Method for estimating radiation quantity of maintenance work order of nuclear power station

Abstract

The invention belongs to the technical field of nuclear power station radiation dose rate prediction, and provides a nuclear power station maintenance work order radiation quantity estimation method, which comprises the steps of collecting data of historical maintenance work orders to establish a historical maintenance work order set, obtaining comparison maintenance work orders in the historical maintenance work order set, and constructing a comparison maintenance work order group; establishing a linear equation related to the radiation dose; obtaining historical data of each comparison maintenance work order, substituting the historical data into the linear equation, and constructing a linear equation set; and solving the linear equation set to obtain a linear regression coefficient, and calculating a radiation dose estimated value of the current maintenance work order based on the linear regression coefficient. Relatively reliable reference basis is provided for actual radiation protection and nuclear power station personnel exposure dose control, and safe and stable operation of the nuclear power station and safety and health of operators are guaranteed.

Description

Technical Field

[0001] The present application belongs to the technical field of nuclear power plant radiation dose rate prediction, and in particular, relates to a method for estimating the radiation dose of a nuclear power plant maintenance work order. Background Art

[0002] The operation and maintenance of nuclear power plants involve the management and control of radiation doses, and ensuring the radiation safety of operators is extremely important. During the overhaul of nuclear power plants, a large amount of radiation dose absorption information of personnel was generated. In addition, during the new round of overhaul, equipment and plant buildings may be located in radiation areas, and the health of on-site operators may be harmed by radiation. For nuclear power plants, the estimation and statistics of personnel radiation doses during overhaul is a forward-looking and practical task.

[0003] At present, the radiation dose prediction based on the maintenance work order of nuclear power plants mainly includes real-time monitoring, empirical estimation, protective measures and other methods, which ensure the safety of operators to a certain extent. However, when the working environment or working type conditions change, the traditional radiation dose prediction method has obvious shortcomings, as follows:

[0004] 1. Unable to fully utilize historical radiation dose data for accurate prediction;

[0005] 2. Lack of dynamic adaptability to changes in the operating environment and types of work;

[0006] 3. The existing prediction methods are not accurate enough to provide a reliable basis for the formulation of safety protection measures. Summary of the invention

[0007] The purpose of this application is to overcome the defects of the prior art and provide a method for estimating the radiation dose of a nuclear power plant maintenance work order. By collecting and analyzing historical work order radiation dose data, the radiation dose value can be predicted before the work order task begins, providing data support for the formulation of measures in the pre-work meeting.

[0008] In order to achieve the above objectives, this application provides the following technical solutions:

[0009] The present application provides a method for estimating radiation levels of a nuclear power plant maintenance work order, comprising:

[0010] Step 1: Collect data of historical maintenance work orders to establish a historical maintenance work order set, obtain comparison maintenance work orders from the historical maintenance work order set, and construct a comparison maintenance work order group;

[0011] Step 2: Develop a linear equation related to radiation dose:

[0012] K i (1) *Rate i (1) +Ki 92) *Rate i 92) +…+K i (m) *Rate i (m) =E i / T i ;

[0013] In the formula, K i (j) To execute the maintenance work order R i The ratio of the working hours in the jth room to the total working hours, Rate i (j) To execute the maintenance work order R i The radiation dose rate to the jth room at time, E i For maintenance work order R i The radiation dose of a single person, T i For maintenance work order R i Total working hours;

[0014] Step 3: Obtain the historical data of each maintenance work order for comparison, substitute the historical data into the linear equation, and construct a linear equation system;

[0015] Step 4: Solve the linear equations to obtain linear regression coefficients, and calculate the estimated radiation dose of the current maintenance work order based on the linear regression coefficients.

[0016] In some embodiments, a historical maintenance work order set is searched for a historical maintenance work order that performs the same task as the current maintenance work order as a comparison maintenance work order.

[0017] In some embodiments, the mean and standard deviation of the working time data in the comparison maintenance work order group are calculated, a screening interval is set based on the mean and standard deviation of the working time data, and the comparison maintenance work orders with working time values ​​outside the screening interval are eliminated.

