Assessment of the level of reliability and safety based on the index of the technical condition of the equipment of energy systems

,


Introduction
Determining the permissible interval of predictive trouble-free operation of power equipment to a greater extent makes it possible to determine the strategy of current operation management.
The complexity of forecasting changes in the technical condition of power equipment has different reasons:  The multiplicity and interrelation of various processes occurring in the operating equipment.
 The relationship of the physical processes occurring in the elements from the operating conditions and operating modes of the power equipment as a whole.
 Different rate of change in the condition of equipment at different stages of operation, which has a nonlinear dependence on the rate of development and self-development of defects.
 Technological failures embedded in the equipment due to design flaws and manufacturing technology.
 And others.In this regard, existing models do not allow us to reliably describe the forecast of changes in the condition of equipment, since the influence of unpredictable impacts and the rate of development of defects of various origins on the modeling results is quite large.

Materials and methods
Improving the reliability of power equipment is ensured by minimizing technological failures in human production and operational activities, which include both technical errors in the design of facilities associated, for example, with incorrect initial data, and inaccuracies in the manufacture of equipment.In addition, technological violations may occur due to violations of technical regulations during the operation of power equipment operating in non-calculation modes, experiencing overload.Thus, the multifactorial nature of technological processes, taking into account the inter-repair period, accident rate, metal condition, residual resource, will allow us to develop an approach to predictive assessment of the operability of each element in order to obtain a reliable and representative sample of failure statistics to assess the reliability of power equipment.To establish the required set of components of technological parameters for assessing reliability indicators, it is necessary to determine the composition of the element base with the ability to predict the resource of power plants using digital technologies, including neural connections and digital twins, to create unique digital energy systems.Monitoring of the technical condition of power equipment reduces the likelihood of technological violations and failures during operation of the equipment.
Faulty and unreliable equipment is restored by carrying out repairs to ensure the operability of the equipment by eliminating malfunctions and the consequences of wear of parts, as well as the causes that caused the malfunction or increased wear.In accordance with the Russian energy digitalization program, a single integral indicator is used to assess the technical condition and determine the physical wear of power equipment -the technical condition index (TCI) [1][2][3][4][5].
To assess the reliability of hydropower facilities [6], two indicators are used: the K g readiness coefficient and the K og operational readiness coefficient.The determination of their actual value is based both on the time when the unit is ready for operation and on its downtime.The following indicators are used to determine the operational readiness coefficient:  Duration of operation in generator mode. Duration of stay in reserve.
 Total downtime in repair. Calendar time.
The K g readiness coefficient differs from the K og operational readiness coefficient in that the planned repairs are taken into account, i.e. they are divided into forced and planned.
The authors [7] performed calculations of the operational readiness coefficient for K og for a number of HPPs, given in Table 1.The types of turbines, the country of manufacture, the design parameters of turbines, as well as the maximum values of the calendar time of operation of hydraulic turbines and the generation coefficient are given.The values of the generation coefficient and the operational readiness coefficient at the time of equipment replacement are also given.
Table 1.The value of the operational readiness coefficient of hydraulic units [7].The analysis of the data presented in Table 1 shows that the coefficient K og by the time of complete replacement of hydraulic equipment can be taken as a basic value, at which it is advisable to raise the question of timely reconstruction of hydraulic machines at the HPP.
Along with the availability coefficient, the main one for analyzing changes in the reliability of equipment is the coefficient of technical use of equipment -the proportion of the time the technical system is in working condition relative to the total duration of operation in a given time interval, including all types of maintenance.
Assessment of the technical condition of the main technological equipment is the process of determining the integral indicator of the technical condition (technical condition index -TCI) [8][9][10][11][12][13][14].
As a result of the assessment, the TCI of power equipment takes values in accordance with Table 2.
Specific values of TCI characterize the current state of power equipment, which will more accurately determine the timing of repairs, improve their quality, reduce the likelihood of emergency shutdowns, reduce the need for scheduled and emergency repairs.
The formation of a comprehensive indicator of energy security and energy efficiency of an energy campaign is carried out as an arithmetic mean of the point estimates of the nature of the situation for each energy security indicator presented in Table 2.
For the convenience of processing the received information, visualization of the level of reliability and safety is used, taking into account the physical wear of the equipment.Threshold values of the company's energy security indicators are proposed in order to determine their actual nature (normal (stable), pre-crisis, crisis) for each indicator presented in Table 3.
The proposed method of using a neural network model is an integrating link that unites all diagnostic monitoring systems operating at an energy enterprise into a single whole for predicting trouble-free operation of power equipment.The TensorFlow open source library for the Python programming language is used to train and use the neural network model.
The source data is presented in the form of special objects called dataframes, which can easily be obtained from files of the common csv format used, among other things, in the ExcelMicrosoft program.The practical coincidence of the histograms of training and test values means the correctness of the templates presented to the neural network used for its training.
Next, the original dataframes undergo "preprocessing" -a procedure that includes, firstly, the normalization of numerical data to bring the values of various parameters to a single range, and secondly, the replacement of lexical parameter values ("available", "absent", "single defect", etc.) with numerical values.
After the preprocessing is completed, the neural network model can be trained.At the same time, it is possible to set hyperparameters of the neural network model -the number of hidden layers, the number of neurons in each layer, the activation function, the learning method, the learning rate, the number of epochs of the learning algorithm (the number of iterations of optimizing neural network parameters).
For the final conclusion about the success of neural network training, a graph of the dependence of ITS values obtained by the neural network on TCI values from the training template is constructed.
The proposed neural network model is based on a methodology for assessing the reliability and calculation of TCI equipment, taking into account the controlled parameters of the object.To analyze the indicators, a choice of hydraulic power equipment was made: a hydraulic turbine (GT), which includes the following functional units: a guide device (NA), a turbine cover, a flow part, an impeller, a turbine bearing and a shaft.
To determine the vehicle of the selected functional units of the equipment, groups of controlled parameters are used in accordance with the methodology [3].At the same time, TCI evaluation is performed in stages in accordance with the above algorithm of the neural network model.A number of calculated parameters are given in Table 4.
During the maintenance of the equipment, the condition parameters are measured, which cause gradual failures resulting from the aging process, which worsens the initial parameters of the equipment.To assess the reliability indicators, the authors identified elements and components that characterize the performance of hydraulic power equipment, in accordance with the results of TCI evaluation.

