

[0032] After this operation, the instruction is upgraded to a first intermediate instruction, which carries a first feature pair and the newly added first network node forwarding trace.

[0033] The first network node then forwards the first intermediate instruction to the next-hop node according to normal routing rules. Through the above operations, the first network node completes the role transformation from transparent transmission to state-aware transmission.

[0034] S1.2: When the first intermediate instruction passes through the second network node, the second network node collects its current second network node load data, current second network node local timestamp, and second network node identifier, and adds the three as the second network node forwarding trace to the first intermediate instruction, generating a first conflict instruction carrying a first feature pair and a first forwarding trace set consisting of the first network node forwarding trace and the second network node forwarding trace.

[0035] In this context, the second network node refers to the forwarding device immediately following the first network node on the transmission path from the first controller to the conflict resolver. Upon receiving the generated first intermediate instruction, this node recognizes that the instruction already contains the forwarding trace of the first network node and continues to perform similar additional operations. The specific process is as follows: The second network node reads the current instantaneous load status, such as the output port queue depth, reflecting the congestion level of the second network node at the time of forwarding; the second network node obtains the current value of its local clock as a timestamp for the forwarding time; and the second network node appends a unique identifier to the forwarding trace.

[0036] After the data collection is completed, the second network node encapsulates the three data points into a second network node forwarding trace data block and appends it to the end of the first intermediate instruction, placing it after the existing first network node forwarding trace.

[0037] At this point, the instruction already contains two forwarding traces arranged in sequence: the first network node forwarding trace (first) and the second network node forwarding trace (second). These two traces, arranged in forwarding order, constitute the prototype of the first forwarding trace set.

[0038] After this operation, the instruction is upgraded to the final form of the first conflict instruction, which carries a first feature pair and a first forwarding trace set consisting of two forwarding traces.

[0039] The second network node then forwards the first conflict command to the next hop. If there are more network nodes in the path, each node will repeat the above process, appending its own forwarding trace to the end of the command in turn, until the command reaches the conflict resolver.

[0040] Ultimately, the first forwarding trace set will contain the state information of all network nodes on the path from the first controller to the conflict resolver, and the order will be strictly consistent with the forwarding path.

[0041] It should be noted that the first forwarding trace set finally obtained by the conflict resolver contains the load data and identifier of each node, which can accurately reconstruct the queuing delay of each hop the instruction went through. This path holographic recording method provides complete data support for subsequent accurate calculation of the delay time and avoids estimation errors caused by missing path information.

[0042] S1.3: The conflict resolver receives the first conflict instruction, records the local received timestamp, and extracts the first feature pair and the first forwarding trace set from the first conflict instruction.

[0043] It should be noted that the conflict resolver, acting as the central node in the distributed control system, continuously monitors the network port. When the first conflict command arrives, the conflict resolver first triggers a hardware or software interrupt to record the precise time of command arrival, i.e., the local received timestamp. This timestamp is based on the conflict resolver's own local clock and is used for subsequent comparison with a corrected clock.

[0044] After completing the reception time recording, the conflict resolver calls the built-in instruction parsing module to perform structured parsing of the data packet of the first conflicting instruction. The parsing process consists of two parallel steps: The instruction parsing module locates the instruction header or a specified field, reads the first local clock stamp and the first physical characteristic parameter, combines them into a first feature pair, and stores it in a temporary storage area. The instruction parsing module traverses the extended area of ​​the instruction data packet, sequentially identifying and extracting each independent forwarding trace data block according to a predefined trace data format. Since the traces are arranged in an appended order, the instruction parsing module maintains this order during extraction, forming an ordered list of forwarding traces. For instructions generated in S1.2, this list contains two elements: the first network node forwarding trace and the second network node forwarding trace. If there are more nodes in the path, the list expands accordingly.

[0045] After extraction, the conflict resolver associates and stores the first feature pair, the first forwarding trace set, and the local received timestamp. The original content of the first conflicting instruction (such as the control instruction itself) is then passed to the execution mechanism or temporary storage area, awaiting processing after the conflict resolution result is determined.

[0046] Ultimately, the conflict resolver processes the second conflict command issued by the second controller in exactly the same manner as or shortly after receiving the first conflict command. During its transmission from the second controller to the conflict resolver, the second conflict command also passes through a series of network nodes (possibly along a different path than the first command). Each node appends its own forwarding trace, forming a second set of forwarding traces. When the conflict resolver receives the second conflict command, it records the local reception timestamp and extracts the second feature pair and the second set of forwarding traces (containing the load data, local timestamps, and node identifiers of each network node along the way).

[0047] It should be noted that the processing of the first and second instructions is independent and completely symmetrical.

[0048] S2: Calculate the first and second dwell times of the first and second conflict instructions in network transmission based on the first and second forwarding trace sets; use the first and second dwell times to correct the first and second local clock stamps respectively to obtain the first corrected clock and the second corrected clock.

