WO2023088355A1 - 一种传感器智能数据重构方法及系统 - Google Patents

一种传感器智能数据重构方法及系统 Download PDF

Info

Publication number
WO2023088355A1
WO2023088355A1 PCT/CN2022/132511 CN2022132511W WO2023088355A1 WO 2023088355 A1 WO2023088355 A1 WO 2023088355A1 CN 2022132511 W CN2022132511 W CN 2022132511W WO 2023088355 A1 WO2023088355 A1 WO 2023088355A1
Authority
WO
WIPO (PCT)
Prior art keywords
sensor
data reconstruction
model
data
reconstruction
Prior art date
Application number
PCT/CN2022/132511
Other languages
English (en)
French (fr)
Inventor
堵树宏
徐钊
于方小稚
苗壮
马心童
张敏
楚济如
洪郡滢
Original Assignee
中国核电工程有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中国核电工程有限公司 filed Critical 中国核电工程有限公司
Publication of WO2023088355A1 publication Critical patent/WO2023088355A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G21NUCLEAR PHYSICS; NUCLEAR ENGINEERING
    • G21CNUCLEAR REACTORS
    • G21C17/00Monitoring; Testing ; Maintaining
    • GPHYSICS
    • G21NUCLEAR PHYSICS; NUCLEAR ENGINEERING
    • G21DNUCLEAR POWER PLANT
    • G21D3/00Control of nuclear power plant
    • GPHYSICS
    • G21NUCLEAR PHYSICS; NUCLEAR ENGINEERING
    • G21DNUCLEAR POWER PLANT
    • G21D3/00Control of nuclear power plant
    • G21D3/04Safety arrangements
    • G21D3/06Safety arrangements responsive to faults within the plant
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02EREDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E30/00Energy generation of nuclear origin
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02EREDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E30/00Energy generation of nuclear origin
    • Y02E30/30Nuclear fission reactors