[0018] In some embodiments, step 2 comprises:

[0019] Create a single maintenance work order R i Radiation dose per person E i Calculation formula:

[0020]

[0021] In the formula, E i For maintenance work order R i The radiation dose of a single person, T i (j) To execute the maintenance work order R i The working hours to go to the jth room, Ratei (j) To execute the maintenance work order R i The radiation dose rate to the jth room at time ;

[0022] The above formula satisfies the time constraint:

[0023]

[0024] Where, T i For maintenance work order R i of working hours;

[0025] The radiation dose for a single person is E i Divide both sides of the calculation formula by T i ,get:

[0026]

[0027] make The above formula is simplified to:

[0028] K i (1) *Rate i (1) +K i (2) *Rate i (2) +…+K i (m) *Rate i (m) =E i / T i .

[0029] In some embodiments, a single work order R is executed centrally for historical maintenance work orders. i The corresponding room list and the corresponding radiation dose rate of the room at that time are respectively represented by {Room i (1) ,Room i (2) ,…,Room i (m)} and {Rate i (1) ,Rate i (2) ,…,Rate i (m)} indicates that the work order R i The work hours involved are expressed as a single T i , while executing the work order task, the radiation dose absorbed is E i , if work order R iThe corresponding rooms are all zero dose rate, so the absorbed dose E corresponding to this work order task i is zero.

[0030] In some embodiments, the system of linear equations is:

[0031] K1 (1) *Rate1 (1) +K1 (2) *Rate1 (2) +…+K1 (m) *Rate1 (m) =E1 / T1

[0032] K2 (1) *Rate2 (1) +K2 (2) *Rate2 (2) +…+K2 (m) *Rate2 (m) =E2 / T2

[0033] …

[0034] K n (1) *Rate n (1) +K n (2) *Rate n (2) +…+K n (m) *Rate n (m) =E n / T n

[0035] In the formula, K i (j) To execute the i-th maintenance work order R i The time spent in the jth room on the maintenance work order R i The proportion of total working hours, Rate i (j) To execute the i-th maintenance work order R i The radiation dose rate to the jth room at time, E i is the i-th maintenance work order R i The radiation dose of a single person, T i is the i-th maintenance work order R i Total working hours;

[0036] Based on the linear equation system, establish linear operations:

[0037]

[0038] In the formula, {K * (1) ,K * (2) ,…,K * (m)} is the linear regression coefficient to be solved.

[0039] In some embodiments, the room set {R i} is the union of the rooms visited by different historical maintenance orders for the same task, {Rate i (j)} is the union {R i} is a dose rate array consisting of the dose rates of each room in .

[0040] In some embodiments, solving the linear equations to obtain linear regression coefficients includes:

[0041] Establish the loss function J(θ):

[0042]

[0043] With the goal of minimizing the loss function J(θ), the linear equations are solved using the gradient descent method to obtain the linear regression coefficients.

[0044] In some embodiments, calculating the radiation dose estimate of the current maintenance work order based on the linear regression coefficient includes: substituting the linear regression coefficient value into the linear equation to obtain the radiation dose calculation formula of the current maintenance work order, obtaining the radiation dose rate and operation time data of the current maintenance work order, and substituting them into the radiation dose calculation formula to obtain the radiation dose estimate of the current maintenance work order.

[0045] In some embodiments, step 4 further includes: when the dose rate array {Rate i (j) When the dimension of} is greater than the number of maintenance work orders, a constant term and a unit matrix are introduced into the linear equation system as bias terms, and the ridge regression analysis method is used to solve the linear equation system to obtain the linear regression coefficients.

[0046] Compared with the prior art, the method for estimating radiation dose of a nuclear power plant maintenance work order provided by this application has the following advantages:

[0047] Beneficial effects:

[0048] This application first collects historical maintenance work order data, and proposes a radiation dose estimation method based on the idea of ​​multivariate linear regression according to the maintenance work orders and exposure data information generated by previous overhauls. A linear equation related to radiation dose is established, and the linear equation is solved using comparative maintenance work orders with the same task, thereby realizing the radiation dose estimation of the current maintenance work order. Ultimately, before the work order is executed, the range of personnel exposure doses that may be generated by the current maintenance work order is informed to the operation and maintenance personnel, providing a relatively reliable reference basis for actual radiation protection and nuclear power plant personnel exposure dose control, and ensuring the safe and stable operation of the nuclear power plant and the safety and health of the operating personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solution of the present application, the following is a brief introduction to the drawings required for the technical description.