[0049] First, the conflict resolver parses the first forwarding trace set, extracting the first network node load sequence and its corresponding first network node identifier sequence arranged in forwarding order. Similarly, it parses the second forwarding trace set to obtain the second network node load sequence and its corresponding second network node identifier sequence. These two sequences are of equal length and their elements correspond one-to-one, reflecting the identifiers of each network node that the instruction passes through sequentially after originating from the first controller, and their corresponding load states.

[0050] S2.1: Read the service rate parameters corresponding to each network node and the sum of the transmission delays of each network link from the system configuration database, and use them as the base transmission delay.

[0051] Specifically, the service rate parameters corresponding to each network node are read from the system configuration database. Simultaneously, based on the first node identifier sequence in the first forwarding trace set, the transmission delay of each network link along the path is queried (e.g., determining the link between adjacent nodes based on the node identifier and obtaining the physical transmission delay of that link from the database), and these link delays are summed to obtain the basic transmission delay corresponding to the first instruction. Similarly, the basic transmission delay of the second instruction is calculated based on the second node identifier sequence. If the link delay is not directly stored in the database, it can also be estimated based on the distance between nodes and the medium type.

[0052] The basic transmission delay is a relatively fixed value, which mainly depends on the link length and the medium type, and remains basically unchanged when the network topology remains unchanged.

[0053] When it is necessary to calculate the dwell time of the first conflict command, the conflict resolver queries the database one by one according to the first network node identifier sequence obtained in S2.1, obtains the service rate parameter corresponding to each identifier, and forms a service rate sequence corresponding to the load sequence. At the same time, it reads the preset basic transmission delay value and records it as... .

[0054] S2.2: Based on the first network node load sequence and the corresponding first network node identifier sequence, the queuing delay of the first conflict instruction at each forwarding network node is determined using the service rate parameters of each network node, and all queuing delays are accumulated to obtain the first dwell time; similarly, the second dwell time is determined based on the second network node load sequence and the corresponding second network node identifier sequence.

[0055] Specifically, for each load value in the first network node load sequence ( (Indicates the order of nodes in the path), the conflict resolver queries the service rate based on the corresponding node identifier. Calculate the queuing delay of this node. The queuing delay is the time required for the node to process all data packets ahead of the current instruction under the current load before it can begin processing the instruction.

[0056] In the queuing theory model, if nodes use a first-in, first-out queue, and the load value represents the queue length (i.e., the number of packets waiting to be processed), then the queuing delay is equal to the queue length divided by the service rate. The service time is already implicitly included here. The conflict resolver calculates the queuing delay for each node in turn and sets all... Accumulate the results to obtain the first length of stay. This latency reflects the total time that the instruction spends waiting in the transmission path due to queuing.

[0057] S2.3: Add the first dwell time to the basic transmission delay to obtain the first total transmission time; subtract the first total transmission time from the first local reception timestamp recorded by the conflict resolver to obtain the first corrected clock.

[0058] The first total transmission time represents the total time elapsed from when the first controller issues a command to when the conflict resolver receives the command. It includes both the fixed delay of link propagation and the dynamic delay caused by network node queuing.

[0059] The first correction clock represents the estimated time when the first conflict command was actually issued, and this time is referenced to the clock of the conflict resolver. Similarly, the second delay time is added to the basic transmission delay to obtain the second total transmission time, and the second local reception timestamp is subtracted from the second total transmission time to obtain the second correction clock.

[0060] The two instructions are processed independently and in the same way, ensuring the symmetry and comparability of the correction results.

[0061] As can be seen, the corrected clock is based on the local clock of the conflict resolver, eliminating the time deviation caused by the asynchronous clocks of each controller and the network transmission delay. This allows the issuance times of two instructions to be directly compared on the same time axis. Furthermore, this invention does not depend on a specific network topology. Regardless of how many nodes are included in the path, it can be completely recorded and accurately calculated through the forwarding trace set, exhibiting good scalability and adaptability.

[0062] S3: Input the first physical feature parameter and the second physical feature parameter into the physical constraint model, and obtain the first theoretical time required to convert the first physical feature parameter to the second physical feature parameter, and the second theoretical time required to convert the second physical feature parameter to the first physical feature parameter.

[0063] It should be noted that in distributed control systems, control commands are typically used to adjust the physical state of target equipment (such as valves, motors, and heaters), for example, changing the opening degree, speed, or temperature setpoint. When commands issued by two controllers target the same or related equipment, it is difficult to determine the true causal relationship solely based on the order of time, because changes in physical state require a certain amount of time to complete. For example, if one command requests to open a valve and another command requests to close it, and a closing command is received shortly after the opening command is issued, the closing command may be a response to the state after opening or it may be an independent event. In existing technologies, conflict resolution often relies solely on timestamp sorting or preset priorities, without considering the dynamic response characteristics of the equipment itself, which may lead to incorrect causal judgments. To solve the above problems, the specific operation of this invention is as follows: S3.1: Based on the first local clock stamp in the first conflict instruction and the second local clock stamp in the second conflict instruction, extract the historical operating condition data of the target equipment at the corresponding time from the historical database to form the first operating condition parameter set and the second operating condition parameter set.