Definitions

  • the disclosure belongs to but not limited to the field of data processing, and in particular, relates to a method and system for reconstructing intelligent sensor data.
  • the safe and reliable operation of nuclear power units is inseparable from the complete function of the instrument and control system, and the integrity of the sensor function is the basis for the instrument and control system to monitor the nuclear power plant's operating status and trigger automatic actions.
  • three technical means are mainly adopted: "higher identification requirements”, “hardware redundant configuration” and “periodic tests or condition-based maintenance during operation and maintenance”.
  • sensors are divided into two categories: “safety level” and “non-safety level”.
  • safety level sensors higher quality assurance, earthquake resistance and identification requirements are put forward.
  • the method of redundant configuration of sensor hardware is generally adopted to further improve the reliability of the data acquisition function of the same measuring point.
  • nuclear power plants In the operation and maintenance phase, in view of possible slow faults on the sensors, nuclear power plants use regular tests and condition-based maintenance to check whether the operating status of the sensors is normal, so as to identify minor faults as early as possible, and improve the functions of the sensors based on preventive ideas completeness.
  • the purpose of the present disclosure is to provide a sensor intelligent data reconstruction method and system, which can quickly and actively identify the faulty sensor after a fast fault occurs in the sensor, and automatically generate The data replaces the value of the fault sensor to ensure that the unit is in a safe and stable state, and buy time for the operator's operation control operation and the maintenance personnel's on-site maintenance and replacement work.
  • the technical solution adopted according to the embodiments of the present disclosure is to provide a sensor intelligent data reconstruction method, including the following steps:
  • step S1 Judging whether the operation condition of the data reconstruction model is satisfied, if the operation condition of the data reconstruction model is satisfied, then enter step S2, otherwise, the data reconstruction step is terminated;
  • step S2. Perform sensor state detection and judge whether the detection result is abnormal. If the detection result is abnormal, enter step S3. If the detection result is normal, the data reconstruction step is terminated;
  • step S3 Input the real-time measurement data of the sensor with abnormal detection results obtained in step S2 into the data reconstruction model to obtain a reconstruction value.
  • step S1 includes the following sub-steps:
  • step S101 Determine whether the nuclear power of the reactor is higher than 20%*Pn, where Pn is the rated active power output by the reactor. If the nuclear power of the reactor is higher than 20%*Pn, go to step S102; otherwise, the data reconstruction step is terminated;
  • step S102 Judging whether the number of virtual sensors in the input set of the sensor X model in the sensor set K that needs data reconstruction meets the requirements, if it meets the requirements, go to step S103; if it does not meet the requirements, the data reconstruction step is terminated, delete sensor X from said sensor set K;
  • step S102 includes: judging whether the number of virtual sensors among the input sensors of the sensor X model in the sensor set K reaches a certain proportion k of the total number of input sensor sets, and if the number of virtual sensors reaches a certain proportion of the total number of input sensor sets , then the data reconstruction step is terminated, and the sensor X is deleted from the sensor set K; if the number of virtual sensors does not reach the ratio k, the sensor X is added to the entity sensor set N and the entity is further judged Whether the number of physical sensors in the sensor set N is greater than 0, if the number of sensors in the physical sensor set N is greater than 0, it is determined that the number of virtual sensors in the sensor set K meets the requirements.
  • step S2 includes the following sub-steps:
  • S202 Input the real-time measurement data into the state monitoring model under the working condition to monitor the operating state. If the operating state is abnormal, add the sensor to the abnormal sensor set L and continue the subsequent data reconstruction step. If If the running state is normal, the sensor data reconstruction step is terminated.
  • step S3 includes the following sub-steps:
  • step S302 is a Monte Carlo method.
  • a sensor intelligent data reconstruction system includes a model operation condition judgment module, a sensor state detection module and a data reconstruction module,
  • the model operation condition judging module is configured to judge whether the data reconstruction model operation condition is met, and if the data reconstruction model operation condition is met, the subsequent data reconstruction is continued, otherwise the data reconstruction is suspended;
  • the sensor state detection module is configured to perform sensor state detection, and determine whether the detection result is abnormal, if the sensor state detection shows that the detection result is abnormal, then continue the subsequent data reconstruction, otherwise the data reconstruction is suspended;
  • the data reconstruction module is configured to input the real-time measurement data of the sensor into a data reconstruction model to obtain a reconstructed value.
  • model commissioning condition judgment module includes a power judgment unit, a virtual sensor quantity judgment unit, and a start request sending unit,
  • the power judging unit is configured to judge whether the nuclear power of the reactor is above 20% Pn, if the nuclear power of the reactor is above 20% Pn, continue the subsequent data reconstruction; otherwise, the data reconstruction is suspended;
  • the virtual sensor quantity judging unit is configured to judge whether the number of virtual sensors in the input set of the sensor X model in the sensor set K that needs data reconstruction meets the requirements, and if it meets the requirements, continue the subsequent data reconstruction; otherwise Data reconstruction aborted;
  • the start request sending unit is configured to push a "start request" to the operator, and after the operator agrees to start, subsequent data reconstruction is started.
  • the sensor state detection module includes a working condition classification unit and an operating state judgment unit,
  • the working condition classification unit is configured to input the real-time measurement data of each sensor in the sensor set K into the working condition classification model in recent several sampling time periods to classify the working conditions, and match the sensors in the working condition
  • the sensor state monitoring model under the condition
  • the operating state judging unit is configured to input the real-time measurement data into the state monitoring model under the working conditions to monitor the operating state, and if the operating state is abnormal, add the sensor to the abnormal sensor set L, Continue subsequent data reconstruction, if the running status is normal, then the sensor data reconstruction is terminated.
  • the data reconstruction module includes a data reconstruction unit and a reconstructed signal verification unit,
  • the data reconstruction unit is configured to substitute the real-time measurement data into the data reconstruction model corresponding to the working condition type, so as to obtain the reconstruction value output by the data reconstruction model;
  • the reconstructed signal verification unit is configured to use an uncertainty analysis method to calculate the uncertainty bandwidth of the reconstructed value, and if the uncertainty bandwidth of the reconstructed value does not exceed a fixed value, the reconstructed If the reconstructed value is valid, the reconstructed value is sent back to the instrument and control system to temporarily replace the abnormal sensor.
  • the sensor intelligent data reconstruction method and system disclosed in the embodiments of the present disclosure can quickly and actively identify the faulty sensor after a fast fault occurs in the sensor, and automatically generate data when relevant conditions are met. Replace the value of the fault sensor to ensure that the unit is in a safe and stable state, and buy time for the operator's operation control operation and maintenance personnel's on-site maintenance and replacement work. It can replace the traditional "periodical test", and uniform and targeted maintenance when the shutdown is planned, thereby reducing the operation and maintenance cost. It can also reduce the number of redundant sensor configurations, thereby reducing the construction cost of the unit.
  • FIG. 1 is a schematic diagram of a sensor reconstruction model in a sensor intelligent data reconstruction method according to an embodiment of the present disclosure
  • FIG. 2 is a flow chart of a method for reconstructing intelligent sensor data according to an embodiment of the present disclosure
  • Fig. 3 is a flow chart of step S1 "judging whether the data reconstruction model operation condition is met" in a sensor intelligent data reconstruction method described in an embodiment of the present disclosure
  • Fig. 4 is a flow chart of step S2 "performing sensor state detection and judging whether the detection result is abnormal" in a method for reconstructing intelligent sensor data according to an embodiment of the present disclosure
  • Fig. 5 is a flow chart of step S3 "performing sensor data reconstruction" in a sensor intelligent data reconstruction method according to an embodiment of the present disclosure.
  • the set K is defined as a set of sensors (hereinafter referred to as “reconstructed sensors”) that need to be reconstructed.
  • each data reconstruction sensor has a corresponding data reconstruction model.
  • the data reconstruction model is trained based on the Auto-Association Kernel Regression Algorithm (AAKR).
  • AAKR Auto-Association Kernel Regression Algorithm
  • a method for reconstructing intelligent sensor data includes the following steps:
  • step S1 Determine whether the data reconstruction model operation conditions are met. If the data reconstruction model operation conditions are met, then enter step S2. Otherwise, the data reconstruction step is terminated and wait for the next judgment cycle to continue judging whether the data reconstruction model is suitable for operation. condition.
  • step S2 Perform sensor state detection and judge whether the detection result is abnormal. If the detection result is abnormal, go to step S3. If the detection result is normal, the data reconstruction step is stopped, and the sensor state detection continues in the next judgment cycle.
  • Step S1 specifically includes the following sub-steps:
  • step S101 Judging whether the nuclear power of the reactor is above 20% Pn, if the nuclear power of the reactor is above 20% Pn, then go to step S102; otherwise, the data reconstruction step is terminated, waiting for the next judgment cycle to continue judging whether the nuclear power of the reactor is 20% %Pn above.
  • Pn is the rated active power output by the reactor, and if the nuclear power of the reactor is above 20% of Pn, it means that the state of the nuclear reactor is relatively stable, and the data reconstruction model can be started.
  • the set K is initialized. If the data reconstruction model is in the starting state, the data reconstruction step will be terminated, and the information of "the model stops running" will be pushed to the operator for notification.
  • step S102 Determine whether the number of virtual sensors in the sensor set K that needs data reconstruction meets the requirements, and if so, go to step S103; if not, stop the data reconstruction step.
  • Judging whether the number of virtual sensors in the sensor set K that needs data reconstruction meets the requirements includes: judging whether the number of virtual sensors in the input sensors of the sensor X model in the sensor set K that needs data reconstruction reaches the input sensor set a certain percentage of the total. If the number of virtual sensors reaches a certain proportion of the total number of input sensor sets, reconstruction cannot be realized, and sensor X is deleted from set K. If the number of virtual sensors does not reach a certain ratio, add the physical sensor set N. If the number of sensors in the physical sensor set N is greater than 0, it is determined that the number of virtual sensors in the sensor set K that needs data reconstruction meets the requirements.