[0050] Figure 1 A flow chart of a method for estimating radiation levels from a nuclear power plant maintenance work order provided for this application;

[0051] Figure 2 Data related to work order dose estimation and overall dose estimation during overhaul period provided for this application and schematic diagram of related sources;

[0052] Figure 3 The overall calculation flow chart provided for this application. DETAILED DESCRIPTION

[0053] The following is further explained in detail through specific implementation methods.

[0054] like Figure 1 As shown, the present application provides a method for estimating radiation dose of a nuclear power plant maintenance work order, comprising the following steps:

[0055] Step 1: Collect data of historical maintenance work orders to establish a historical maintenance work order set, search the historical maintenance work order set for the same task as the current maintenance work order as a comparison maintenance work order, and construct a comparison maintenance work order group;

[0056] Step 2: Develop a linear equation related to radiation dose:

[0057] K i (1) *Rate i (1) +K i (2) *Rate i (2) +…+K i (m) *Rate i (m) =E i / T i;

[0058] In the formula, K i (j) To execute the maintenance work order R i The proportion of the working hours in the jth room to the total working hours, so that the maintenance work order time K for performing the same task in different maintenance periods i (j) Remain unchanged, Rate i (j) To execute the maintenance work order R i The radiation dose rate when going to the jth room, executing the maintenance work order R i There are m rooms to go to, j = 1, 2, ..., m, E i For maintenance work order R i The radiation dose of a single person, T i For maintenance work order R i Total working hours;

[0059] Step 3: Obtain the historical data of each maintenance work order for comparison, substitute the historical data into the linear equation, and construct a linear equation system;

[0060] Step 4: Solve the linear equations to obtain linear regression coefficients, and calculate the estimated radiation dose of the current maintenance work order based on the linear regression coefficients.

[0061] During the overhaul process of a nuclear power plant, maintenance work will generate thousands of maintenance work orders, each of which is associated with an engineer team and a series of work rooms. In order to complete the work order task, the engineer team needs to carry out on-site operations in a designated series of rooms. After the work order is completed, the time-consuming data of the overhaul work order and the exposure dose data of the overhaul work order personnel will be exported. The collection of such data generated by multiple work orders forms the historical maintenance work order set required to be established by this method, which is the data input required by this method. In addition, since the radiation-carrying areas are mainly located in the radiation control area. Therefore, in order to reduce the complexity of multivariate linear regression calculations to a certain extent, in the actual overhaul work order personnel exposure dose estimation process, only the radiation-carrying rooms located in the radiation control area will be assigned the actual value of the room dose rate before the work order is executed, and the remaining areas will be assigned zero.

[0062] Regarding the radiation dose estimation needs during the overhaul of the production area of ​​a nuclear power plant, before establishing the radiation dose estimation calculation model, in order to facilitate the establishment of a unified calculation paradigm, based on the actual operation and maintenance scenarios and rules of the nuclear power plant, the following two prerequisites are proposed for the dose estimation of the operation and maintenance tasks during the overhaul:

[0063] 1. During the two overhauls DX1 and DX2, two maintenance work orders R1 and R2 performed the same task. Both work orders R1 and R2 went to a certain room during the execution. j, then during the execution of work order R1 and work order R2, the operator team is executing work order R1 in Room j The proportion of the waiting time in the total time of work order R1 is K1 (j) , the operator team is executing work order R2 in Room j The proportion of the waiting time in the total time of work order R2 is K2 (j) ,K1 (j) With K2 (j) equal.

[0064] 2. Perform a task R during the overhaul period in the historical data i , the room set to which this task is performed is {Room j}, in the new round of overhaul to perform tasks R i The room set we go to is {Room j *}. Collection {Room j} and {Room j * There must be an intersection. According to operation and maintenance experience and rules, the room sets that are completed for the same work order during different overhaul periods are generally roughly the same, and the best case is that they are exactly the same.