[0064] If there is no data in the historical database that corresponds to the exact time, then linear interpolation or nearest neighbor interpolation methods are used to obtain it.

[0065] The historical database continuously records real-time operating data of the target equipment, including but not limited to load rate (the ratio of the equipment's current load to its rated load), medium temperature (such as fluid temperature), and wear coefficient (characterizing the degree of equipment aging). Each data record is timestamped and based on the local clock.

[0066] In this embodiment of the invention, when calculating the theoretical duration, the conflict resolver first retrieves the first local clock stamp carried by the first conflict instruction. It should be noted that since the historical database uses a global clock (such as the conflict resolver clock) to record timestamps, while the first local clock stamp is the local clock of the first controller, there may be a discrepancy. Therefore, the conflict resolver uses the first corrected clock calculated in S2 (based on the conflict resolver clock) as the estimated global clock value for the time when the first instruction is issued, and uses the first corrected clock as the query index to extract the corresponding operating condition data from the historical database. For higher accuracy, the deviation between the local clock stamp and the corrected clock can be recorded simultaneously for subsequent conversion. Similarly, for the second instruction, the second corrected clock is used.

[0067] The conflict resolver searches the historical database for the data record with the closest timestamp based on the first correction clock. If an exact match exists, it directly reads the load rate, medium temperature, and wear coefficient at that moment to form the first operating condition parameter set C1. If no exact match exists, it searches for the two closest timestamps before and after the first correction clock and their corresponding operating condition data, using a linear interpolation method to calculate the operating condition value at the first correction clock time. For example, regarding the load rate, if the previous time... load rate The next moment load rate Then the first corrected clock time load rate :

[0068] The same treatment applies to medium temperature and wear coefficient. If the first correction clock exceeds the time range of the historical database, boundary values ​​are used or nearest neighbor interpolation is employed. This ultimately yields the first operating condition parameter set C1.

[0069] Similarly, the second operating condition parameter set C2 is extracted from the historical database based on the second correction clock.

[0070] It can be seen that by using the corrected clock as the basis for querying, it can be ensured that the extracted operating condition data matches the time when the command is issued as closely as possible.

[0071] S3.2: Perform a weighted average of the first and second operating condition parameter sets to obtain the current operating condition parameter set.

[0072] The weights can be determined based on the confidence level of the corrected clock or its proximity to the current conflict resolution time. It should be noted that the estimation of dwell time in S2 may contain errors, such as due to the instantaneous nature of node load data or the approximation of service rate parameters. If based on the confidence level of the corrected clock, these errors can be quantified by the estimated variance of the dwell time. This estimated variance can be calculated as the sample variance based on the deviation between the actual observed dwell time and the theoretical estimate in historical conflict cases. The variance of the total dwell time can be approximated as the sum of the variances of each node (assuming independence).

[0073] If the estimated variance of the first corrected clock is small, it indicates a high confidence level, and C1 is given a larger weight during fusion; conversely, if the variance is large, C1 is given a larger weight. The conflict resolver in S2 can simultaneously calculate the variance of the dwell time (e.g., based on the variance formula of the queuing model) and store it for later use. If the determination is based on the time proximity to the current conflict resolution time, the conflict resolver records the current conflict processing time (based on its own clock) and calculates the time differences Δt1 and Δt2 between the first and second corrected clocks and the current conflict processing time. The smaller the time difference, the newer the data for that operating condition, and the more representative it is of the current equipment state; therefore, it is given a larger weight. The weight can be the reciprocal of the time difference or after normalization.

[0074] In this embodiment, a weighting method based on the corrected clock confidence level is preferred, which directly reflects the estimation accuracy of step S2; if the confidence level data is unavailable (e.g., the variance is not calculated), a weighting method based on time proximity is used.

[0075] If neither method is used, or if the confidence levels and time differences are equal, then the arithmetic mean is taken.

[0076] S3.3: The first physical characteristic parameter is used as the initial state value, and the second physical characteristic parameter is used as the target state value. The current operating condition parameter set is input into the equipment state transition duration calculation model. The equipment state transition duration calculation model predicts the dynamic characteristic parameters of the equipment at the current moment online based on the current operating condition parameter set, and determines the first theoretical duration required for the change from the initial state value to the target state value based on the dynamic characteristic parameters.

[0077] The equipment state transition time calculation model is a first-order inertial plus pure time-delay model. The model parameters are corrected online based on the current operating condition parameter set, and the theoretical time required for the change from the initial state value to the target state value is calculated based on the step response inverse calculation principle. The transfer function of this model is as follows:

[0078] in, For gain, It is a time constant. For pure time delay, For the Plas operator.

[0079] During the offline phase, the model stores a set of baseline parameters. , , This corresponds to the dynamic characteristics of the equipment under standard operating conditions (such as rated load, standard temperature, and new equipment). However, in actual operation, the equipment's operating conditions (load rate, medium temperature, wear coefficient) will affect its dynamic response. For example, an increased load rate may lead to a slower response (increased time constant), and increased wear may lead to increased hysteresis. Therefore, the model needs to be adjusted online based on the current set of operating condition parameters.