  • the information will be pushed to the operator in the form of an alarm, and the information includes: "Sensor X reconfiguration step failure" and "Estimated consequences after the failure of the sensor signal”.
  • S103 Push a "startup request" to the operator, the content of the "startup request” includes: a startup request and a list of sensors that meet the prerequisites for monitoring and reconstruction. After the operator agrees to start, start the subsequent steps of data reconstruction.
  • step S2 includes the following sub-steps:
  • S201 Input the data collected by the sensor X in several recent sampling time periods (for example, 60s) into the corresponding working condition classification model, classify the working conditions, obtain the working condition type of the real-time measurement sample, and select the sensor under this working condition sensor condition monitoring model.
  • the working condition classification model is trained based on the density clustering algorithm, and the working condition we need is obtained by bringing the sensor measurement data of normal operation under a certain working condition into the working condition classification model for model training classification model.
  • the working condition classification model includes a power-up model and a power-down model. Classifying the working conditions of sensors is beneficial to understand the signal change trend of sensors and improve the accuracy of data reconstruction.
  • S202 Input the real-time measurement data of all sensors in the sensor set K that needs to be reconstructed into the condition monitoring model under the above working conditions, and output the operating status of the sensors. If the operating state is abnormal, add the sensor to the abnormal sensor set L, and continue the subsequent data reconstruction step; if the operating state is normal, stop the sensor data reconstruction step.
  • condition monitoring model is trained based on the auto-associative neural network, and the condition monitoring model we need is obtained by bringing the sensor measurement data under a certain working condition into the condition monitoring model for model training in advance.
  • Step S3 includes the following sub-steps:
  • sensor data reconstruction that is: for the real-time measurement data collected by sensor X within a certain period of time according to its working condition type, and select the data reconstruction model of the sensor under this working condition; The collected real-time measurement data is substituted into the data reconstruction model of the working condition type to obtain the output reconstruction value.
  • the model reconstruction signal is verified and sent back, that is, the uncertainty bandwidth of the sensor reconstruction value is calculated by using an uncertainty analysis method (such as Monte Carlo method, etc.). If the uncertainty bandwidth of the sensor reconstructed value does not exceed the fixed value, the reconstructed value is valid, and the value is sent back to the instrument and control system to temporarily replace the abnormal sensor.
  • an uncertainty analysis method such as Monte Carlo method, etc.
  • the information is pushed to the operator in the form of an alarm
  • the information includes "the sensor X is abnormal”, “the sensor X has been reconfigured and sent back", “the reconfigured value of the sensor X may be invalid after 4 hours” and the The expected consequences of failure of the reconstructed signal from the sensor. If the uncertainty bandwidth exceeds the fixed value, the reconstruction of the sensor will be stopped, and the information will be pushed to the operator in the form of an alarm.
  • the information includes "abnormal sensor X", “failure of sensor X reconstruction” and the expected Loss of function timing and consequences.
  • the operator starts to judge whether the operation conditions of the data reconstruction model are met, and the data reconstruction model performs the following sub-steps every 60s (this period may be extended or shortened according to the actual application scenario requirements):
  • the model "start-up requirement" is met if the current power is determined to be above 20% of the full power of the rig loop.
  • the measurement data of each sensor for 60s is substituted into the respective working condition classification model, and all are detected as the power-up working condition.
  • the reconstructed value of the outlet temperature of the heating section of the primary circuit is sent back to the corresponding sensor, and the information is pushed to the operator in the form of an alarm.
  • the information includes "the temperature sensor at the outlet of the heating section of the primary circuit is abnormal", " The outlet temperature sensor has been reconfigured and sent back” and "The reconfigured value of the outlet temperature sensor of the heating section of the primary circuit may fail after 4 hours.”
  • a sensor intelligent data reconstruction method provided by the embodiments of the present disclosure, it can quickly and actively identify the faulty sensor after a fast fault occurs in the sensor, and automatically generate data when the relevant conditions are met. Replace the value of the fault sensor to ensure that the unit is in a safe and stable state, and buy time for the operator's operation control operation and maintenance personnel's on-site maintenance and replacement work. It can replace the traditional "periodical test", and uniform and targeted maintenance when the shutdown is planned, thereby reducing the operation and maintenance cost. It can also reduce the number of redundant sensor configurations, thereby reducing the construction cost of the unit.
  • a sensor intelligent data reconstruction system includes a model operation condition judgment module, a sensor state detection module, and a data reconstruction module.
  • the model operation condition judgment module is configured to judge whether the data reconstruction model operation condition is met. If the data reconstruction model operation condition is met, the subsequent data reconstruction will continue. Otherwise, the data reconstruction will be suspended and wait for the next judgment cycle to continue. Determine whether the data reconstruction model operation conditions are met.
  • the sensor state detection module is configured to detect the sensor state and determine whether the detection result is abnormal. If the sensor state detection shows that the detection result is abnormal, continue the subsequent data reconstruction. If the sensor state detection shows that the detection result is normal, the data reconstruction is suspended. , and wait for the next judgment cycle to continue sensor status detection.
  • the data reconstruction module is configured to input the data measured by the sensor in real time into the data reconstruction model to obtain the reconstructed value.
  • the module for judging the model operation conditions includes a power judging unit, a virtual sensor quantity judging unit and a start request sending unit.
  • the power judging unit is configured to judge whether the nuclear power of the reactor is above 20% Pn, if the nuclear power of the reactor is above 20% Pn, then proceed to the subsequent data reconstruction step; otherwise, the data reconstruction is suspended, and the next judging period is waited for to continue judging Whether the nuclear power of the reactor is above 20% Pn.
  • the set K is initialized. If the reconstructed model is in the starting state, the reconstructed model will automatically stop running, and the information of "model stopped running” will be pushed to the operator to know.
  • the judging unit for the number of virtual sensors is configured to judge whether the number of virtual sensors in the model input set of sensors in the set K meets the requirements, and if it meets the requirements, continue the subsequent data reconstruction; if it does not meet the requirements, the data reconstruction is suspended.
  • Judging whether the number of virtual sensors in the sensor set K that needs data reconstruction meets the requirements includes: judging whether the number of virtual sensors in the input sensors of the sensor X model in the sensor set K that needs data reconstruction reaches the input sensor set a certain percentage of the total. If the number of virtual sensors reaches a certain proportion of the total number of input sensor sets, reconstruction cannot be realized, and sensor X is deleted from set K. If the number of virtual sensors does not reach a certain ratio, add the physical sensor set N. If the number of sensors in the physical sensor set N is greater than 0, it is determined that the number of virtual sensors in the sensor set K that needs data reconstruction meets the requirements.
  • the information will be pushed to the operator in the form of an alarm, and the information includes: "Sensor X reconfiguration step failure" and "Estimated consequences after the failure of the sensor signal”.
  • the start request sending unit is configured to push a "start request" to the operator, and the content of the "start request” includes: a start request and a list of sensors that meet the prerequisites for monitoring and reconstruction. After the operator agrees to start, start the data reconstruction model.
  • the sensor state detection module includes a working condition classification unit and a running state judgment unit.
  • the working condition classification unit is configured to input the data collected by the sensor X in several recent sampling time periods (for example, 60s) into the working condition classification model to perform working condition classification, obtain the working condition type to which the real-time measurement sample belongs, and select the sensor in the The sensor state monitoring model under this working condition.
  • the running state judging unit is configured to input all the data collected by the sensors in the sensor input set into the state monitoring model under the above working conditions, and output the running state of the sensors. If the running state is abnormal, add the sensor to the abnormal sensor set L, and continue the subsequent data reconstruction; if the running state is normal, stop the sensor data reconstruction.
  • the data reconstruction module includes a data reconstruction unit and a reconstruction signal verification unit.
  • the data reconstruction unit is configured to select the data reconstruction model of the sensor under the working condition for the real-time measurement data collected by the sensor X within a certain period of time according to the type of working condition to which it belongs; input the data reconstruction model into the data set Substitute into the data reconstruction model of the working condition type to obtain the output reconstruction value of the data reconstruction model.
  • the reconstructed signal checking unit is configured to calculate the uncertainty bandwidth of the reconstructed value of the sensor by using an uncertainty analysis method (such as Monte Carlo method, etc.). If the uncertainty bandwidth of the sensor reconstructed value does not exceed the fixed value, the reconstructed value is valid, and the value is sent back to the instrument and control system to temporarily replace the abnormal sensor.
  • an uncertainty analysis method such as Monte Carlo method, etc.
  • the information is pushed to the operator in the form of an alarm
  • the information includes "the sensor X is abnormal”, “the sensor X has been reconfigured and sent back", “the reconfigured value of the sensor X may be invalid after 4 hours” and the The expected consequences of failure of the reconstructed signal from the sensor. If the uncertainty bandwidth exceeds the fixed value, the reconstruction of the sensor will be stopped, and the information will be pushed to the operator in the form of an alarm.
  • the information includes "abnormal sensor X", “failure of sensor X reconstruction” and the expected Loss of function timing and consequences.

