This study examines the challenges encountered when using wireless technologies based on random multiple access for voltage monitoring in low voltage (LV) electrical grids. The introduction of photovoltaic modules on the roofs of buildings creates the need to monitor the dynamics of the node voltages of the grid. We consider two related objects -an LV electrical grid and a communication network with devices in the nodes of the grid. The communication network carries out a one-way transmission of monitoring data from the devices to the LV grid operator and is based on LoRaWAN technology. Transmission is carried out via a common data transmission medium. We presume that each message contains information about the node voltage. The messages from each node in the grid are sent to the grid operator at random time intervals and independently of each other. If the airtimes of messages from two or more nodes overlap, a collision occurs and none of the messages reach the operator. The operator has to monitor the random voltage behavior over time to avoid exceeding a certain level (an upper voltage limit). Under the specifics of the voltage random process, we study the problem of choosing parameters of message transmission to uniformly minimize the probability of such excess among all nodes. The work precisely formulates the optimization problem and proposes an algorithm for its solution for the particular case.