[0080] The correction method can employ a pre-calibrated correction coefficient table or a mapping relationship established through a machine learning model (such as a neural network). In this embodiment, a combination of a pre-calibrated correction coefficient table and interpolation techniques is used to ensure the determinism and repeatability of the correction process. The specific technical solution is as follows: For load rate The system searches the load rate-gain correction factor table for a precisely matching load rate value. If a match is found, the corresponding correction factor is read directly; otherwise, the load rate is located. Two adjacent load rate values ​​in the table and and its corresponding correction coefficient and The calculation is performed using a linear interpolation formula:

[0081] like If the value exceeds the minimum or maximum value in the table, the correction coefficient corresponding to the boundary value is taken (i.e., nearest neighbor interpolation is used during extrapolation to avoid introducing unreliability).

[0082] Similarly, for medium temperature, the correction coefficient can be obtained by looking up the table and interpolating from the temperature-time constant correction coefficient table; similarly, for wear coefficient, the correction coefficient can be obtained by looking up the table and interpolating from the wear coefficient-hysteresis time correction coefficient table. It should be noted that the above three correction coefficient tables are pre-constructed discrete point sets pre-built by measuring the actual dynamic characteristic parameters of the target equipment under different operating conditions (including different load rates, different medium temperatures, and different wear coefficients) during the offline calibration phase, calculating the ratios to the reference parameters (gain under rated load, time constant at standard temperature, and hysteresis time under virgin conditions), and storing them in the system configuration database. These tables are used to obtain the correction coefficient corresponding to the current operating condition through table lookup and linear interpolation during real-time operation, thereby correcting the model parameters online.

[0083] Furthermore, after obtaining the three correction coefficients, the reference parameters are used respectively. , , Multiply by the three correction factors to calculate the current dynamic characteristic parameters (current gain, current time constant, and current lag time).

[0084] Obtain the corrected model parameters and 、 After that, the model starts to calculate the theoretical duration. Given the starting state value and the target state value , it is necessary to solve for the time required to change from to . For a first-order inertia plus pure dead-time link, under a step input, the output response is: when t < L, the output remains at the initial value ; where t is the time variable, representing the elapsed time starting from the instant when the step input is applied; when t ≥ L, the output approaches the steady-state value according to an exponential law, where is the equivalent step input amplitude. In practice, is usually determined by the controller output, but here the starting and target states are known, and the required equivalent input can be deduced backwards.

[0085] Assume that the device was originally in a stable state , and the controller applies a step input to make the final steady-state value become , and it is desired that the output reaches at a certain moment.

[0086] where .

[0087] From this, it is solved that

[0088] But here is unknown. It should be noted that the positive or negative gain does not affect the calculation of the theoretical duration. In fact, is related to the target steady-state value. If the desired final steady-state value is , then is required, so . Substituting gives , but it is meaningless. This reflects the problem: if the target value is the steady-state value, then theoretically it takes an infinite time to reach exactly, but in practice it is usually considered that when the output enters a certain error band (such as ±2%) of the steady-state value, it is considered to have reached. Therefore, the model needs to allow an error band , for example, taking (when ), then the time within the range of reaching is .

[0089] A more general approach is: Let the desired target value be However, allowable error Then solve:

[0090] in, (Because the final steady state is P2).

[0091] Substituting, we get:

[0092]

[0093]

[0094]

[0095]

[0096] Among the requirements ,Right now .like ,but (Instantly reached within the error band).

[0097] Therefore, the model calculates based on the preset error band δ (usually taken as 1% to 5% of the steady-state value). ,like ,but This indicates that the state has not changed.

[0098] It should be noted that in the above calculations, it is necessary to ensure that and Same sign, otherwise it cannot be achieved. If If the input exceeds the reachable range (e.g., the gain is negative or the input is limited), the model returns infinity or a maximum value.

[0099] Through the above reverse deduction, the model obtains the first theoretical duration. .

[0100] It should be noted that this invention is based on the FOPDT model and error band inverse calculation, which can quantify the time required for changes in physical state, and the model parameters are adaptively adjusted according to the working conditions, thus improving the accuracy of calculation.

[0101] S3.4: Exchange the initial state value and the target state value, take the second physical characteristic parameter as the new initial state value and the first physical characteristic parameter as the new target state value, keep the current operating condition parameter set unchanged, input it into the equipment state transition duration calculation model, and obtain the second theoretical duration.

[0102] That is, to As the initial state As the target state, the remaining parameters (current operating condition parameter set, model parameters) , , Keeping the time constant, repeat the calculation process of S3.3 to obtain the second theoretical duration. Note that at this time... The signs may be opposite, but the calculation process is the same, and the error band is... They can take the same value. If ,but .

[0103] By calculating the theoretical durations for both the forward and reverse directions, the required time for both directions was obtained, providing a benchmark for subsequent causal matching.