[0065] In the actual operation and maintenance process, the basis for calculating the estimated dose of a single work order is derived from all the work orders in the overhaul history. In a single work order task, the estimation method relies on data including the number of workers, working hours, a list of rooms involved in the work order, and the historical real-time dose rate of a single room. The overall dose of a historical overhaul is determined by the cumulative estimated doses of all work orders. All work order information of the overhaul plan can be obtained from the overhaul planning platform, so the estimated overall dose of the expected overhaul can be obtained by accumulating the doses of a single work order. Based on the above discussion, the data related to the work order dose estimation and the overall dose estimation during the overhaul period and the related sources are as follows: Figure 2 shown.

[0066] Based on K i (j)The radiation data and working time data of each historical maintenance work order can be extracted as a room dose rate vector and a room time ratio vector of the total time ratio. Therefore, the personnel exposure dose data of each historical work order can be expressed as the dot product result of these two vectors. In the process of constructing a linear equation system, the work order screening strategy adopted by this method is to traverse the work order sets of each overhaul, calculate the similarity between the expected work order description items and the historical work order description items, select the historical work order corresponding to the highest similarity, and obtain the linear equation representation of the work order in the form of the dot product result. The equation representations of all dot product results are combined together to form a linear equation system to be solved. Solving this equation system is to output a linear regression coefficient array. Since the work order with the highest degree of similarity in the description items of all overhauls is selected as the data source for the expected work order personnel exposure dose, the calculation basis is more reliable and the estimated results are more accurate.

[0067] In step 1, historical maintenance work orders that perform the same task as the current maintenance work order are searched in the historical maintenance work order set as comparison maintenance work orders, and a comparison maintenance work order group is constructed, which also includes:

[0068] Calculate the mean and standard deviation of the working time data in the comparison maintenance work order group, set the filtering range based on the mean and standard deviation of the working time data, and eliminate the comparison maintenance work orders with working time values ​​outside the filtering range.

[0069] The same task corresponds to different maintenance work orders during different overhaul periods. The number and engineer team of each maintenance work order may not be the same. Usually, the work orders in previous overhauls are scanned according to the description items of the expected work order, and the work order with the highest description item similarity is found as the historical comparison maintenance work order set. At the same time, since the exposure dose value of personnel is strongly correlated with the duration of the work order, and the duration of the work order may be interfered by human factors. In order to screen out the interference of human factors on the calculation process as much as possible, in the historical data processing, it is necessary to calculate the mean u and standard deviation σ of the working time array in the comparison maintenance work order set, establish a screening interval, and eliminate the maintenance work orders with working time values ​​outside the screening interval to avoid the influence of outlier data items on the overall calculation. The historical comparison maintenance work order set obtained by this step is used as the source data for constructing a linear equation system. The screening interval set in this embodiment is [u-3σ, u+3σ]. The 3σ judgment basis used in this embodiment is that about 99.7% of the data are located in the [u-3σ, u+3σ] interval, and the eliminated outliers can be regarded as necessarily located outside this data interval.

[0070] In step 2, a linear equation related to radiation dose is established, specifically:

[0071] Create a single maintenance work order R i Radiation dose per person E i Calculation formula:

[0072]

[0073] In the formula, E i For maintenance work order R i The radiation dose of a single person, T i (j) To execute the maintenance work order R i The working hours to go to the jth room, Rate i (j) To execute the maintenance work order R i The radiation dose rate when going to the jth room, executing the maintenance work order R i There are m rooms to go to, j = 1, 2, ..., m;

[0074] The above formula satisfies the time constraint:

[0075]

[0076] Where, T i For maintenance work order R i of working hours;

[0077] The radiation dose for a single person is E i Divide both sides of the calculation formula by T i ,get:

[0078]

[0079] make The above formula is simplified to:

[0080] K i (1) *Rate i (1) +K i (2) *Rate i (2) +…+K i (m) *Rate i (m) =E i / T i .

[0081] In the overhaul history data, a single work order is completed by a designated engineer team. Assume that an engineer team consists of N people, and each person works on site for t i , i∈[1,n], where t i satisfy After completing the work order, the total radiation dose received by the engineer team members who executed the work order is Rate. Then the work order execution process can be described as N people entering the work site in turn, and the i-th person works for t i, and then the i+1th person takes over the work of the former and continues to work t i+1 until the work order is completed.

