CN-122006196-A - Intelligent operation and maintenance management control system for fire-fighting equipment full life cycle

CN122006196ACN 122006196 ACN122006196 ACN 122006196ACN-122006196-A

Abstract

The invention belongs to the technical field of fire-fighting equipment control and management, and discloses a fire-fighting equipment full life cycle intelligent operation and maintenance management control system; the method comprises the steps of acquiring real-time on-site state data, executing active excitation and response tests under a non-fire state to acquire dynamic response data, adopting an end-to-end multi-task learning framework to process the on-site state data and the dynamic response data, outputting a structural diagnosis result comprising a pump group, a valve and a pipe network health state, establishing an initial health base line of each key device, dynamically generating an adaptive alarm threshold value, generating a quantized risk level in combination with a fire safety level to generate an operation and maintenance strategy instruction comprising test operation and maintenance parameters, automatically executing device active-standby switching and robust health detection, preferentially guaranteeing a fire protection function when receiving a fire alarm or manual forced instruction, and executing graded alarm according to the quantized risk level. The intelligent operation and maintenance management control of the fire-fighting equipment full life cycle is realized.

Inventors

Ge Xiaocen
GE SHUHAN
GE YANBIN
GE YANPING

Assignees

吉林省鸿星建设集团有限公司

Dates

Publication Date: 20260512
Application Date: 20260105

Claims (10)

