基于不确定性用气量的输气管网供气可靠度计算方法

由于未来用气量是一个随机变量,过去采用点预测的方式无法客观体现出其随机性的特征,由此得到的供气可靠性评价结果也难以客观反映实际情况。为了准确预测用气量,在调研输气管网供气量预测模型研究进展的基础上,提出了一种采用基于小波分解的神经网络模型来预测随机性用气量的供气可靠度计算方法,通过对比分析模型预测结果与实际用气量的误差,确定用气量所服从的分布类型及参数,结合管道的最大允许输气量,以供气可靠度最高作为目标函数,建立了优化流量分配及计算管网供气可靠度的数学模型,进而利用该模型对某一虚拟管网进行了供气可靠性评价。结果表明:(1)所建模型求解的流量分配方案优先保障权重较大/较为重要的用户,但会牺牲其他用户的供气可靠度,管网总体供气可靠度亦会下降;(2)取消权重后,管网总体供气可靠度提高,优先保障距离气源地较近的用户,若不要求完全满足用户的用气量,则在降低一定的标准后,所有用户的供气可靠度都能达标。结论认为:所提出的计算方法结合了用气量随机的特性,能够更加客观地评价管网的供气可靠度,同时由于引入了用户的权重,在计算管网供气能力的时候能优先满足重要的用户,更加符合实际情况,其评价结果可以指导输气管网的高效运行。

A computation model for gas supply reliability analysis in a gas pipeline network based upon the uncertainty of gas consumption

Because the future gas consumption is a random variable, the previously used point forecast method cannot fully embody its randomicity, from which the resulted gas supply reliability assessment fails to objectively reflect the practical situation. In view of this, we first investigated the research progress in gas supply forecast modeling in gas pipeline network systems, and thus presented a gas-supply reliability calculation methodology for predicting the random gas consumption assisted with artificial neural network based on wavelet decomposition. Through an error analysis between the forecasted results from the presented method and the actual gas consumption, we determined the distribution types and parameters with which the gas consumption complies should comply. Then, in combination with the maximum allowable gas flow rate and with the maximum gas-supply reliability as objective function, we established a mathematical model of optimizing flowrate distribution and calculating the gas-supply reliability, and finally applied this model to the analog computation and analysis of a simulated gas pipeline network. The following findings were obtained. (1) The flowrate distribution scheme resulted from the model gives priority to more important gas users or those with greater weight coefficients but this will sacrifice other users' gas-supply reliability, as a result, this will lower the overall reliability of the gas network system. (2) Without considering the weight factor, the overall reliability will be improved to guarantee such gas users closer to gas source locations; thus if it is unnecessary to fully satisfy all the gas users, the overall reliability will reach the standard which can be possibly lowered to some degree. It is concluded that this calculation methodology combined with the randomicity of gas consumption can objectively assess the gas supply reliability of a gas pipeline network system and due to the introduction of the weight of users, more important users will be satisfied preferentially while the gas supply capacity is computed, as a result, this will conform to the actual situation and the assessment results will be taken as reference to the high-efficiency operation of gas pipeline network systems.