[0104] As can be seen, this invention calculates both the forward and reverse theoretical durations simultaneously, providing a symmetrical reference for subsequent matching. This ensures that regardless of which instruction is the cause, a corresponding theoretical duration can be found. Furthermore, the model parameters are corrected online based on real-time operating conditions, adapting to factors such as equipment aging and environmental changes. The theoretical duration calculation is closer to reality.

[0105] S4: Calculate the clock difference between the first corrected clock and the second corrected clock, match the clock difference with the first theoretical duration and the second theoretical duration, and select the reserved instruction as the conflict resolution result output according to the causal order.

[0106] S4.1: The conflict resolver calculates the clock difference between the first corrected clock and the second corrected clock. This difference reflects the temporal order of the estimated issuance times of the first and second instructions: if This indicates that the first instruction was issued before the second instruction; if This indicates that the second instruction was issued before the first instruction; if This indicates that both are emitted simultaneously (theoretically, the probability is extremely low). absolute value This indicates the time interval between the issuance of two instructions. It should be noted that the unified reference for the corrected clock makes... It can accurately reflect the true temporal relationship.

[0107] S4.2: Obtain network latency fluctuation data in the most recent n historical conflict resolution cases, and determine the network latency fluctuation range based on the statistical distribution of the fluctuation data; obtain the prediction deviation data of the device state transition duration calculation model in historical conflict resolution cases, and determine the model prediction error range based on the statistical distribution of the prediction deviation data.

[0108] For network latency fluctuations, this embodiment quantifies them based on the statistical characteristics of latency duration in historical conflict cases. Specifically, the conflict resolver records the first and second latency durations calculated for each conflict resolution, and maintains a sliding window to store the latency duration data of the most recent n conflict cases. The sample standard deviation of these latency durations is calculated. .Pick Several times (e.g., 3 times) as the range of network latency fluctuation This represents the fluctuation range of network latency in most cases. If historical data is insufficient (e.g., n < 30), a conservative value preset by the system is used as the initial fluctuation range, which is dynamically updated after sufficient data is accumulated. In addition, changes in network round-trip time measured periodically by network monitoring can also be used to assist in verification, but the primary method is based on statistics of the latency itself.

[0109] Regarding model prediction errors, after each conflict resolution, if subsequent observations show changes in equipment status, the deviation between the theoretical predicted duration and the actual change time can be calculated and recorded in the historical case database. The conflict resolver then calculates the standard deviation by summarizing the absolute values ​​of these deviations over the most recent n historical cases. Take 3 As the model prediction error range .

[0110] comprehensive and Obtain the uncertainty tolerance range The synthesis method can use the root mean square (RMS) as an example: Or take the sum of the two. Alternatively, the maximum value can be taken. This embodiment preferably uses the root mean square (RMS) because it can reasonably reflect the combined effect of two independent error sources. If historical data is insufficient (e.g., n < 30), a system-preset conservative value is used as the initial tolerance range, which is dynamically updated after sufficient data has been accumulated.

[0111] S4.3: By combining the range of network latency fluctuations and the range of model prediction errors, the current uncertainty tolerance range is obtained.

[0112] Specifically, let the first theoretical duration be... The second theoretical duration is For the first theoretical duration, the calculation is based on... Centered on, with radius interval It is important to note the clock difference. It can be positive or negative, but the theoretical duration is always non-negative, so this needs to be considered during matching. The relationship between the absolute value and the theoretical duration. The specific method for determining this is: calculate... and The absolute value of the difference: ,like Then it is considered that the clock difference falls within the range of The tolerance interval centered on this determines the first theoretical duration matching. Similarly, the calculation... and The absolute value of the difference ,like Then the second theoretical duration is matched.

[0113] It should be noted here that the following is used Instead This is because theoretical duration describes time intervals without direction; while the sign of clock difference (i.e.) The sign (positive or negative) is used separately in subsequent causality determination to indicate which instruction came first. Therefore, in the matching phase, only whether the interval size matches the theoretical duration is of concern.

[0114] like or If the value is infinity (meaning that it is physically impossible to convert), then a mismatch is directly determined.

[0115] S4.4: Determine whether the clock difference falls within the uncertainty tolerance range based on the first theoretical duration. If so, determine that the first theoretical duration matches. Determine whether the clock difference falls within the uncertainty tolerance range based on the second theoretical duration. If so, determine that the second theoretical duration matches.

[0116] In this embodiment of the invention, causality is determined based on the matching result of S4.3 and the sign of the clock difference. The specific rules are as follows: If only the first theoretical duration matches (i.e.) and The deviation is within U, while... The deviation exceeds If so, further investigation is needed. The symbol: If This indicates that the first instruction was issued before the second instruction, and the time interval is consistent with the theoretical time required to transition from the first state to the second state. Therefore, the first instruction is determined to be the cause and the second instruction to be the effect. This indicates that the second instruction was issued earlier than the first instruction, but the time interval is consistent with the theoretical time from the first to the second, which is physically unreasonable (because if the second instruction is earlier, then it may be the cause, but the time interval corresponds to from the first to the second, which is a contradiction). Therefore, the first instruction should not be judged as the cause, but should be regarded as a mismatch.

