Space Model Fee Estimation of Random Uncertainties Optimization Comprehensive
Under complex technology and external environmental conditions, random uncertainty involved in three cost estimates and characterization parameters - - Standard poor. Through the smallest linear variance combination, three cost estimates are achieved. Optimization of random uncertainty. The corresponding solution method is also given to the case of the non-zero association between the method.
The expected value of the three estimation results and its random uncertainty
x 1 , x 2 , x 3 estimated expected value of the CER model, qualitative prediction, and risk expense estimation, according to regression analysis, CER model The uncertainty produced in random disturbances can be represented by the standard deviation of its distribution. Sub> 1 . In a qualitative method, a plurality of predicted values of a set of parameters (number of test times and seconds) are given. According to the statistical analysis of the sample, the standard deviation of the random uncertainty can be used in the standard difference E 2 To represent. In the risk analysis method, the probability value is obtained on the basis of the historical data of various models, so the random uncertainty of its estimated value can also be used by its corresponding distribution of standard differences E 3 is expressed.
Optimization of the three estimates and its random uncertainties
existing x 1 , E 1 , x 2 , e 2 , x 3 , e 3 The estimated value and standard deviation of three methods, random uncertainty of three methods are generally obedient. Studies have shown that the development of aerospace models and various divisions have formed a number of relatively independent cost expenditure units. When the number of units is 10 or more (the actual number is mostly 10 or more), most models full system cost estimation output is approximately Normal distribution.
Three results are performed according to estimation results of respective methods ( x i ) and accuracy (E i ) Comprehensive and desirable results: Estimation accuracy can be improved and unpiased, ie: The final result can be represented by a comprehensive expected value, and its variance is minimized.
Indicates that the integrated process can reduce the uncertainty of the empty space model estimation, that is, E should be less than E 1 , e 2 , e < SUB> 3 . It should be noted that the improvement of the E value is not the effect of the uncertainty factor in reality, and the objective situation will not be naturally better due to the improvement of the method, nor does it show the respective uncertainty of the three methods. It was reduced. But because the cost estimation value is more objective after optimization of the comprehensive process, the uncertainty of the final estimate is reduced, which makes the cost estimates more accurate.
Voltage Monitoring Point Optimization Configuration of Fault Radio Uncertainties
When a short circuit fault occurs, the fault resistance objective is present and has a random uncertainty. In order to improve the engineering applicability of the voltage suspected monitoring point configuration scheme under different fault resistance conditions, the key parameters of the fault resistor will be introduced into the monitoring point optimization configuration model. First, a critical fault resistance matrix is constructed based on network parameters and short-circuit computing, on this, based on the minimum number of monitoring points, the minimum voltage temporary scandals is constrained by the short circuit of each fault point, and the monitoring point for the random uncertainty of fault resistors is established. Optimize the configuration model and solve the optimal configuration scheme with the genetic algorithm. Application This method simulates the IEEE30 node test system, the results show that the method can effectively monitor the voltage temporary drop caused by non-metallic short-circuit failures, compared to traditional methods more engineered practical value.
Crown fault resistance matrix
For a short circuit fault point in the system, various fault critical fault resistors corresponding to the voltage threshold VTH can be obtained according to the formula. For asymmetric short-circuit faults, the maximum value in three-phase critical fault resistors is taken as a critical fault resistance value of the fault point.
Defining a critical fault resistance matrix formula, t is a fault type; n is the number of fault points set in the whole network; B < / i> is the number of whole network nodes. R ' cri any element T CRI, IJ takes the value of i > T When short-circuit fault, the critical fault resistance corresponding to the node j .
Monitoring point Optimization Configuration Model
The same is based on D as a decision vector, a decision plan corresponding to d Fault resistance matrix
r t D any element r t T / sup> D, IJ value is
r t d, ij = r t c ri, ij d j
Defining voltage temporary detectivity is
Voltage temporary reduction rate = number of voltage temporarons is monitored ÷ Short-circuit fault causes voltage temporary total failure resistance with random uncertain characteristics, if you use F ( R T i ) Indicates T When short-circuit fault, fault resistor r t i is probability density function, a ' CRI, I and a ' d, i represent r cri and r t D of i row element, r ' min < / sub> represents the minimum of the short circuit fault resistance of T .
Cross Probability and Variation Probability By adaptation changes, and minimize individuals, cross probability, and variation probability of individual monitoring points, so that the excellent individual is more easily protected into the next generation. For individuals, intersection probability and variation probability of total monitoring points, the individual is easier to phase out. Consider the voltage suspense monitoring point optimization configuration model for the fault resistor random uncertainty.
Start → Enter System Parameters → Based on System Parameters and Short Calculation Build Critical Fault Resistance Matrices → Establishing Public Notice IMF in accordance with critical fault resistance matrix and fault resistor random distribution characteristics → at least the number of monitoring points, based on Genetic Algorithm Solving the Excellent Configuration Scheme of Monitoring Points to Meeting inequality Constraints → End.