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Plasma & Fusion (AREA)
  • General Engineering & Computer Science (AREA)
  • High Energy & Nuclear Physics (AREA)
  • Business, Economics & Management (AREA)
  • Emergency Management (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

一种传感器智能数据重构方法及系统。方法包括:S1、判断是否满足数据重构模型投运条件,如果满足数据重构模型投运条件,则进入步骤S2,否则数据重构步骤中止;S2、进行传感器状态检测并判断检测结果是否异常,如果显示检测结果异常,则进入步骤S3,如果显示检测结果正常,则数据重构步骤中止;S3、将步骤S2中得到的检测结果异常传感器的实时测量数据输入数据重构模型得到重构数值。通过依次判断是否满足数据重构模型投运条件、传感器状态检测结果是否异常,将检测结果异常传感器的实时测量数据输入数据重构模型得到重构数值,可在传感器发生快故障后,快速主动识别故障传感器。

Description

一种传感器智能数据重构方法及系统
相关公开的交叉引用
本公开要求在2021年11月17日提交国家知识产权局、申请号为202111360560.4、发明名称为“一种传感器智能数据重构方法及系统”的中国专利申请的优先权,该申请的全部内容通过引用结合在本公开中。
技术领域
本公开属于但不限于数据处理领域,具体地说,涉及一种传感器智能数据重构方法及系统。
背景技术
核电机组的安全可靠运行离不开仪控系统的功能完备,而传感器功能的完好性是仪控系统核电厂运行状态监测、触发自动动作的基础。为提高传感器的可靠性,在现有核电机组设计中,主要采取了“更高鉴定要求”、“硬件冗余配置”和“运维期间的定期试验或视情维修”等3种技术手段。在核电厂中,传感器分为“安全级”和“非安全级”两大类。对于核安全级传感器提出了更高的质保、抗震和鉴定等要求。针对重要传感器,一般还采用了传感器硬件冗余配置的方法,以进一步提高同一测点数据采集功能的可靠性。在运维阶段,针对传感器上可能发生的慢故障,核电厂采用定期试验和视情维修等方式来检查传感器的运行状态是否正常,以尽早识别微小故障,基于预防性的思路来提高传感器的功能的完备性。
尽管采用了多种技术手段提升核电厂传感器的可靠性,但仍难以避免传感器故障(如:突然性的功能丧失)的发生。在传感器故障后,可能导致系统基于错误的机组状态误触发自动动作(如:误动或拒动),威胁机组安全,导致机组运行的鲁棒性仍然不高,部分传感器故障仍需要人员及 时干预。
并且在发生由共模故障导致的冗余/非冗余传感器感知数据丧失时,仍会导致基于相关传感器的数据感知功能突然丧失。进而,导致在相关工况下,机组运行的鲁棒性不足,使得在运维人员尚未能够有效干预时,就引发不适当的控制系统拒动或误动,削弱了机组的安全性和经济性。
为解决此问题,亟需提出一种可在紧急情况下,替代故障传感器产生有效替代数据的数据重构方法及系统,用于在确保安全的情况下稳定机组运行状态,为运行控制和维修替换争取时间。
发明内容
针对现有技术中存在的缺陷,本公开的目的在于提供一种传感器智能数据重构方法及系统,可在传感器发生快故障后,快速主动识别故障传感器,并在满足相关条件的情况下自动生成数据,替代故障传感器数值,保证机组处于安全稳定状态,为操纵员的运行控制操作和维修人员的现场维修替换工作争取时间。
为达到以上目的,根据本公开的实施例采用的技术方案是提供一种传感器智能数据重构方法,包括以下步骤:
S1、判断是否满足数据重构模型投运条件,如果满足数据重构模型投运条件,则进入步骤S2,否则数据重构步骤中止;
S2、进行传感器状态检测并判断检测结果是否异常,如果显示检测结果异常,则进入步骤S3,如果显示检测结果正常,则数据重构步骤中止;
S3、将步骤S2中得到的所述检测结果异常传感器的实时测量数据输入数据重构模型得到重构数值。
进一步,步骤S1包括如下子步骤:
S101、判断反应堆核功率是否高于20%*Pn,其中Pn为反应堆输出的额定有功功率,若反应堆核功率高于20%*Pn,则转至步骤S102;否则, 数据重构步骤中止;
S102、判断需要进行数据重构的传感器集合K中传感器X模型的输入集合中,虚拟传感器的数量是否符合要求,若符合要求,则转至步骤S103;如不符合要求,数据重构步骤中止,从所述传感器集合K中删除传感器X;
S103、向操纵员推送“启动请求”,操纵员同意启动后,继续后续数据重构步骤。
进一步,步骤S102包括:判断所述传感器集合K中,传感器X模型的输入传感器中,虚拟传感器的数量是否达到输入传感器集合总数的一定比例k,若虚拟传感器的数量达到输入传感器集合总数的一定比例,则数据重构步骤中止,从所述传感器集合K中删除所述传感器X;若虚拟传感器的数量未达到所述比例k,则将所述传感器X加入实体传感器集合N并进一步判断所述实体传感器集合N中实体传感器的数量是否大于0,若所述实体传感器集合N中的传感器数量大于0,则确定所述传感器集合K中,虚拟传感器的数量符合要求。
进一步,步骤S2包括如下子步骤:
S201、将所述传感器集合K中的每个传感器最近若干采样时间段内的实时测量数据输入工况分类模型进行工况分类,并匹配所述传感器在所述工况下的传感器状态监测模型;
S202、将所述实时测量数据输入所述工况下的状态监测模型中进行运行状态监测,若运行状态为异常,则将所述传感器加入异常传感器集合L中,继续后续数据重构步骤,若运行状态为正常,则所述传感器数据重构步骤中止。
进一步,步骤S3包括如下子步骤:
S301、将所述实时测量数据代入所述工况类型对应的数据重构模型中,以得到所述数据重构模型输出的重构数值;
S302、采用不确定度分析法计算所述重构数值的不确定度带宽,若所述重构数值的不确定度带宽不超过定值,则重构数值有效,将所述重构数值回传到仪控系统中,用于暂时替代异常传感器。
进一步,步骤S302中的所述不确定度分析法为蒙特卡洛法。
一种传感器智能数据重构系统,所述系统包括模型投运条件判断模块、传感器状态检测模块以及数据重构模块,
所述模型投运条件判断模块,被配置成判断是否满足数据重构模型投运条件,如果满足数据重构模型投运条件,则继续后续数据重构,否则数据重构中止;
所述传感器状态检测模块,被配置成进行传感器状态检测,并判断检测结果是否异常,如果传感器状态检测显示检测结果异常,则继续后续数据重构,否则数据重构中止;
所述数据重构模块,被配置成将所述传感器的实时测量数据输入数据重构模型得到重构数值。
进一步,所述模型投运条件判断模块包括功率判断单元、虚拟传感器数量判断单元以及启动请求发送单元,