[0117] In reality, when only the first theoretical duration matches, the time interval must be consistent with the theoretical duration. Since the theoretical duration itself is positive, the earlier instruction must be the cause. Therefore, it can be simplified to: if only the first theoretical duration matches, and (If the first precedes the second), then the first is the cause; if only the first theoretically matches in duration but... If this is a logical contradiction, it should be considered a mismatch and proceed to dynamic criticality assessment. Similarly, if only the second theoretical duration matches, and (If the second precedes the first), then the second is the cause; if The contradiction then leads to a critical assessment.

[0118] If only the first theoretical duration matches and If so, then the first instruction is determined to be the cause and the second instruction to be the effect.

[0119] If only the second theoretical duration matches and If so, then the second instruction is determined to be the cause and the first instruction to be the effect.

[0120] If both the first and second theoretical durations match, it means that regardless of which instruction is the cause, the time interval conforms to the corresponding theoretical duration, and causality cannot be distinguished. If neither matches, it means that the time interval is inconsistent with any theoretical duration and cannot be explained by a physical model. Both of these cases require dynamic criticality assessment.

[0121] Further, dynamic key assessments include: (1) Obtain the current real-time operating data of the target equipment. The real-time operating data shall include at least the load rate, medium temperature and wear coefficient. The real-time operating data shall be collected directly from the field to ensure that the assessment is based on the latest status and is timely.

[0122] (2) Assess the criticality of the first and second conflicting commands based on real-time operating data: The criticality is determined by analyzing the deviation of the first or second physical characteristic parameter involved in the command from the safety boundary under the current operating conditions, the urgency of the action represented by the command, and the sensitivity of the equipment's current health status to the command execution. Specifically, the process for determining the deviation is as follows: For the first physical characteristic parameter, calculate the relative deviation from the current safety boundary. For example, if the first physical characteristic parameter is the desired load rate, the degree of deviation can be quantified as follows: if the first physical characteristic parameter is greater than the upper limit of the load rate, then the deviation is calculated by normalizing the load rate upper limit. If the deviation is less than the lower limit of the load rate, the normalization is calculated based on the lower limit of the load rate as the degree of deviation; if the first physical characteristic parameter is within the safe range, the degree of deviation = 0. Similarly, the degree of deviation for the second instruction is calculated. The greater the deviation, the more the instruction tends to push the equipment towards a dangerous area, and therefore the more critical it is. The process of determining the urgency level is as follows: if the state change required by the instruction is large, or the rate of change is fast (which can be compared with historical change rates), the urgency level is considered high. For example, calculate the absolute value of the first physical characteristic parameter and the current actual state value of the equipment; the larger the difference, the higher the urgency level. Alternatively, if the current state of the equipment is close to the danger boundary, and the instruction attempts to pull it back to the safe zone, the urgency level is also high. In this embodiment, the magnitude of change is used as the urgency level indicator. Calculate the same After normalization, the responses are compared. The process for determining response sensitivity is as follows: the larger the wear coefficient (the newer the equipment), the more sensitive the response, and the better the command execution effect; therefore, the command may be more important. Conversely, the smaller the wear coefficient, the slower the equipment response, the worse the execution effect of the same command, and the relatively lower the importance of the command. In this embodiment, response sensitivity... The wear coefficient can be directly taken (since it applies to the same device, the response sensitivity of the two commands is the same). However, in some scenarios, different commands may correspond to different actuators, in which case the wear coefficient of the corresponding actuator needs to be queried separately. and .

[0123] Furthermore, the conflict resolver combines the three indices of the first and second instructions into a vector. and According to the lexicographical order comparison rule: first compare the degree of deviation. ,like If the first instruction has higher priority, then the first instruction is selected; if If so, then select the second instruction; if Then compare the urgency. ,like Then choose the first one. Then choose the second option; if Then compare response sensitivity ,like Then choose the first one. Then choose the second option; if If the specified number is not specified, then a random instruction will be selected as the reserved instruction. Random selection can be achieved by generating random numbers, or by selecting any one of the instructions (such as the one with the smaller number).

[0124] As can be seen, the layered comparison in the above technical solution avoids the weighting issues of weighted summation, and its logic is clear and consistent with decision-making intuition: safety first, timeliness second, and execution effect third. If all indicators are the same, it means that the two instructions are equally critical, and random selection will not have a substantial impact.

[0125] The first and second conflicting instructions are compared in the following order: degree of deviation, urgency, and sensitivity of response. The instruction with higher priority is selected as the reserved instruction. If all indicators are the same, one of them is randomly selected as the reserved instruction. Random selection can be achieved by generating a random number or selecting an instruction with a smaller number.