Results and Discussion
In this regard, graphical dependencies have been obtained that allow us to obtain forecast values of ITS and reliability indicators depending on the conditions and operating modes of hydraulic equipment.
All the presented graphs visually demonstrate the relationship on TCI (the probability of equipment failure) of individual functional units of hydraulic turbine equipment and TCI hydraulic power equipment as a whole of two interdependent parameters.The degree of influence of individual parameters depends on a number of factors.First of all, it depends on the weighting factor of the parameter that affects the TCI of this group of parameters.Secondly, it depends on the weight coefficient of the functional node itself.And, thirdly, it depends on whether the selected parameters characterize one functional node or different ones.
The graph in Figure 1 shows a change in the probability of failure of the hydraulic power equipment as a whole and the functional unit "Hydroturbine" to a critical one with an increase in the pressure difference in the cavities of the servo motors of the guiding device (NA).
On the graph (curve 1) characterizes the change in the probability of failure of the "hydroturbine" node in the absence of a difference in the turn of the blades at the same opening ON after processing the signals to "add" and "subtract"; (curve 2) -the change in the probability of failure of the "hydroturbine" node with the maximum difference in the turn of the blades at the same opening on after working out the signals to "add" and "subtract"; (curve 3) -change in the probability of failure of hydraulic power equipment in the absence of a difference in the turn of the blades at the same opening ON after testing the signals to "add" and "decrease"; (curve 4) -change in the probability of failure of hydraulic power equipment at the maximum difference in the turn of the blades at the same opening on after processing the signals to "add" and "subtract.The graphical dependence shows the effect of the synchronicity of the turning of the feathers on the general technical condition of the guiding device, which is confirmed by the pressure difference in the cavities of the servo motor and allows you to unambiguously determine the cause of defects.In the future, according to the indications of the pressure difference, it will be possible to judge the development of the defect and will determine the critical condition of the equipment.
The graph in Figure 2 shows the change in the TCI of the hydraulic unit as a whole and its functional unit "Hydroturbine" with an increase in the pressure difference in the cavities of the servomotors and the lack of synchronicity of the blades turning ON.On the graph (curve 1) characterizes the change in the TCI of the "hydroturbine" node in the absence of a pressure difference in the cavities of the servomotors on; (curve 2) -change of ITS of the "hydroturbine" node with an unacceptable pressure difference in the cavities of servomotors on; (curve 3) -change of TCI hydraulic power equipment in the absence of pressure difference in the cavities of servomotors on; (curve) -change of ITS hydraulic power equipment with an unacceptable pressure difference in the cavities of servomotors ON.The combined effect of increasing the pressure difference in the cavities of servomotors and increasing the pressure difference in the cavities of servomotors significantly reduces TCI BY increasing the probability of failure up to 55%.
A noticeable complex effect of these two parameters arises due to the fact that both indicators relate to the evaluation of one functional node.
To assess the effect of vibrations in a turbine bearing, the probability of its trouble-free operation was calculated depending on the technical condition on the example of a number of parameters by which the TCI of this node is determined.
Figures 3 and 4 show the dependences that characterize the influence of the turbine bearing condition on the vibration characteristics in the area of the support bearing (podpyatnik) at various loads of the hydraulic unit.Due to the fact that the selected parameters characterize different functional nodes, the graph shows that the turbine power does not affect the TCI of the node.With an increase in vibration in the turbine bearing area to the maximum permissible value, the TCI of the turbine bearing decreases.
Figure 4 shows the effect of the shaft battle on the probability of failure of the hydraulic unit and the Turbine bearing and Shaft assembly.On the graph (line 1) characterizes the change in the probability of failure of hydraulic power equipment at 100% turbine power; (line 2) -a change in the probability of operation of hydraulic power equipment at 83% turbine power; (line 3) -a change in the probability of failure of the hydro turbine unit.
In order to assess the effect of shaft vibrations together with the development of cavitation erosion of the blades on the TCI of the impeller, the probability of its trouble-free operation was calculated.
Figures 5 and 6 show graphs that characterize the effect of vertical vibration of the hydraulic unit shaft during the development of cavitation erosion of the blades.These parameters also relate to different functional nodes, which makes it impossible to assess their cross-influence on the corresponding functional node.However, it allows us to identify the mutual influence on the general condition of hydraulic power equipment.
In graph 5 (curve 1) characterizes the change of TCI node "flow part of the turbine"; (curve 2) -the change of TCI hydraulic power equipment in the absence of shaft vibration; (curve 3) -change of TCI hydraulic power equipment at maximum vibration of the shaft.When the vibration of the shaft increases to the maximum value, the TCI of the impeller decreases significantly, which is confirmed by the data presented in Figure 6.In Graph 6 (curve 1) characterizes the change in TCI hydraulic power equipment in the absence of cavitation erosion of the blades of the RC; (curve 2) -a change in ITS hydraulic power equipment with unacceptable cavitation erosion of the blades of the RC; (curve 3) -a change in TCI node "Hydraulic turbine".
The analysis of the presented dependencies allows us to conclude that the effect of vibration of the shaft of the hydraulic unit during the development of cavitation erosion of the blades reduces the TCI of hydraulic power equipment from 90 to 80.
The neural network model allows us to evaluate analytical dependencies characterizing the influence of specific parameters on the TCI of individual nodes and power equipment, as well as the joint influence of a number of GA nodes on the TCI and the overall technical condition of the hydraulic unit.