[0082] Obviously, this process can be abstracted into a model in which the operators of an engineering team complete a work order task in series. Since the level of radiation absorption of a person has nothing to do with the person himself, it is only considered to be related to the working hours. Therefore, a single engineering team completes a work order, which is equivalent to: in order to complete a work order, one person works for t hours and absorbs a dose of Rate. It can be seen that in the dose estimation process, the numerical calculation results are irrelevant to the specific number of people in the engineering team.

[0083] The overall overhaul plan consists of several work orders. This method first calculates a single work order as the basis to achieve radiation dose estimation based on multivariate linear regression. i The corresponding room list and the corresponding radiation dose rate of the room at that time are respectively given by (Room i )*1) ,Room i (2) ,…,Room i (m)} and {Rate i (1) ,Rate i (2) ,…,Rate i (m)} indicates that the work order R i The work hours involved are expressed as a single T i , while executing the work order task, the radiation dose absorbed is E i If the work order R i The corresponding rooms are all zero dose rate, so the absorbed dose E corresponding to this work order task i Must be zero. Based on a single task R in an overhaul i , the single-person radiation dose calculation formula can establish a linear equation. Considering that the total time spent on the same work order task during different overhaul periods is different, this method assumes that the proportion of the length of time spent in each room in the total time spent on the work order remains unchanged. Therefore, let The linear equation can be simplified.

[0084] The linear equations in step 3 are:

[0085] K1 (1) *Rate1 (1) +K1 (2) *Rate1 (2) +…+K1 (m) *Rate1(m) =E1 / T1

[0086] K2 (1) *Rate2 (1) +K2 (2) *Rate2 (2) +…+K2 (m) *Rate2 (m) =E2 / T2

[0087] …

[0088] K n (1) *Rate n (1) +K n (2) *Rate n (2) +…+K n (m) *Rate n (m) =E n / T n

[0089] In the formula, K i (j) To execute the i-th maintenance work order R i The time spent in the jth room on the maintenance work order R i The proportion of total working hours, Rate i (j) To execute the i-th maintenance work order R i The radiation dose rate of the jth room at the time of the comparison maintenance work order is n, i = 1, 2, ..., n, and the comparison maintenance work order R performing the same task is i There are m rooms to go to, j = 1, 2, ..., m, E i is the i-th maintenance work order R i The radiation dose of a single person, T i is the i-th maintenance work order R i Total working hours;

[0090] Based on the linear equation system, establish linear operations:

[0091]

[0092] In the formula, {K * (1) ,K * (2) ,…,K * (m)} is the linear regression coefficient to be solved.

[0093] In the linear operation representation, each row corresponds to a historical work order data of the same task, and Ratei(j) represents the room dose rate of the jth room corresponding to the work order at that time.

[0094] In the linear system, the room set {R i} is the union of the rooms visited by different historical maintenance orders for the same task, {Rate i (j)} is the union {R i} is a dose rate array consisting of the dose rates of each room in .

[0095] Considering that the rooms to which the maintenance orders for the same task may go during different overhaul periods may not be exactly the same, the {Rate i (j)} is the union of the room dose rates involved in different historical work orders of the same task.

[0096] Specifically, consider two maintenance work orders for the same task, R1 and R2. The room set involved in work order R1 is {Room1, Room2, Room3}, and the room set involved in work order R2 is {Room2, Room3, Room4}. Then in the dose rate matrix {Rate i (j)}, the row corresponding to work order R1 also needs to be aligned to the union {Room1, Room2, Room3, Room4}, but the dose rate value corresponding to Room4 of work order R1 is 0. For work order R2, the same alignment and zeroing operations are also required.

[0097] In step 4, solve the linear equations to obtain the linear regression coefficients, which are:

[0098] Establish the loss function J(θ):

[0099]

[0100] With the goal of minimizing the loss function J(θ), the linear equations are solved using the gradient descent method to obtain the linear regression coefficients.

[0101] The radiation dose rate of the room in linear operation Rate i (j) 、E i and T i are all known variables. Therefore, the radiation dose estimation problem to be solved by this method can be transformed into solving a set of linear regression coefficient arrays {K * (j)}, this method sets the loss function J(θ) and the coefficient array {K * (j)The solution is completed by gradient descent method. Under the premise that the historical maintenance work order data is used as the calculation sample, the loss function J(θ) is minimized and the multivariate linear programming is implemented by gradient descent method to solve the required coefficient {K * (j)}}.