[0126] Finally, the conflict resolver encapsulates the selected retained instructions (whether causal instructions obtained from causal determination or critical instructions obtained from criticality assessment) into an output format and sends it to the actuator of the target device through the distributed control system's communication network. The actuator performs the corresponding operation according to the instruction content. Simultaneously, the conflict resolver stores the process data of this conflict resolution (including corrected clock, theoretical duration, matching results, critical indicators, etc.) in a historical case database for subsequent updates to uncertainty parameters.

[0127] The method also includes one or more processors and memory.

[0128] The memory is used to store operable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations, including the flow of a distributed timing logic conflict resolution method for a distributed control system according to the foregoing embodiments, particularly... Figure 1 The flowchart of the method is shown.

[0129] Other aspects disclosed in the embodiments of the present invention also propose a computer-readable medium for storing software including instructions executable by one or more computers, which, upon execution, cause the one or more computers to perform operations including the flow of a distributed timing logic conflict resolution method for a distributed control system according to the foregoing embodiments, particularly... Figure 1 The flowchart of the method is shown.

[0130] It should be recognized that embodiments of the present invention may be implemented or carried out by computer hardware, a combination of hardware and software, or by computer instructions stored in a non-transitory computer-readable storage medium.

[0131] The method can be implemented using standard programming techniques, including a non-transitory computer-readable storage medium configured with a computer program in the computer program, wherein the storage medium is configured such that the computer operates in a specific and predefined manner.

[0132] Each program can be implemented in a high-level procedural or object-oriented programming language to communicate with the computer system; however, if required, the program can be implemented in assembly or machine language.

[0133] In any case, the language can be either compiled or interpreted.

[0134] Furthermore, for this purpose, the program can run on a programmed application-specific integrated circuit.

[0135] The processes described herein (or variations and / or combinations thereof) can be executed under the control of one or more computer systems configured with executable instructions, and can be implemented by hardware or a combination thereof as code (e.g., executable instructions, one or more computer programs, or one or more applications) that commonly executes on one or more processors. The computer program includes a plurality of instructions executable by one or more processors.

[0136] Furthermore, the method can be implemented in any suitable computing platform, including but not limited to personal computers, minicomputers, mainframes, workstations, networked or distributed computing environments, standalone or integrated computer platforms, or in communication with charged particle tools or other imaging devices.

[0137] Various aspects of the present invention can be implemented in machine-readable code stored on a non-transitory storage medium or device, whether portable or integrated into a computing platform, such as a hard disk, optical read and / or write storage medium, RAM, ROM, etc., such that it can be read by a programmable computer, and when the storage medium or device is read by the computer, it can be used to configure and operate the computer to perform the processes described herein.

[0138] Furthermore, machine-readable code, or parts thereof, can be transmitted via wired or wireless networks.

[0139] When such media includes instructions or programs that combine with a microprocessor or other data processor to implement the steps described above, the invention described herein includes these and other different types of non-transitory computer-readable storage media.

[0140] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for resolving distributed sequential logic conflicts in a distributed control system, characterized in that: include: The system acquires a first conflict instruction and a second conflict instruction issued by the first controller and the second controller respectively when there is a timing conflict; the first conflict instruction includes a first feature pair consisting of a first local clock stamp and a first physical feature parameter, and a first forwarding trace set; the second conflict instruction includes a second feature pair consisting of a second local clock stamp and a second physical feature parameter, and a second forwarding trace set. The first and second dwell times of the first and second conflict instructions in network transmission are calculated based on the first and second forwarding trace sets. The first local clock stamp and the second local clock stamp are corrected using the first retention time and the second retention time, respectively, to obtain the first corrected clock and the second corrected clock; Input the first physical feature parameter and the second physical feature parameter into the physical constraint model to obtain the first theoretical time required to transform the first physical feature parameter into the second physical feature parameter, and the second theoretical time required to transform the second physical feature parameter into the first physical feature parameter; Calculate the clock difference between the first corrected clock and the second corrected clock, match the clock difference with the first theoretical duration and the second theoretical duration, and select the reserved instruction as the conflict resolution result output according to the causal order.

2. The distributed sequential logic conflict resolution method for distributed control systems as described in claim 1, characterized in that: When the first conflict instruction passes through the first network node, the first network node collects its current first network node load data, current first network node local timestamp, and first network node identifier, and adds the three together as the first network node forwarding trace to the first conflict instruction to generate the first intermediate instruction. When the first intermediate instruction passes through the second network node, the second network node collects its current second network node load data, current second network node local timestamp, and second network node identifier, and combines the three as a second network node forwarding trace to be added to the first intermediate instruction, generating a first conflict instruction carrying a first feature pair and a first forwarding trace set consisting of the first network node forwarding trace and the second network node forwarding trace; The conflict resolver receives the first conflict instruction, records the local reception timestamp, and extracts the first feature pair and the first forwarding trace set from the first conflict instruction.

3. The distributed sequential logic conflict resolution method for distributed control systems as described in claim 2, characterized in that: The second conflict command issued by the second controller is executed with steps symmetrical to the first conflict command to obtain a second feature pair, a second forwarding trace set, and the corresponding local reception timestamp.