所述功率判断单元,被配置成判断反应堆核功率是否在20%Pn以上,若反应堆核功率在20%Pn以上,则继续后续数据重构;否则数据重构中止;
所述虚拟传感器数量判断单元,被配置成判断需要进行数据重构的传感器集合K中传感器X模型的输入集合中,虚拟传感器的数量是否符合要求,若符合要求,则继续后续数据重构;否则数据重构中止;
所述启动请求发送单元,被配置成向操纵员推送“启动请求”,所述操纵员同意启动后,启动后续数据重构。
进一步,所述传感器状态检测模块包括工况分类单元以及运行状态判断单元,
所述工况分类单元,被配置成将所述传感器集合K中的每个传感器最近若干采样时间段内的实时测量数据输入工况分类模型进行工况分类,并匹配所述传感器在所述工况下的传感器状态监测模型;
所述运行状态判断单元,被配置成将所述实时测量数据输入所述工况下的状态监测模型中进行运行状态监测,若运行状态为异常,则将所述传 感器加入异常传感器集合L中,继续后续数据重构,若运行状态为正常,则所述传感器数据重构中止。
进一步,所述数据重构模块包括数据重构单元以及重构信号校验单元,
所述数据重构单元,被配置成将所述实时测量数据代入所述工况类型对应的数据重构模型中,以得到所述数据重构模型输出的重构数值;
所述重构信号校验单元,被配置成采用不确定度分析法计算所述重构数值的不确定度带宽,若所述重构数值的不确定度带宽不超过定值,则所述重构数值有效,将所述重构数值回传到仪控系统中,用于暂时替代异常传感器。
本公开的效果在于:本公开的实施例公开的一种传感器智能数据重构方法及系统,可在传感器发生快故障后,快速主动识别故障传感器,并在满足相关条件的情况下自动生成数据,替代故障传感器数值,保证机组处于安全稳定状态,为操纵员的运行控制操作和维修人员的现场维修替换工作争取时间。可以代替传统“定期试验”,在计划停堆时,统一有针对性的维修,从而降低了运维成本。也可以减少冗余传感器配置数量,从而减少机组建造成本。
避免核电厂系统基于错误的机组状态信号触发自动动作误动或拒动,无需操纵员及时干预,可在传感器共模故障条件下自动稳定机组运行状态,提高核电厂运行的安全性和稳定性。
附图说明
图1为根据本公开的实施例所述的一种传感器智能数据重构方法中传感器重构模型的示意图;
图2为根据本公开的实施例所述的一种传感器智能数据重构方法的流程图;
图3为本公开的实施例所述的一种传感器智能数据重构方法中步骤S1 “判断是否满足数据重构模型投运条件”的流程图;
图4为本公开的实施例所述的一种传感器智能数据重构方法中步骤S2“进行传感器状态检测,并判断检测结果是否异常”的流程图;
图5为本公开的实施例所述的一种传感器智能数据重构方法中步骤S3“进行传感器数据重构”的流程图。
具体实施方式
下面结合附图和具体实施方式对本公开作进一步描述。
实施例一
为便于表述本公开的技术方案,定义集合K为需进行数据重构的传感器(以下简称“重构传感器”)集合。在不同机组运行工况下,每个数据重构传感器都有对应的数据重构模型。在本实施例中,数据重构模型基于自联想核回归算法(AAKR)训练而成,通过事先将某个工况下正常运行的传感器测量数据带入数据重构模型进行模型训练得到我们所需的数据重构模型。如图1所示,传感器X的重构模型输入为对应工况下相关传感器的实时测量数据,输出为传感器X的重构值。
如图2所示,根据本公开的实施例的一种传感器智能数据重构方法包括以下步骤:
S1:判断是否满足数据重构模型投运条件,如果满足数据重构模型投运条件,则进入步骤S2,否则数据重构步骤中止,等待下一个判断周期继续判断是否满足数据重构模型投运条件。
S2:进行传感器状态检测,并判断检测结果是否异常,如果显示检测结果异常,则进入步骤S3,如果显示检测结果正常,则数据重构步骤中止,等待下一个判断周期继续进行传感器状态检测。
S3:将传感器实时测量的数据输入数据重构模型得到重构数值。
如图3所示,在电厂运行过程中,按照一定周期,对数据重构模型投运条件进行判断。步骤S1具体包括如下子步骤:
S101:判断反应堆核功率是否在20%Pn以上,若反应堆核功率在20%Pn以上,则转至步骤S102;否则,数据重构步骤中止,等待下一个判断周期继续判断反应堆核功率是否在20%Pn以上。
这里,Pn为反应堆输出的额定有功功率,反应堆核功率在20%Pn以上表示核反应堆的状态相对稳定,可启动数据重构模型。
当反应堆核功率低于20%Pn时,初始化集合K。若数据重构模型处于启动状态,则数据重构步骤中止,并将“模型停止运行”的信息推送给操纵员知悉。
S102:判断需要进行数据重构的传感器集合K中,虚拟传感器的数量是否符合要求,若符合要求,则转至步骤S103;如不符合要求则数据重构步骤中止。
判断需要进行数据重构的传感器集合K中,虚拟传感器的数量是否符合要求包括:判断需要进行数据重构的传感器集合K中,传感器X模型的输入传感器中,虚拟传感器的数量是否达到输入传感器集合总数的一定比例。若虚拟传感器的数量达到输入传感器集合总数的一定比例,则无法实现重构,从集合K中删除传感器X。若虚拟传感器的数量未达到一定比例,则加入实体传感器集合N。若实体传感器集合N中的传感器数量大于0,则确定需要进行数据重构的传感器集合K中,虚拟传感器的数量符合要求。
若集合K中存在无法重构的传感器,则报警的形式,将信息推送给操纵员,信息包括:“传感器X重构步骤失效”和“与该传感器信号失效后预计产生的后果”。
S103:向操纵员推送“启动请求”,“启动请求”的内容包括:启动请求和满足监测重构前提条件的传感器列表。操纵员同意启动后,启动数据重构后续步骤。
按照一定周期,对K集合中的每个传感器分别进行状态检测,筛选异常传感器并存入异常传感器集合L。如图4所示,步骤S2包括如下子步骤:
S201:将传感器X最近若干采样时间段(例如,60s)内采集的数据输入对应的工况分类模型,进行工况分类,得到实时测量样本所属的工况 类型,并选取传感器在该工况下的传感器状态监测模型。
在本实施例中,工况分类模型基于密度的聚类算法训练而成,通过事先将某个工况下正常运行的传感器测量数据带入工况分类模型进行模型训练得到我们所需的工况分类模型。
工况分类模型包括升功率模型以及降功率模型。对传感器进行工况分类有利于了解传感器的信号变化趋势,提高数据重构准确度。
S202:将需要进行数据重构的传感器集合K中所有传感器实时测量数据输入上述工况下的状态监测模型中,输出传感器的运行状态。若运行状态为异常,则将该传感器加入异常传感器集合L中,继续后续数据重构步骤,若运行状态为正常,则将该传感器数据重构步骤中止。
在本实施例中,状态监测模型基于自联想神经网络训练而成,通过事先将某个工况下正常运行的传感器测量数据带入状态监测模型进行模型训练得到我们所需的状态监测模型。
如图5所示对异常传感器集合L中的每个异常传感器进行数据重构。步骤S3包括如下子步骤:
S301,传感器数据重构,即:对传感器X在一定时间段内采集的实时测量数据按照其工况类型,并选取传感器在该工况下的数据重构模型;将传感器X在一定时间段内采集的实时测量数据代入该工况类型的数据重构模型中,得到输出的重构数值。
S302,模型重构信号校验回传,即:采用不确定度分析法(如蒙特卡洛法等)计算传感器重构值的不确定度带宽。若传感器重构值的不确定度带宽不超过定值,则重构数值有效,将该数值回传到仪控系统中,用于暂时替代异常传感器。
同时,以报警的形式将信息推送给操纵员,该信息包括“传感器X出现异常”、“传感器X已经完成重构回传”、“传感器X的重构数值可能在4小时后失效”和该传感器的重构信号失效后预计产生的后果。若不确定度带宽超过定值,则停止该传感器的重构,并以报警的形式将信息推送给操纵员,信息包括“传感器X出现异常”、“传感器X重构失败”和该传感器 的预计功能丧失时间和后果。
以一套核电厂多回路耦合能量传输系统台架在从6kw-7kw的升功率过程中,一回路加热上升段温度发生1%的零点漂移的传感器异常的传感器数据重构为例,对本专利公开的实施方法进行说明:
存在启动数据重构模型需求时,操纵员启动判断是否满足数据重构模型投运条件,数据重构模型每60s(该周期可能根据实际应用场景需求延长或缩短)执行一遍以下子步骤:
如果确定当前功率在台架回路满功率的20%以上,则满足模型“启动要求”。
由于所有传感器都是实时测量传感器,因此每个传感器的输入集合的重构传感器数量为0,所有传感器均满足监测重构前提条件。
向操纵员发送“启动请求”。接收到确认信号后,启动数据重构后续步骤。
将每个传感器60s的测量数据代入各自的工况分类模型,均检测为升功率工况。
将每个传感器60s的测量数据代入各自的升功率工况下的状态监测模型中,检测出一回路加热段出口温度出现异常。
将一回路加热段出口温度代入升功率工况下的重构模型中,得到回路加热段出口温度正常升功率运行的重构值。
计算一回路加热段出口温度重构值的不确定度带宽,带宽较小。
将一回路加热段出口温度的重构数值回传到对应的传感器中,并以报警的形式将信息推送给操纵员,信息包括“一回路加热段出口温度传感器出现异常”、“一回路加热段出口温度传感器已经完成重构回传”和“一回路加热段出口温度传感器的重构数值可能在4小时后失效”。
通过上述实施例可以看出,根据本公开实施例提供的一种传感器智能数据重构方法,可在传感器发生快故障后,快速主动识别故障传感器,并在满足相关条件的情况下自动生成数据,替代故障传感器数值,保证机组 处于安全稳定状态,为操纵员的运行控制操作和维修人员的现场维修替换工作争取时间。可以代替传统“定期试验”,在计划停堆时,统一有针对性的维修,从而降低了运维成本。也可以减少冗余传感器配置数量,从而减少机组建造成本。
避免核电厂系统基于错误的机组状态信号触发自动动作误动或拒动,无需操纵员及时干预,可在传感器共模故障条件下自动稳定机组运行状态,提高核电厂运行的安全性和稳定性。
实施例二
根据本公开的实施例所述的一种传感器智能数据重构系统,包括模型投运条件判断模块、传感器状态检测模块以及数据重构模块。
模型投运条件判断模块,被配置成判断是否满足数据重构模型投运条件,如果满足数据重构模型投运条件,则继续后续数据重构,否则数据重构中止,等待下一个判断周期继续判断是否满足数据重构模型投运条件。
传感器状态检测模块,被配置成进行传感器状态检测,并判断检测结果是否异常,如果传感器状态检测显示检测结果异常,则继续后续数据重构,如果传感器状态检测显示检测结果正常,则数据重构中止,等待下一个判断周期继续进行传感器状态检测。
数据重构模块,被配置成将传感器实时测量的数据输入数据重构模型得到重构数值。
其中模型投运条件判断模块包括功率判断单元、虚拟传感器数量判断单元以及启动请求发送单元。
功率判断单元,被配置成判断反应堆核功率是否在20%Pn以上,若反应堆核功率在20%Pn以上,则继续后续数据重构步骤;否则,数据重构中止,等待下一个判断周期继续判断反应堆核功率是否在20%Pn以上。
当功率低于20%Pn时,初始化集合K。若重构模型处于启动状态,则重构模型自动停止运行,并将“模型停止运行”的信息推送给操纵员知悉。
虚拟传感器数量判断单元,被配置成判断集合K中传感器的模型输入 集合中,虚拟传感器的数量是否符合要求,若符合要求,则继续后续数据重构;如不符合要求,数据重构中止。
判断需要进行数据重构的传感器集合K中,虚拟传感器的数量是否符合要求包括:判断需要进行数据重构的传感器集合K中,传感器X模型的输入传感器中,虚拟传感器的数量是否达到输入传感器集合总数的一定比例。若虚拟传感器的数量达到输入传感器集合总数的一定比例,则无法实现重构,从集合K中删除传感器X。若虚拟传感器的数量未达到一定比例,则加入实体传感器集合N。若实体传感器集合N中的传感器数量大于0,则确定需要进行数据重构的传感器集合K中,虚拟传感器的数量符合要求。
若集合K中存在无法重构的传感器,则报警的形式,将信息推送给操纵员,信息包括:“传感器X重构步骤失效”和“与该传感器信号失效后预计产生的后果”。
启动请求发送单元,被配置成向操纵员推送“启动请求”,“启动请求”的内容包括:启动请求和满足监测重构前提条件的传感器列表。操纵员同意启动后,启动数据重构模型。
传感器状态检测模块包括工况分类单元以及运行状态判断单元。
工况分类单元,被配置成将传感器X最近若干采样时间段(例如,60s)内采集的数据输入工况分类模型,进行工况分类,得到实时测量样本所属的工况类型,并选取传感器在该工况下的传感器状态监测模型。
运行状态判断单元,被配置成将传感器的输入集合中所有传感器采集数据输入上述工况下的状态监测模型中,输出传感器的运行状态。若运行状态为异常,则将该传感器加入异常传感器集合L中,继续后续数据重构,若运行状态为正常,则将该传感器数据重构中止。
数据重构模块包括数据重构单元以及重构信号校验单元。
数据重构单元,被配置成对传感器X的一定时间段内采集的实时测量数据按照其所属的工况类型,选取传感器在该工况下的数据重构模型;将数据重构模型输入数据集合代入该工况类型的数据重构模型中,得到数据重构模型输出重构数值。
重构信号校验单元,被配置成采用不确定度分析法(如蒙特卡洛法等)计算传感器重构值的不确定度带宽。若传感器重构值的不确定度带宽不超过定值,则重构数值有效,将该数值回传到仪控系统中,用于暂时替代异常传感器。
同时,以报警的形式将信息推送给操纵员,该信息包括“传感器X出现异常”、“传感器X已经完成重构回传”、“传感器X的重构数值可能在4小时后失效”和该传感器的重构信号失效后预计产生的后果。若不确定度带宽超过定值,则停止该传感器的重构,并以报警的形式将信息推送给操纵员,信息包括“传感器X出现异常”、“传感器X重构失败”和该传感器的预计功能丧失时间和后果。
本公开所述的方法并不限于具体实施方式中所述的实施例,本领域技术人员根据本公开的技术方案得出其他的实施方式,同样属于本公开的技术创新范围。