[0102] In addition, in linear operations, the linear coefficient array {K * (j)} will be used as the solution object. When the number of maintenance work orders screened for comparison is greater than the number of variables in the dose rate coefficient array to be calculated, the dose rate matrix {Rate i (j)} has fewer columns than rows. In this case, we only need to align the linear operations and then use the gradient descent method to solve the problem with the minimum loss function to get the exact solution of the linear regression coefficient.

[0103]

[0104] Among them, Rate is the dose rate matrix after alignment, Rate T is the transposed matrix of the Rate matrix, and y is the linear combination value array of the linear equation system {E j / T i}.

[0105] In step 4, the estimated radiation dose of the current maintenance work order is calculated based on the linear regression coefficient, specifically:

[0106] Substitute the linear regression coefficient value into the linear equation to obtain the radiation dose calculation formula of the current maintenance work order, obtain the radiation dose rate and operation time data of the current maintenance work order, and substitute them into the radiation dose calculation formula to obtain the radiation dose estimate of the current maintenance work order.

[0107] After calculating the linear regression coefficient array, bring it into the linear equation and multiply it by the radiation dose rate of the corresponding room under the current maintenance work order, that is, the radiation dose absorbed per unit time when the operator completes the work order. This value is multiplied by the expected duration of the current maintenance work order to obtain the radiation dose value absorbed by the team during the execution of the work order. The overall calculation process is as follows: Figure 3 shown.

[0108] Solve to get the coefficient array {K * (j)}, where the element values ​​of the coefficient array correspond to the room numbers one by one. Figure 3In the step of returning the solution result, it is necessary to find the coefficient value corresponding to each room number involved in the expected work order in the coefficient array, accumulate the product of the coefficient value and the real-time dose rate corresponding to the room number, and finally obtain the E of the current maintenance work order. * / T * The estimated value of E * / T * Estimated value and expected working time T of the current maintenance work order * Multiply them together to get the estimated result value E of the radiation dose received by the maintenance personnel in the current maintenance order. * . The estimated value E * Added historical data samples of personnel exposure doses.

[0109] After testing, the historical personnel exposure dose data was used as the overall control sample, and the average value of the overall control sample was calculated. Estimated value and mean of personnel exposure dose data estimated by this method The comparison deviation percentage is about ±1.5%, which meets the high accuracy requirement. If the radiation dose of the associated historical work orders is added to the estimated radiation dose of the expected work orders as the overall sample, the proportion of the expected work orders deviating from the center of the overall sample mean is basically no more than ±1%.

[0110] Since the total personnel radiation dose of an overhaul is accumulated from the personnel radiation doses of several expected work orders, it is possible to further reduce the personnel radiation dose error of a single expected work order to reduce the overall error of the estimated personnel radiation dose of all work orders of an expected overhaul. In the calculation process of a single estimated work order, the estimated radiation exposure value of the expected work order is finally output by the algorithm. Combined with the historical similar data of the work order, the confidence interval of the estimated radiation dose of the expected work order is calculated.

[0111] Furthermore, step 4 also includes:

[0112] When the dose rate array {Rate i (j) When the dimension of} is greater than the number of maintenance work orders, a constant term and a unit matrix are introduced into the linear equation system as bias terms, and the ridge regression analysis method is used to solve the linear equation system to obtain the linear regression coefficients.

[0113] When the dimension of the linear regression coefficient array is greater than the number of selected maintenance work orders, the linear equation system will have multiple applicable solutions. In this case, the mean method is a more convenient method. However, this method considers the use of ridge regression analysis as a feasible solution to this type of non-unique solution problem and discusses it.

[0114] Since the radiation dose rate of the room involved in the maintenance work order selected by the previous overhaul data may remain basically unchanged, under this premise, the dose rate matrix corresponding to the linear operation {Rate i (j)} has a linear correlation with the column vectors, which may be a more common situation in the actual dose estimation process. In addition, it is also feasible to select a maintenance work order with the highest degree of similarity in description items from each overhaul data as historical data. However, there is a situation where, after the array of radiation rooms involved in the work order is aligned, the array dimension of the generated room dose rate array {Rate*(i)} is greater than the number of historical data selected. At this time, the dose rate matrix corresponding to the linear operation is often not full rank. In this case, solving the linear operation by gradient descent method may have problems with matrix operations.