Conclusion
 It is proposed to use a neural network calculation model to evaluate and predict the reliability indicators of power equipment and evaluate their impact on the main reliability indicators of the generating system.
 Comparative assessments of the reliability of power equipment based on the presented mathematical dependencies and data obtained on the basis of neural network modeling have been carried out.
 Estimates of the influence of a number of parameters of the components and elements of the HA on their ITS, on the general technical condition of the hydraulic unit, which showed that the most confirmed are the combined effect of an increase in the pressure difference in the cavities of servomotors and an increase in the pressure difference in the cavities of servomotors, the combined effect of shaft vibrations with the development of cavitation erosion of the blades.In this regard, it is necessary to create modern digital control systems of power equipment necessary for the continuous collection, storage, archiving of data, taking into account the actual condition of a particular element of a hydraulic unit.
To build a neural network, a topology was developed, a learning and testing mechanism was defined.During the research, a sample of input data was created, an algorithm was built.Training and test data files have been created to train the neural network model.The volume of the training file (training template) is 6000 records, each of which corresponds to any possible combination of parameters characterizing the technical condition of the object.The test file with a volume of 1000 records is formed similarly to the training template, but this set is not presented to the model at the training stage.It is used to evaluate the correctness of training a neural network model, implementing the principle of the so-called "deferred sampling".

Fig. 1 .
Fig. 1.The effect of the rotation of the blades on the change in the probability of failure of the hydraulic power equipment as a whole and the hydraulic turbine.

Fig. 2 .
Fig. 2. Impact of turning of blades change in probability of failure of hydraulic power equipment in general and hydraulic turbine.

Fig. 3 .
Fig. 3.The impact of the shaft battle on the index of the technical condition of the Turbine bearing and shaft assembly.

Figure 3
Figure 3 shows the effect of the state of the turbine bearing on the shaft fight when the power of the hydraulic turbine changes from 83 to 100%.On the graph (line 1) characterizes the change in TCI of the Turbine Bearing and Shaft node; (line 2) -the change in TCI of hydraulic power equipment at 100% turbine power; (line 3) -the change in TCI of hydraulic power equipment at 83% turbine power.With an increase in vibration in the turbine bearing area to the maximum permissible value, the TCI of the turbine bearing decreases.Figure4shows the effect of the shaft battle on the probability of failure of the hydraulic unit and the Turbine bearing and Shaft assembly.

Fig. 4 .
Fig. 4. The impact of the shaft battle on the probability of failure of the hydraulic unit and the Turbine bearing and shaft assembly.

Fig. 5 .
Fig. 5. Effect of changes in the cavitation erosion of the blades on TCI hydraulic power equipment and the flow part of the turbine.

Fig. 6 .
Fig. 6.The effect of shaft vibration changes on ITS hydraulic power equipment and turbine covers.

Table 2 .
Color indication of assessment of reliability and safety levels of power equipment.

Table 3 .
Threshold values of energy security indicators.

Table 4 .
Initial data for assessing the technical condition of hydraulic power equipment.Class "Hydraulic turbine".