4. The distributed sequential logic conflict resolution method for distributed control systems as described in claim 3, characterized in that: The conflict resolver parses the first forwarding trace set and extracts the first network node load sequence and the corresponding first network node identifier sequence arranged in forwarding order; similarly, it parses the second forwarding trace set to obtain the second network node load sequence and the corresponding second network node identifier sequence. Read the service rate parameters corresponding to each network node from the system configuration database, as well as the sum of the transmission delays of each network link, as the base transmission delay; Based on the first network node load sequence and the corresponding first network node identifier sequence, the queuing delay of the first conflict instruction at each forwarding network node is determined using the service rate parameters of each network node, and all queuing delays are accumulated to obtain the first dwell time. Similarly, the second dwell time is determined based on the second network node load sequence and the corresponding second network node identifier sequence.

5. The distributed sequential logic conflict resolution method for distributed control systems as described in claim 4, characterized in that: Add the first dwell time to the basic transmission delay to obtain the first total transmission time; The first corrected clock is obtained by subtracting the first total transmission duration from the first local reception timestamp recorded by the conflict resolver. Similarly, the second dwell time is added to the basic transmission delay to obtain the second total transmission time, and the second local reception timestamp is subtracted from the second total transmission time to obtain the second corrected clock.

6. The distributed sequential logic conflict resolution method for distributed control systems as described in claim 5, characterized in that: Based on the first local clock stamp in the first conflict instruction and the second local clock stamp in the second conflict instruction, the historical operating condition data of the target equipment at the corresponding time are extracted from the historical database to form the first operating condition parameter set and the second operating condition parameter set. If there is no data in the historical database that corresponds to the exact time, then linear interpolation or nearest neighbor interpolation methods are used to obtain it; The current working condition parameter set is obtained by taking a weighted average of the first working condition parameter set and the second working condition parameter set. The first physical characteristic parameter is used as the initial state value, the second physical characteristic parameter is used as the target state value, and the current operating condition parameter set is input into the equipment state transition time calculation model. The equipment state transition time calculation model predicts the equipment dynamic characteristic parameters at the current moment online based on the current operating condition parameter set, and determines the first theoretical time required for the change from the initial state value to the target state value based on the equipment dynamic characteristic parameters; The initial state value and the target state value are exchanged. The second physical characteristic parameter is used as the new initial state value, and the first physical characteristic parameter is used as the new target state value. The current operating condition parameter set remains unchanged and is input into the equipment state transition duration calculation model to obtain the second theoretical duration.

7. The distributed sequential logic conflict resolution method for distributed control systems as described in claim 6, characterized in that: The equipment state transition time calculation model is a first-order inertial plus pure time delay model. The model parameters are corrected online according to the current operating condition parameter set, and the theoretical time required to change from the initial state value to the target state value is calculated based on the step response inverse calculation principle.

8. The distributed sequential logic conflict resolution method for distributed control systems as described in claim 7, characterized in that: The conflict resolver calculates the clock difference between the first corrected clock and the second corrected clock; Obtain network latency fluctuation data from the most recent n historical conflict resolution cases, and determine the network latency fluctuation range based on the statistical distribution of the fluctuation data; obtain the prediction deviation data of the device state transition duration calculation model in historical conflict resolution cases, and determine the model prediction error range based on the statistical distribution of the prediction deviation data; By combining the network latency fluctuation range and the model prediction error range, the current uncertainty tolerance range is obtained; Determine whether the clock difference falls within the uncertainty tolerance range based on the first theoretical duration. If so, determine that the first theoretical duration matches. Determine whether the clock difference falls within the uncertainty tolerance range based on the second theoretical duration. If so, determine that the second theoretical duration matches.

9. The distributed sequential logic conflict resolution method for a distributed control system as described in claim 8, characterized in that: Determine the causal order based on the matching results: If only the first theoretical duration matches, then the first conflicting instruction is determined to be the cause and the second conflicting instruction to be the effect; if only the second theoretical duration matches, then the second conflicting instruction is determined to be the cause and the first conflicting instruction to be the effect; if both the first and second theoretical durations match or neither matches, then proceed to dynamic criticality assessment. The retained instructions selected based on causal order determination or dynamic criticality assessment will be output to the actuators of the distributed control system as the conflict resolution result.

10. The distributed sequential logic conflict resolution method for a distributed control system as described in claim 9, characterized in that: The dynamic criticality assessment includes: Acquire the current real-time operating condition data of the target device, wherein the real-time operating condition data includes at least load rate, medium temperature and wear coefficient; The criticality of the first conflict command and the second conflict command are assessed based on the real-time operating data. The criticality level is determined by analyzing the degree of deviation of the first or second physical characteristic parameter involved in the instruction from the safety boundary under the current operating conditions, the urgency of the action represented by the instruction, and the sensitivity of the equipment's current health status to the execution of the instruction. The first and second conflicting instructions are compared in the following order: degree of deviation, urgency, and sensitivity of response. The instruction with higher priority is selected as the reserved instruction. If all indicators are the same, one of them is randomly selected as the reserved instruction.

Images