[0115] In summary, it is a more reliable method to use the ridge regression analysis method as a supplement to the radiation dose estimation method. At this time, by introducing an additional constant λ and the unit matrix E as bias terms, a definite solution of the coefficient array obtained by linear operation is It can be expressed as:

[0116]

[0117] Where y is the linear combination value array of the linear equation system {E j / T i}, λ and E are temporary in Rate T , a diagonal matrix with a constant eigenvalue λ added to the Rate diagonal. Adding this extra bias term facilitates the adjustment of the intermediate matrix variable Rate T Rate+λE performs inverse operation (Rate T Rate+λE)- 1 .

[0118] In the above formula, the choice of λ value is usually related to the application scenario of ridge regression analysis method, and its selection direction usually includes two aspects: 1. By dividing the historical sample data into multiple data subsets, training and verifying the model on each subset respectively, to determine the relatively optimal λ value; 2. Directly set a smaller λ value to minimize the prediction error.

[0119] Since different λ values ​​only affect the degree of penalty for overfitting, if the historical maintenance work order data is not rich enough in time sequence, it is not enough to divide the subsets according to Direction 1 to complete training and verification. Then a smaller λ value can be selected according to Direction 2 to ensure Rate TThe reversibility of Rate+λE can ensure the output of a set of definite solutions. Since the radiation prediction algorithm of this embodiment is applied in the scenario where only the linear equations represented by the linear operation are solved, it is sufficient to directly select direction 2 and select a smaller λ value.

[0120] This application analyzes the distribution characteristics and structural features of the relevant data based on the work orders and exposure data information generated by previous overhauls, and proposes a radiation dose estimation method based on the idea of ​​multivariate linear regression. Before the work order is executed, the range of exposure doses that may be generated by the work order task is informed to the operation and maintenance personnel, providing a relatively reliable reference basis for actual radiation protection and exposure dose control of nuclear power plant personnel. The ultimate goal is to ensure the safe and stable operation of the nuclear power plant and the safety and health of the operators.

[0121] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed in the present application should be covered within the protection scope of the present application.

Claims

1. A method for estimating radiation levels of a nuclear power plant maintenance work order, characterized in that: include: Step 1: Collect data of historical maintenance work orders to establish a historical maintenance work order set, obtain comparison maintenance work orders from the historical maintenance work order set, and construct a comparison maintenance work order group; Step 2: Develop a linear equation related to radiation dose: In the formula, K i (j) To execute the maintenance work order R i The ratio of the working hours in the jth room to the total working hours, Rate i (j) To execute the maintenance work order R i The radiation dose rate to the jth room at time, E i For maintenance work order R i The radiation dose of a single person, T i For maintenance work order R i Total working hours; Step 3: Obtain the historical data of each maintenance work order for comparison, substitute the historical data into the linear equation, and construct a linear equation system; Step 4: Solve the linear equations to obtain linear regression coefficients, and calculate the estimated radiation dose of the current maintenance work order based on the linear regression coefficients.

2. The method for estimating radiation dose of a nuclear power plant maintenance work order according to claim 1, characterized in that: In step 1, a historical maintenance work order set is searched for a historical maintenance work order that performs the same task as the current maintenance work order as a comparison maintenance work order.

3. The method for estimating radiation dose of a nuclear power plant maintenance work order according to claim 1, characterized in that: In step 1, the mean and standard deviation of the working time data in the comparison maintenance work order group are calculated, a screening interval is set based on the mean and standard deviation of the working time data, and the comparison maintenance work orders with working time values ​​outside the screening interval are eliminated.