Claims (10)

  1. 一种传感器智能数据重构方法,包括以下步骤:
    S1、判断是否满足数据重构模型投运条件,如果满足数据重构模型投运条件,则进入步骤S2,否则数据重构步骤中止;
    S2、进行传感器状态检测并判断检测结果是否异常,如果显示检测结果异常,则进入步骤S3,如果显示检测结果正常,则数据重构步骤中止;
    S3、将步骤S2中得到的所述检测结果异常传感器的实时测量数据输入数据重构模型得到重构数值。
  2. 如权利要求1中所述的一种传感器智能数据重构方法,其中,步骤S1包括如下子步骤:
    S101、判断反应堆核功率是否高于20%*Pn,其中Pn为反应堆输出的额定有功功率,若反应堆核功率高于20%*Pn,则转至步骤S102;否则,数据重构步骤中止;
    S102、判断需要进行数据重构的传感器集合K中传感器X模型的输入集合中,虚拟传感器的数量是否符合要求,若符合要求,则转至步骤S103;如不符合要求,数据重构步骤中止,从所述传感器集合K中删除传感器X;
    S103、向操纵员推送“启动请求”,操纵员同意启动后,继续后续数据重构步骤。
  3. 如权利要求2中所述的一种传感器智能数据重构方法,其中
    步骤S102包括:判断所述传感器集合K中,传感器X模型的输入传感器中,虚拟传感器的数量是否达到输入传感器集合总数的一定比例k,若虚拟传感器的数量达到输入传感器集合总数的一定比例,则数据重构步骤中止,从所述传感器集合K中删除所述传感器X;若虚拟传感器的数量未达到所述比例k,则将所述传感器X加入实体传感器集合N并进一步判断所述实体传感器集合N中实体传感器的数量是否大于0,若所述实体传 感器集合N中的传感器数量大于0,则确定所述传感器集合K中,虚拟传感器的数量符合要求。
  4. 如权利要求2或3所述的一种传感器智能数据重构方法,其中,步骤S2包括如下子步骤:
    S201、将所述传感器集合K中的每个传感器最近若干采样时间段内的实时测量数据输入工况分类模型进行工况分类,并匹配所述传感器在所述工况下的传感器状态监测模型;
    S202、将所述实时测量数据输入所述工况下的状态监测模型中进行运行状态监测,若运行状态为异常,则将所述传感器加入异常传感器集合L中,继续后续数据重构步骤,若运行状态为正常,则所述传感器数据重构步骤中止。
  5. 如权利要求4中所述的一种传感器智能数据重构方法,其中,步骤S3包括如下子步骤:
    S301、将所述实时测量数据代入所述工况类型对应的数据重构模型中,以得到所述数据重构模型输出的重构数值;
    S302、采用不确定度分析法计算所述重构数值的不确定度带宽,若所述重构数值的不确定度带宽不超过定值,则重构数值有效,将所述重构数值回传到仪控系统中,用于暂时替代异常传感器。
  6. 如权利要求5中所述的一种传感器智能数据重构方法,其中,步骤S302中的所述不确定度分析法为蒙特卡洛法。
  7. 一种传感器智能数据重构系统,其中,所述系统包括:模型投运条件判断模块、传感器状态检测模块以及数据重构模块,
    所述模型投运条件判断模块,被配置成判断是否满足数据重构模型投 运条件,如果满足数据重构模型投运条件,则继续后续数据重构,否则数据重构中止;
    所述传感器状态检测模块,被配置成进行传感器状态检测,并判断检测结果是否异常,如果传感器状态检测显示检测结果异常,则继续后续数据重构,否则数据重构中止;
    所述数据重构模块,被配置成将所述传感器的实时测量数据输入数据重构模型得到重构数值。
  8. 如权利要求7中所述的一种传感器智能数据重构系统,其中,所述模型投运条件判断模块包括功率判断单元、虚拟传感器数量判断单元以及启动请求发送单元,
    所述功率判断单元,被配置成判断反应堆核功率是否在20%Pn以上,若反应堆核功率在20%Pn以上,则继续后续数据重构;否则数据重构中止;
    所述虚拟传感器数量判断单元,被配置成判断需要进行数据重构的传感器集合K中传感器X模型的输入集合中,虚拟传感器的数量是否符合要求,若符合要求,则继续后续数据重构;否则数据重构中止;
    所述启动请求发送单元,被配置成向操纵员推送“启动请求”,所述操纵员同意启动后,启动后续数据重构。
  9. 如权利要求8中所述的一种传感器智能数据重构系统,其中,所述传感器状态检测模块包括工况分类单元以及运行状态判断单元,
    所述工况分类单元,被配置成将所述传感器集合K中的每个传感器最近若干采样时间段内的实时测量数据输入工况分类模型进行工况分类,并匹配所述传感器在所述工况下的传感器状态监测模型;
    所述运行状态判断单元,被配置成将所述实时测量数据输入所述工况下的状态监测模型中进行运行状态监测,若运行状态为异常,则将所述传感器加入异常传感器集合L中,继续后续数据重构,若运行状态为正常, 则所述传感器数据重构中止。
  10. 如权利要求9中所述的一种传感器智能数据重构系统,其中所述数据重构模块包括数据重构单元以及重构信号校验单元,
    所述数据重构单元,被配置成将所述实时测量数据代入所述工况类型对应的数据重构模型中,以得到所述数据重构模型输出的重构数值;
    所述重构信号校验单元,被配置成采用不确定度分析法计算所述重构数值的不确定度带宽,若所述重构数值的不确定度带宽不超过定值,则所述重构数值有效,将所述重构数值回传到仪控系统中,用于暂时替代异常传感器。
PCT/CN2022/132511 2021-11-17 2022-11-17 一种传感器智能数据重构方法及系统 WO2023088355A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111360560.4A CN114283956A (zh) 2021-11-17 2021-11-17 一种传感器智能数据重构方法及系统
CN202111360560.4 2021-11-17