1. The utility model provides a fire control facility full life cycle intelligence fortune dimension management control system which characterized in that includes: The system comprises a data acquisition and active response test module, a dynamic response data acquisition module and a dynamic response data acquisition module, wherein the data acquisition and active response test module is used for carrying out multi-source sensing acquisition on the running states of a fire-fighting water pump, a motor, a pipe network and a valve actuator to acquire real-time on-site state data; The health diagnosis and hidden fault identification module adopts an end-to-end multitask learning framework to process the on-site state data and the dynamic response data and output a structural diagnosis result containing the health states of the pump group, the valve and the pipe network; the full life cycle management module is used for establishing an initial health baseline of each key device, constructing a health index history sequence based on a structural diagnosis result, fitting a device aging track function and dynamically generating a self-adaptive alarm threshold; the risk assessment and strategy generation module is used for generating a quantized risk level in combination with the fire safety level and generating an operation and maintenance strategy instruction containing test operation and maintenance parameters; The intelligent coordinated control execution module is used for converting operation and maintenance strategy instructions into control instructions, automatically executing equipment active-standby switching and healthy detection, and preferentially guaranteeing a fire protection function when receiving a fire alarm or manual forced instruction; And the self-adaptive optimization and alarm module is used for self-adaptively adjusting a data acquisition strategy according to the equipment state stability and executing hierarchical alarm according to the quantized risk level.
2. The intelligent life cycle operation and maintenance management control system of a fire-fighting equipment according to claim 1, wherein the multi-source sensing and acquisition of the operation states of the fire-fighting water pump, the motor, the pipe network and the valve actuator are performed, and the method comprises: a multipoint acceleration sensor is arranged on a bearing seat of the fire-fighting water pump and a pump body shell, and vibration signals of the bearing and the pump body are collected; collecting three-phase current, voltage and active power waveforms of a fire pump motor through an electric parameter sensor; Arranging a pressure sensor and a flowmeter at key nodes of a fire-fighting pipe network, and collecting pressure and flow data of the pipe network; And acquiring a travel switch state, an encoder position signal and an execution torque curve of the electric valve actuator through the position and torque sensor.
3. The fire protection facility full life cycle intelligent operation and maintenance management control system of claim 2, wherein the method of performing active incentive and response tests comprises: judging whether the pressure of the main fire-fighting pipeline is not lower than a preset lower limit, if so, The bypass water drain valve is controlled to be opened by regulating the rotation speed of the fire pump in a short time or controlling the bypass water drain valve in a short time, and controlled flow or pressure disturbance is injected into the fire pipeline; executing small-stroke reciprocating opening and closing actions of a preset angle or applying small-stroke torque loading to the key valve; during the active excitation execution, multi-source sensing acquisition is carried out to acquire dynamic response data; stopping executing the active excitation and response test when the pressure of the main fire-fighting pipeline is monitored to be lower than the preset lower limit; And the real-time on-site state data refers to data acquired through multi-source sensing acquisition during full life cycle management of the fire protection facility.
4. The fire protection facility full life cycle intelligent operation and maintenance management control system of claim 3, wherein the end-to-end multi-task learning framework comprises a shared input layer for receiving field state data and dynamic response data, 1 shared encoder, and 3 dedicated task decoders comprising a pump group health head, a valve identification head, and a pipe network diagnostic head, wherein, The pump set health head is designed based on a multi-layer perceptron and comprises an input layer, a hidden layer and an output layer, wherein the input layer is used for receiving the output of a shared encoder, the number of neurons of each layer is gradually decreased from 2 to 4 fully connected layers, the hidden layer is activated by using a ReLU function, the output layer is used for outputting a bearing wear grade and a pump body corrosion degree, the predefined bearing wear grade is N, the number of neurons of the output layer is N+1, a bearing wear grade task is activated by using a Softmax function, the probability of each grade is output, the pump body corrosion degree task is activated by using a Sigmoid function, and a numerical value between 0 and 1 is output; The valve identification head is designed based on a comparison learning structure and comprises a double-branch twin network, wherein each branch is a 3-4-layer multi-layer perceptron, the two branches share weight, the input of the first branch is the stroke feedback signal characteristic of a valve, the input of the second branch is the pressure-flow response characteristic of a pipe network in the same time period, the two sub-networks respectively output vectors with the same dimension, and then, the distance value between the two branch vectors is calculated and is converted into an inconsistency confidence coefficient with the value range of 0-1 through a full connection layer and a Sigmoid function; The pipe network diagnosis head is designed based on a graph neural network decoder, firstly, a fire pipe network is abstracted into a graph, each node represents the physical position of a pipe point, node characteristics comprise static properties and dynamic data, edges represent pipe connection, the edge characteristics comprise pipe scalar values, a 2-4-layer graph rolling network is deployed, the layer number design is related to the pipe network topology diameter, after the graph rolling, each node obtains a feature vector containing global topology information, at the moment, a full connection layer is connected for each node, a scalar value is output for each node, the node is converted into the fault confidence of the node after being processed by a Sigmoid function, a probability list is output, the list length is equal to the node number, and each value in the list corresponds to the fault confidence of one node; The multi-task learning framework generates a structured diagnosis result comprising a pump group health state evaluation result, a valve consistency recognition result and a pipe network node fault confidence list by collecting the output of the three task decoders.
5. The fire protection facility full life cycle intelligent operation and maintenance management control system of claim 4, wherein the method for dynamically generating the adaptive alarm threshold comprises: extracting quantitative characteristics of health states of each key device from the structural diagnosis result to serve as a core health index; At the initial stage of operation of the fire-fighting equipment, core health indexes under multiple working conditions are collected, and initial health baseline vectors of all key equipment are established; In the running process of the equipment, core health indexes are recorded according to time sequence to form a health index history sequence; Fitting an aging trend of the health state of the equipment along with the change of the running time through regression analysis based on the health index history sequence to obtain an equipment aging track function; Stripping reversible working condition fluctuation from the real-time health index, and outputting corrected health index reflecting the inherent performance attenuation of the equipment; based on the equipment aging track function, the corresponding self-adaptive alarm threshold value is dynamically calculated and updated for each core health index.
6. The intelligent life cycle operation and maintenance management control system of a fire protection facility according to claim 5, wherein the method for stripping reversible condition fluctuations from real-time health indicators and outputting corrected health indicators reflecting equipment inherent performance attenuation comprises: establishing a working condition influence lookup table taking seasons, water sources and environmental humiture as input and typical deviation of health indexes as output; And searching the offset of the current working condition through the working condition influence lookup table, correcting the real-time core health index, and recording the corrected health index after stripping the reversible working condition fluctuation.
7. The fire protection system full life cycle intelligent operation and maintenance management control system of claim 6, wherein the method for generating the operation and maintenance policy instructions including the commissioning and maintenance parameters in combination with the fire protection security level generates a quantized risk level, comprises: Obtaining an aging speed represented by an equipment aging track function; Comparing the corrected health index with the self-adaptive alarm threshold value in combination with the current fire safety level requirement, and evaluating the aging speed, thereby quantifying the failure risk of the single equipment and the whole fire protection system and outputting quantified risk level; According to the quantized risk level, inquiring a pre-stored operation and maintenance strategy library, and generating a corresponding operation and maintenance strategy instruction, wherein the operation and maintenance strategy instruction at least comprises a test operation configuration parameter for the fire water pump and an overhaul suggestion parameter for the equipment with the identified hidden fault.
8. The fire protection facility full life cycle intelligent operation and maintenance management control system of claim 7, wherein the method of converting operation and maintenance policy instructions into control instructions comprises: When the corrected health index is lower than the corresponding self-adaptive alarm threshold or the quantized risk level reaches the preset high-risk level, the main equipment and the standby equipment are automatically switched, and the main equipment after switching is subjected to enhanced health detection; when a fire alarm signal or a manual forced pump start instruction is received, all active excitation tests are immediately stopped.
9. The intelligent operation and maintenance management control system for the full life cycle of the fire-fighting equipment according to claim 8, wherein the method for adaptively adjusting the data acquisition strategy according to the equipment state stability comprises the steps of judging that the state of the fire-fighting equipment is stable when the variation amplitude of the corresponding monitoring value of the core health index of a plurality of continuous cycles is smaller than a first preset threshold value, and the data acquisition is operated with a low sampling rate; When the amplitude or current value of the vibration signal acquired in real time exceeds a second preset threshold value, the abnormal sign is judged to appear, the sampling rate is automatically increased, and the active excitation test is triggered.
10. The fire protection facility full life cycle intelligent operation and maintenance management control system of claim 9, wherein the method of performing the hierarchical alarm according to the quantized risk level comprises: pushing the low-level risk to an operation and maintenance work order, informing an operator on duty of the medium-level risk through a mobile message, triggering a fire alarm signal by the high-level risk, and uploading the fire alarm signal to the upper-level management platform.