4. The method for estimating radiation dose of a nuclear power plant maintenance work order according to claim 1, characterized in that: Step 2 includes: Create a single maintenance work order R i Radiation dose per person E i Calculation formula: In the formula, E i For maintenance work order R i The radiation dose of a single person, T i (j) To execute the maintenance work order R i The working hours to go to the jth room, Rate i (j) To execute the maintenance work order R i The radiation dose rate to the jth room at time ; The above formula satisfies the formula in terms of time constraints: Where, T i For maintenance work order R i of working hours; The radiation dose for a single person is E i Divide both sides of the calculation formula by T i ,get: make The above formula is simplified to: K i (1) *Rate i (1) +K i (2) *Rate i (2) +…+K i (m) *Rate i (m) =E i / T i 。 5. The method for estimating radiation dose of a nuclear power plant maintenance work order according to claim 1, characterized in that: In step 2, a single work order R is executed for the historical maintenance work orders. i The corresponding room list and the corresponding radiation dose rate of the room at that time are respectively represented by {Room i (1) ,Room i (2) ,…,Room i (m) } and {Rate i (1) ,Rate i (2) ,…,Rate i (m) } indicates that the work order R i The work hours involved are expressed as a single T i , while executing the work order task, the radiation dose absorbed is E i , if work order R i The corresponding rooms are all zero dose rate, so the absorbed dose E corresponding to this work order task i is zero.

6. The method for estimating radiation dose of a nuclear power plant maintenance work order according to claim 1, characterized in that: The linear equations in step 3 are: K1 (1) *Rate1 (1) +K1 (2) *Rate1 (2) +…+K1 (m) *Rate1 (m) =E1 / T1 <h2 style=";text-align:left;direction:ltr">K2<h2 style=";text-align:left;direction:ltr"> (1) <h2 style=";text-align:left;direction:ltr"> *Rate2<h2 style=";text-align:left;direction:ltr"> (1) <h2 style=";text-align:left;direction:ltr"> +K2<h2 style=";text-align:left;direction:ltr"> (2) <h2 style=";text-align:left;direction:ltr"> *Rate2<h2 style=";text-align:left;direction:ltr"> (2) <h2 style=";text-align:left;direction:ltr"> +…+K2<h2 style=";text-align:left;direction:ltr"> (m) <h2 style=";text-align:left;direction:ltr"> *Rate2<h2 style=";text-align:left;direction:ltr"> (m) <h2 style=";text-align:left;direction:ltr"> (E2 / T2) … K n (1) *Rate n (1) +K n (2) *Rate n (2) +…+K n (m) *Rate n (m) =E n / T n In the formula, K i (j) To execute the i-th maintenance work order R i The time spent in the jth room on the maintenance work order R i The proportion of total working hours, Rate i (j) To execute the i-th maintenance work order R i The radiation dose rate to the jth room at time, E i is the i-th maintenance work order R i The radiation dose of a single person, T i is the i-th maintenance work order R i Total working hours; Based on the linear equation system, establish linear operations: In the formula, {K * (1) ,K * (2) ,…,K * (m) } is the linear regression coefficient to be solved.

7. The method for estimating radiation dose of a nuclear power plant maintenance work order according to claim 1, characterized in that: In step 3, the room set {R i } is the union of the rooms visited by different historical maintenance orders for the same task, {Rate i (j) } is the union {R i } is a dose rate array consisting of the dose rates of each room in .

8. The method for estimating radiation dose of a nuclear power plant maintenance work order according to claim 1, characterized in that: In step 4, solve the linear equations to obtain the linear regression coefficients, including: Establish the loss function J(θ): With the goal of minimizing the loss function J(θ), the linear equations are solved using the gradient descent method to obtain the linear regression coefficients.

9. The method for estimating radiation dose of a nuclear power plant maintenance work order according to claim 1, characterized in that: Calculating the estimated radiation dose of the current maintenance work order based on the linear regression coefficient in step 4 includes: substituting the linear regression coefficient value into the linear equation to obtain the radiation dose calculation formula of the current maintenance work order, obtaining the radiation dose rate and operation time data of the current maintenance work order, and substituting them into the radiation dose calculation formula to obtain the estimated radiation dose of the current maintenance work order.

10. The method for estimating radiation dose of a nuclear power plant maintenance work order according to claim 1, characterized in that: Step 4 also includes: when the dose rate array {Rate i (j) When the dimension of} is greater than the number of maintenance work orders, a constant term and a unit matrix are introduced into the linear equation system as bias terms, and the ridge regression analysis method is used to solve the linear equation system to obtain the linear regression coefficients.

Images