Publications (1)

Publication Number Publication Date
WO2023088355A1 true WO2023088355A1 (zh) 2023-05-25

Family

ID=80869285

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/132511 WO2023088355A1 (zh) 2021-11-17 2022-11-17 一种传感器智能数据重构方法及系统

Country Status (2)

Country Link
CN (1) CN114283956A (zh)
WO (1) WO2023088355A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114283956A (zh) * 2021-11-17 2022-04-05 中国核电工程有限公司 一种传感器智能数据重构方法及系统

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000081991A (ja) * 1998-07-09 2000-03-21 Toyota Central Res & Dev Lab Inc フェ―ルセ―フ機能付き情報処理装置
CN107153414A (zh) * 2017-07-17 2017-09-12 北京航空航天大学 一种基于动静混合策略的故障检测与重构系统及方法
CN112466496A (zh) * 2020-11-17 2021-03-09 中广核工程有限公司 核电厂智能技术规范方法及系统
CN113157992A (zh) * 2021-01-29 2021-07-23 河北建投新能源有限公司 风机传感器数据重构方法、装置与计算机可读存储介质
CN113221435A (zh) * 2021-02-08 2021-08-06 中国核电工程有限公司 传感器的筛选方法及装置、传感器数据重构方法及系统
CN114283956A (zh) * 2021-11-17 2022-04-05 中国核电工程有限公司 一种传感器智能数据重构方法及系统

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR102274139B1 (ko) * 2020-03-10 2021-07-06 조선대학교산학협력단 원자력 발전소의 운영기술지침서 이행 실시간 감시시스템
CN113010843B (zh) * 2021-03-05 2024-04-09 中国核电工程有限公司 测量参数集的确定方法及装置、验证方法及故障诊断方法
CN113654476B (zh) * 2021-08-12 2022-10-25 上海交通大学 一种耐辐照的水下变形视觉测量单元及测量方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000081991A (ja) * 1998-07-09 2000-03-21 Toyota Central Res & Dev Lab Inc フェ―ルセ―フ機能付き情報処理装置
CN107153414A (zh) * 2017-07-17 2017-09-12 北京航空航天大学 一种基于动静混合策略的故障检测与重构系统及方法
CN112466496A (zh) * 2020-11-17 2021-03-09 中广核工程有限公司 核电厂智能技术规范方法及系统
CN113157992A (zh) * 2021-01-29 2021-07-23 河北建投新能源有限公司 风机传感器数据重构方法、装置与计算机可读存储介质
CN113221435A (zh) * 2021-02-08 2021-08-06 中国核电工程有限公司 传感器的筛选方法及装置、传感器数据重构方法及系统
CN114283956A (zh) * 2021-11-17 2022-04-05 中国核电工程有限公司 一种传感器智能数据重构方法及系统

Also Published As

Publication number Publication date
CN114283956A (zh) 2022-04-05

Similar Documents

Publication Publication Date Title
CN103163397B (zh) 核电站通用模拟量板件检测系统和方法
WO2021159710A1 (zh) 一种bbu故障诊断方法、装置、电子设备及存储介质
WO2022166466A1 (zh) 传感器的筛选方法及装置、传感器数据重构方法及系统
WO2023088355A1 (zh) 一种传感器智能数据重构方法及系统
CN101556679A (zh) 一种综合前端系统故障处理方法及计算机设备
WO2014173276A1 (zh) 通过hra判定dcs人机界面的可靠性的方法、系统
EP4113539A1 (en) Method and system for intelligent monitoring of state of nuclear power plant
CN109243652B (zh) 一种核电站系统压缩空气流量数据有效性判断系统及方法
CN202421448U (zh) 核电站继电器检测装置
WO2023207635A1 (zh) 燃料电池车辆及排氢阀/排水阀故障诊断方法和装置
CN110287510B (zh) 基于水轮机主阀开度数据关联的诊断控制方法、系统、存储介质和终端
CN116384526A (zh) 一种核电厂系统或设备的重要参数故障预警方法及系统
CN109581109A (zh) 一种电梯状态多信息融合诊断方法
CN111175041B (zh) 一种阀门监测方法及系统
Yu et al. On Line Monitoring of Sensor Failure in Nuclear Power Plant Based on Sequential Probability Ratio Test
CN118053275B (zh) 一种激光甲烷报警器运行管理系统
CN108919169A (zh) 一种电能表的故障自诊断方法
CN117348466A (zh) 一种二次设备管控的集中式监管控制平台及方法
CN216052738U (zh) 一种汽轮机监测仪表自检测系统
CN102509247A (zh) Ems系统一键智能体检方法
CN110286698B (zh) 基于集水井水位数据关联的诊断控制方法、系统、存储介质和终端
CN208736472U (zh) 一种饮用水在线监测设备环境异常诊断与处理系统
Jian et al. Design of Fault Injection System Based on Hardware-In-the-Loop Simulation
Zhou et al. Functional safety analysis and promotion for relay protection device platform
CN116705369A (zh) 基于安全功能导向的核电站事故后监控与诊断方法和装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22894887

Country of ref document: EP

Kind code of ref document: A1