Description

Intelligent operation and maintenance management control system for fire-fighting equipment full life cycle Technical Field The invention relates to the technical field of fire-fighting equipment control and management, in particular to a full life cycle intelligent operation and maintenance management control system of a fire-fighting equipment. Background In the field of modern industry, in particular in process plants with continuous production and warehouses storing large amounts of materials, fire protection facilities as final safety barriers are directly related to the security risk of personnel life and property. The core requirement of such environments for fire protection facilities is constantly, i.e. during their full life cycle, in an immediately available state of health at all times. Failure of any facility, particularly in the event of a fire emergency, can lead to catastrophic results. The traditional operation and maintenance management mode mainly relies on periodic manual inspection and planned maintenance. However, manual inspection is a discrete point-in-time behavior that does not allow for continuous, real-time perception of the device state. During the interval between two rounds, degradation of the equipment status cannot be found in time, so that the facility is in fact in an unknown risk dead zone for a long period of time. Therefore, intelligent detection is a trend of operation and maintenance management of the existing fire-fighting equipment, but the parameters monitored by the existing system are mostly operation result parameters (such as whether the pressure at the tail end of a pipe network reaches the standard or not), but not health state parameters of the fire-fighting equipment (such as pump shaft abrasion vibration characteristics, motor winding insulation aging degree and valve actuator torque change curve), so that hidden faults such as corrosion inside a pump body of a fire-fighting water pump, false opening or false closing of a key electric or pneumatic valve (i.e. an actuator feedback signal is normal but a valve core is not actually in place), local blockage inside the pipe network, slow aging of a sealing element and the like cannot be found through appearance inspection or simple function start-stop test. The system is hidden in the sun, but the sudden shutdown or the loss of functions of the fire-fighting equipment is directly caused when the system is started in an emergency, so that the whole fire-fighting system is similar to a dummy fire-fighting system, and therefore, the system can realize real-time on-line monitoring and early warning on hidden faults such as pump body corrosion, valve false opening and the like by utilizing advanced technology in the field of automatic control, and the system is a unique technical problem which is urgent to be solved in the field but is not solved well. In view of the above description, the design provides a fire-fighting equipment full life cycle intelligent operation and maintenance management control system. Disclosure of Invention In order to overcome the defects in the prior art and achieve the purposes, the invention provides the following technical scheme that the full life cycle intelligent operation and maintenance management control system of the fire-fighting facility comprises: The system comprises a data acquisition and active response test module, a dynamic response data acquisition module and a dynamic response data acquisition module, wherein the data acquisition and active response test module is used for carrying out multi-source sensing acquisition on the running states of a fire-fighting water pump, a motor, a pipe network and a valve actuator to acquire real-time on-site state data; The health diagnosis and hidden fault identification module adopts an end-to-end multitask learning framework to process the on-site state data and the dynamic response data and output a structural diagnosis result containing the health states of the pump group, the valve and the pipe network; the full life cycle management module is used for establishing an initial health baseline of each key device, constructing a health index history sequence based on a structural diagnosis result, fitting a device aging track function and dynamically generating a self-adaptive alarm threshold; the risk assessment and strategy generation module is used for generating a quantized risk level in combination with the fire safety level and generating an operation and maintenance strategy instruction containing test operation and maintenance parameters; The intelligent coordinated control execution module is used for converting operation and maintenance strategy instructions into control instructions, automatically executing equipment active-standby switching and healthy detection, and preferentially guaranteeing a fire protection function when receiving a fire alarm or manual forced instruction; And the self-adaptive optimization and a